Genetic instrument selection for Mendelian randomization explorations of drug target effects

I read with interest the recent work by Zhouet al., where the authors aimed to explore the causal relationship between metformin use and risk of lung cancer with Mendelian randomization (MR).1 Growth differentiation factor 15 (GDF-15) circulating levels increase as a result of metformin use. Following this rationale, the authors use single nucleotide polymorphisms (SNPs) associated with GDF-15 levels as genetic proxies (instruments) for metformin use. Employing two-sample MR, which uses information on genetic variants from one sample and explores associations with outcomes in another independent sample, the authors then explore associations between these SNPs and risk of lung cancer and find no significant associations. Based on this analysis, the authors conclude that metformin use does not cause lung cancer.
Source: International Journal of Epidemiology - Category: Epidemiology Source Type: research