In this chapter, we consider the interpretation of causal effects assessed and estimated in Mendelian randomization, and address the question of under what circumstances a Mendelian randomization estimate may be a reliable guide to the effect of an intervention on the exposure of interest in practice.


In Mendelian randomization, differences in the exposure distribution due to genetic variation are materially distinct from the change due to any proposed therapeutic intervention on the exposure, and so may affect the outcome differently. Consequently, it may be misleading to generalize the magnitude of a Mendelian randomization association to the effect of a potential intervention on the exposure in practice. Awareness of this is important for the use of Mendelian randomization in target-based drug development.

Relevant papers to chapter:

S. Burgess, A. Butterworth, A. Malarstig, S.G. Thompson. Use of Mendelian randomisation to assess potential benefit of clinical intervention. Brit. Med. J. 2012; 345:e7325.

S. Burgess, A. Malarstig. Using Mendelian randomization to assess and develop clinical interventions: limitations and benefits (editorial). J. Comp. Eff. Res. 2013; 2(3):209-212.