Accelerometer-Measured Moderate-to-Vigorous Physical Activity and Cancer Risk: Dose–Response from Observational and Nonlinear Mendelian Randomization in UK Biobank

Background: Physical activity is a well-established modifiable lifestyle factor associated with reduced cancer risk; however, the optimal weekly volume of moderate-to-vigorous physical activity (MVPA) for cancer prevention, particularly when assessed using objective measures, remains unclear. Most existing evidence relies on self-reported physical activity, which may introduce measurement bias and obscure accurate dose–response relationships. Methods: We analyzed data from UK Biobank participants with valid accelerometer measurements to quantify habitual MVPA. Observational associations between MVPA and incident cancer were evaluated using multivariable Cox proportional hazards regression and restricted cubic splines. One-sample Mendelian randomization (MR) analyses, including both linear and nonlinear approaches, were conducted to evaluate potential causal associations and explore possible dose–response patterns. Results: Higher MVPA was associated with lower total cancer risk (HR 0.971, 95% CI 0.954–0.988, p = 0.001). Consistent associations were observed for several site-specific cancers, particularly lung, colorectal, breast, kidney, and bladder cancer. MR analyses supported a directionally consistent association between genetically predicted MVPA and lower total cancer risk (HR 0.977, 95% CI 0.962–0.992, p = 0.002). Nonlinear MR analyses suggested a potential nonlinear association, with lower cancer risk observed at a model-derived exploratory point of approximately 5 h of weekly MVPA. Conclusions: These findings provide supportive evidence that higher accelerometer-measured MVPA is associated with lower total cancer risk and contribute to a better understanding of the dose–response relationship between MVPA and cancer incidence.