Fig. 1
From: Osteoporosis and osteoarthritis: a bi-directional Mendelian randomization study
Workflow of bi-directional MR analysis. A, the fundamental idea of MR analysis: If we cannot randomize the exposure, we can find a randomized instrumental variable to disentangle. â‘ Instrumental variables were highly correlated with exposure factors. â‘¡ Instrumental variables were independent of the outcome. â‘¢ Instrumental variables were not correlated with confounding factors. B, Workflow of our bi-directional MR analysis. MR: Mendelian randomization; BMD: bone mineral density; OA: osteoarthritis; FN: femoral neck; LS: lumbar spine; TB: total body