From randomized controlled trial to Mendelian randomization:principles and applications of instrumental variables
10.3969/j.issn.1672-8467.2025.04.021
- VernacularTitle:从随机对照试验到孟德尔随机化:工具变量的原理与应用
- Author:
Xing LIU
1
;
Wei-bing WANG
1
Author Information
1. 复旦大学公共卫生学院流行病学教研室-公共卫生安全教育部重点实验室 上海 200032
- Publication Type:Journal Article
- Keywords:
Mendelian randomization(MR);
instrumental variable(IV);
randomized controlled trial(RCT);
directed acyclic graph(DAG)
- From:
Fudan University Journal of Medical Sciences
2025;52(4):611-616
- CountryChina
- Language:Chinese
-
Abstract:
Mendelian randomization(MR)is an epidemiological method that uses genetic variants as instrumental variables(IVs).This article begins by elucidating the basic principles of IVs through causal diagrams,directed acyclic graph(DAG),and then systematically reviews the key assumptions,common challenges,and practical considerations in MR studies by comparing them with randomized controlled trials(RCTs).As an observational study,MR can theoretically help avoid confounding,but its validity in practice critically depends on confirming(ⅰ)a strong association between the genetic instrument and the exposure,(ⅱ)independence of the instrument from confounders,and(ⅲ)exclusion of alternative pathways from the instrument to the outcome;over-interpretation of findings should therefore be avoided.The paper also highlights the value of MR in drug target discovery and in exploring disease mechanisms,while reminding investigators that common issues in observational studies,such as the temporal order of exposure and outcome,still require rigorous attention.
