Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 3 of the 3 references it cites were found in a registry or an index.

Mendelian Randomization Studies in the Elderly

Boef, Anna G. C., le Cessie, Saskia, Dekkers, Olaf M. · 2015 · Epidemiology

Abstract

To the Editor: Mendelian randomization studies use genetically determined variation in exposure levels to study causal effects. Unmeasured confounding and reverse causation, which hamper the estimation of causal effects of exposures in conventional analysis, can thereby be circumvented.1 The genetic variant used as an instrument should fulfill the following conditions: (1) it is associated with the exposure; (2) it affects the outcome only through the exposure; and (3) it is not related to other factors that affect the outcome.1–3 If these main assumptions are fulfilled, the effect of the genetic variant on the outcome can be attributed to the effect of the exposure on the outcome. An overview of numerous scenarios in which these assumptions are violated was published recently, with a discussion of the consequences and suggestions for alternative approaches to Mendelian randomization studies.3 Another scenario in which the assumptions could be violated is if a Mendelian randomization study is performed in an elderly population. Let us think of a Mendelian randomization study in which the effect of exposure X on outcome Y is studied in a population aged > 80 years, using genetic variant G as the genetic instrument to circumvent the confounding by unmeasured factors U1 of the X–Y relation. If the genetic variant G, through its effect on the exposure X has affected survival up to age 80, collider stratification bias (selection bias) may occur. This is illustrated in the Figure: both genotype G (through exposure X) and other risk factors U2 affect survival up to age 80. Survival S is therefore a collider, and restriction of the population to those who have survived up to age 80 results in collider stratification bias by inducing an association between G and U2. The intuitive interpretation of this phenomenon is as follows. We assume that genotype G increases mortality rates. If a person with a genotype G is still alive at age 80, this person will be less likely to have other risk factors for mortality (high blood pressure, smoking, etc.) compared with people without genotype G. This means that in the population aged over 80 the genetic variant is associated with other factors that affect the outcome, violating assumption 3. The effect of the genetic variant G on the outcome Y can therefore no longer be solely attributed to the effect of the exposure X on outcome Y.FIGURE: Genetic variant G is an instrument for the effect of exposure X on outcome Y, with unmeasured factors U1 confounding the X–Y relation. Both X and unmeasured risk factors U2 affect survival S, and U2 also affects Y. Because of selection on S, there is an association between G and U2.An example in which this collider stratification bias might occur is a Mendelian randomization study in subjects aged over 80 using APOE variants (G) as an instrument for cholesterol level (X), with (for example) myocardial infarction as the outcome (Y). APOE variants are known to cause variation in cholesterol levels. Because cholesterol levels will have influenced survival up to age 80 (S) and hence selection into the study population, an association of APOE variants with other risk factors that influenced survival up to age 80 (U2, eg, smoking) is introduced. For example, among those with a cholesterol increasing APOE-variant who have survived up to age 80, there will be fewer smokers than among those without the variant. If these other factors also affect the risk of myocardial infarction, the genotype–outcome relation will be biased; in this example, the bias in the estimated effect of genetically increased cholesterol will be towards a lower risk of myocardial infarction due to the inverse relation with smoking. This also applies if the outcome is survival (from age 80). The bias introduced by selection on survival will, of course, be most prominent for Mendelian randomization studies investigating exposures that strongly affect survival. Anna G. C. Boef Saskia le Cessie Olaf M. Dekkers Department of Clinical Epidemiology Leiden University Medical Centre The Netherlands [email protected]

Overview

Integrity signals

Registry

Crossref · 10.1097/ede.0000000000000243

Cites

3 of 3 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Verification

Claims the paper makes

This paper’s full text isn’t openly available, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 3 references

Cited by

See all 39

Metadata

Authors
Boef, A. G. C., le Cessie, S., & Dekkers, O. M.
Affiliations
Leiden University (NL)
Venue
Epidemiology
Published
2015-03
Volume
26
Issue
2
Pages
e15-e16
ISSN
1044-3983 (ISSN-L)
OpenAlex
W2332303329
Cited by
39
References
3
Citations by year
2025: 1 · 2024: 5 · 2023: 8 · 2022: 2 · 2021: 2 · 2020: 3 · 2019: 5 · 2018: 3 · 2017: 6 · 2016: 4
FWCI
2.96
Type
journal article
Language
en
Open access
closed
Locations
  • Epidemiology (publishedVersion)
  • PubMed (publishedVersion)
  • Data Archiving and Networked Services (DANS) (submittedVersion)
  • Leiden Repository (Leiden University) (submittedVersion)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetic Associations and Epidemiology
  • Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
Keywords
Mendelian randomization, older adults, collider stratification bias, selection bias, unmeasured confounding, reverse causality, genetic instruments, exposure-outcome association
MeSH terms
Age Factors, Aged, Aged, 80 and over, Genotype, Humans, Epidemiologic Research Design, Bias, Causality, Mendelian Randomization Analysis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Mendelian Randomization Studies in the Elderly. https://app.cochat.ai/papers/10.1097/ede.0000000000000243
The paper
Boef, A. G. C., le Cessie, S., & Dekkers, O. M. (2015). Mendelian Randomization Studies in the Elderly. Epidemiology. https://doi.org/10.1097/ede.0000000000000243

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.