Methodological Challenges in Mendelian Randomization Analysis

• Activity Number: 149
• Type: Invited
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Tyler J VanderWeele*+
• Companies: Harvard School of Public Health
• Keywords: Causal inference ; Instrumental Variables ; Mendelian Randomization ; Sensitivity Analysis

Abstract

Critical attention is given to the assumptions underlying Mendelian randomization (MR) analysis and their biological plausibility. Several scenarios violating the MR assumptions are described including settings with inadequate phenotype definition, time-varying exposures, the presence of gene-environment interaction, the existence of measurement error, the possibility of reverse causation, and the presence of linkage disequilibrium. Data analysis examples are given illustrating that the use of instrumental variable (IV) techniques when the MR assumptions are violated can lead to biases of enormous magnitude. The original proposal of Katan (1986) for MR was not to use IV techniques to obtain estimates, but merely to examine genotype-outcome associations to test for the presence of an effect of the exposure on the outcome. It is shown that this more modest goal and approach can circumvent many, though not all of, the potential biases described. Further discussion is given to the use of sensitivity analysis in evaluating the consequences of violations in the assumptions and to the possibility that negative, rather than positive, MR results may turn out to be more reliable.