Two-sample Mendelian randomization: avoiding the downsides of a powerful, widely applicable but potentially fallible technique

Summary

This 2016 review by Hartwig and colleagues examines two-sample Mendelian randomization, a genetic epidemiological method for inferring causality from observational data. The authors articulate the power and broad applicability of the technique whilst systematically identifying assumptions that may be violated in practice—including pleiotropy, population stratification, and weak instrument bias—and propose safeguards and sensitivity analyses to mitigate fallible inferences. The paper serves as a methodological guidance document for researchers seeking to apply or interpret MR findings robustly.

UK applicability

This methodological framework is directly applicable to UK biobank and cohort researchers using genetic data to examine food-system and health associations. The guidance on assumption-checking and bias mitigation is essential for UK-based studies employing MR to investigate causal links between dietary exposures, nutrient status, and health outcomes.

Key measures

Methodological assessment of MR assumptions (instrumental variable validity, no pleiotropy, no confounding); evaluation of bias sources; comparison of sensitivity analyses and robustness checks

Outcomes reported

The paper examines the methodological foundations and potential pitfalls of two-sample Mendelian randomization (MR) as a causal inference technique in epidemiological research. It addresses sources of bias, violation of core assumptions, and strategies to strengthen the validity of MR findings.

Topic tags