You are here

Article Text

Article menu
Download PDFPDF
EBM learning
How to interpret studies using Mendelian randomisation
  1. Min Seo Kim1,
  2. Minku Song1,
  3. Jae Il Shin2,3,4,
  4. Hong-Hee Won1,5
  1. 1 Samsung Advanced Institute for Health Sciences & Technology (SAIHST), Sungkyunkwan University, Samsung Medical Center, Seoul, Korea (the Republic of)
  2. 2 Department of Pediatrics, Yonsei University College of Medicine, Seoul, Korea (the Republic of)
  3. 3 The Center for Medical Education Training and Professional Development, Yonsei Donggok Medical Education Institute, Seoul, Korea (the Republic of)
  4. 4 Meta-Research Innovation Center at Stanford (METRICS), Stanford University, Stanford, California, USA
  5. 5 Samsung Genome Institute, Samsung Medical Center, Seoul, Korea (the Republic of)
  1. Correspondence to Professor Hong-Hee Won, SAIHST, Seoul 06351, Korea (the Republic of); wonhh{at}skku.edu; Professor Jae Il Shin, Department of Pediatrics, Yonsei University College of Medicine, Seoul, South Korea; SHINJI{at}yuhs.ac

https://doi.org/10.1136/bmjebm-2022-112149

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction to Mendelian randomisation

    The importance of causal inference in biomedical science cannot be overstated. It can help determine the direction of effect, the mechanisms behind observed associations and potential interventions.1 Although observational studies are the most common type in the biomedical field, they primarily suggest an association, and less likely infer causation, because of the potential for confounding and bias associated with this type of study. Recent advances in genetic epidemiology, such as new statistical techniques, large genotyped datasets and advanced computational methods for data analysis, have enabled the rapid development of methods such as Mendelian randomisation (MR), which can strengthen causal inference in observational research.1

    MR uses genetic variants to assess whether an observed association between a risk factor and an outcome is consistent with a causal relationship.1 2 This is achieved by examining the random assortment of genetic variants that occurs during meiosis. This process conceptually resembles the random allocation of participants in a randomised controlled trial (RCT).1 2 The basic concept of MR is straightforward: genetic variation between people can lead to differences in their health outcomes without any confounding or reverse-causation bias associated with observational methods.3 4 For instance, people who carry variants associated with a risk factor (ie, genetic variants associated with body mass index (BMI)) are compared with those who do not, in terms of the development of an outcome (eg, coronary heart disease (CAD)).2 Since these genetic variants were determined before exposure to any environmental factors, genetic instruments are usually not influenced by confounding factors and any difference in the risk of CAD could be attributed to the difference in BMI.1 2

    To conduct a solid MR study, researchers should evaluate the MR assumptions before they conduct the analysis. The relevance, independence and exclusion restriction assumptions are three key …

    View Full Text

    Footnotes

    • JIS and H-HW are joint senior authors.

    • Twitter @MinSeoKim_MD

    • Contributors MSK, JIS and H-HW contributed to the concept and design of this study. MSK drafted the manuscript. JIS and H-HW supervised the study. MSK, MS, JIS and H-HW participated in the critical revision of the manuscript.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Disclaimer All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare the following: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

    • Competing interests None declared.

    • Provenance and peer review Commissioned; externally peer reviewed.