Genetic evidence does not support causal associations between brain structures and fall risk: evidence from Mendelian randomization.

Background falls are a growing global public health concern that requires our attention and ongoing exploration. Observational studies suggest that changes in certain brain structures are associated with falls and risk factors influencing falls, yet the causal relationships between different brain structures and falls remain unclear. Methods using large publicly available genome-wide association studies of predominantly European ancestry, independent genetic variants associated with hippocampal volume, parietal lobe volume, cerebellum I-IV volume, brainstem volume, and cerebral white matter integrity were selected as instrumental variables to investigate causal relationships with falls. Univariate Mendelian randomization analyses were performed using the inverse-variance weighted, Egger regression, weighted median, simple mode, and weighted mode methods. Multivariable Mendelian randomization analyses were conducted using the inverse-variance weighted, Egger regression, least absolute shrinkage and selection operator, and the median approach. Results univariable MR suggested nominal associations of falls with right cerebellum I–IV (OR = 1.011, 95% CI [1.001, 1.020], p = 0.023), after adjusting for outliers, a significant association was found between brainstem volume and falls (OR = 1.0079, 95% CI [1.0010, 1.0149], p = 0.024). However, these associations did not remain significant after FDR correction. MVMR analyses were further limited by weak instrument strength, precluding reliable estimation of direct, independent effects. Conclusions our genetically informed analysis does not support a direct causal link between brain structures and fall risk, suggesting that previously observed associations may be largely attributable to confounding rather than neuroanatomical causation.