Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Comment
. 2016 Jun;45(3):908-15.
doi: 10.1093/ije/dyw127. Epub 2016 Jul 17.

Commentary: Two-sample Mendelian randomization: opportunities and challenges

Debbie A Lawlor  1
Affiliations
Comment

Commentary: Two-sample Mendelian randomization: opportunities and challenges

Debbie A Lawlor. Int J Epidemiol. 2016 Jun.
No abstract available

PubMed Disclaimer

Figures

Figure 1.
Figure 1.
DAG of instrumental variable analyses in an RCT and MR study exploring the effect of LDLc on CHD. These are directed acyclic graphs (DAGs), thus the absence of an arrow between any two variables (nodes) indicates we do not consider it plausible that there is a causal effect between those two. Figure shows DAGs of instrumental variable (IV) analyses to test the causal effect of low-density lipoprotein cholesterol (LDLc) on CHD. In Figure 1 a and b, the IV is randomization to receiving a statin or not (i.e this is an example of IV analyses in an RCT). In Figure 1 c and d, the IV is genetic variants that are robustly related to LDLc (i.e. this is a Mendelian randomization study). Figure 1 a and c both illustrate the three key assumptions of IV analyses:

  1. i. that the IV ‘Z’ (randomization to statins in Figure 1 a and genetic variants related to LDLc in Figure 1 c) is (or is plausibly) causally related to the risk factor (LDLc in all figures);

  2. ii. that confounding factors for the risk factor-outcome ‘X’-’Y’ association (here LDLc on CHD in all figures) are not related to the instrumental variable;

  3. iii. that the instrumental variable ‘Z’ only affects the outcome ‘Y’ (CHD) through its effect on the risk factor ‘X’ (LDLc).

In the RCT example we know that assumption (i) is true, and if the RCT is well conducted, then assumption (ii) will be true (other than chance associations). If, however, statins are directly (independently of LDLc) related to other factors which then affect CHD, assumption (iii) will be violated and the estimated causal effect will be a biased estimate of the true effect of LDLc. There is some evidence that statins do relate to a wide range of lipids and fatty acids in addition to LDLc, though whether these are caused by the statins independent of LDLc and affect CHD is currently unknown. If they do (as shown in Figure 1 b) then the estimate of the LDLc effect on CHD is likely to be biased. In the MR example, selecting variants from large GWAS consortia where there is replication means that assumption (i) is likely to be correct. For assumption (ii) there is evidence that this is likely to be true. As with the RCT example, in MR we are often most worried about assumption (iii) being violated through directional (horizontal) pleiotropy—i.e. the LDLc genetic variants influencing other factors independently of LDLc which in turn (independently of LDLc) affect CHD ( Figure 1 d). If the IV assumptions are correct (as illustrated in Figure 1 a and c) it can be seen that the magnitude of effect of LDLc on CHD can be easily calculated by the following : e ffect of LDLc on odds of CHD = log odds CHD on Z ÷ β LDLc on Z, where Z = randomization to statins (in the RCT example) or genetic variants for LDLc (in the MR example). For example, if in a well-conducted RCT randomization to a standard dose of statins reduces LDLc by 4 mmol/l and CHD by a relative reduction of 20% (odds ratio 0.80), then the causal effect of LCLc on CHD is a relative reduction of 5% (OR 0.95) per 1 mmol/l. It can also be seen that if assumption (iii) (the exclusion restriction criteria) is violated (as illustrated in Figure 1 b), then this estimate is biased as it is the combined effect of LDLc and any other lipids or fatty acids that are independently affected by statins and influence CHD.
See this image and copyright information in PMC

Comment on

  • Mendelian randomization study of adiposity-related traits and risk of breast, ovarian, prostate, lung and colorectal cancer.
    Gao C, Patel CJ, Michailidou K, Peters U, Gong J, Schildkraut J, Schumacher FR, Zheng W, Boffetta P, Stucker I, Willett W, Gruber S, Easton DF, Hunter DJ, Sellers TA, Haiman C, Henderson BE, Hung RJ, Amos C, Pierce BL, Lindström S, Kraft P; the Colorectal Transdisciplinary Study (CORECT); Discovery, Biology and Risk of Inherited Variants in Breast Cancer (DRIVE); Elucidating Loci Involved in Prostate Cancer Susceptibility (ELLIPSE); Follow-up of Ovarian Cancer Genetic Association and Interaction Studies (FOCI); and Transdisciplinary Research in Cancer of the Lung (TRICL). Gao C, et al. Int J Epidemiol. 2016 Jun;45(3):896-908. doi: 10.1093/ije/dyw129. Epub 2016 Jul 17. Int J Epidemiol. 2016. PMID: 27427428 Free PMC article.

References

    1. Gao C, Patel CJ, Michailidou K, et al. . Mendelian randomization study of adiposity-related traits and risk of breast, ovarian, prostate, lung and colorectal cancer . Int J Epidemiol 2016. ; 45 : 896 – 908 . - PMC - PubMed
    1. Davey Smith G, Ebrahim S . ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol 2003. ; 32 : 1 – 22 . - PubMed
    1. Davey Smith G, Hemani G . Mendelian randomization: genetic anchors for causal inference in epidemiological studies . Hum Mol Genet 2014. ; 23 : R89 – 98 . - PMC - PubMed
    1. Burgess S, Scott RA, Timpson NJ, Davey Smith G, Thompson SG, ; Consortium E-I . Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors . Eur J Epidemiol 2015. ; 30 : 543 – 52 . - PMC - PubMed
    1. Davey Smith G . Randomised by (your) god: robust inference from an observational study design . J Epidemiol Community Health 2006. ; 60 : 382 – 88 . - PMC - PubMed

Publication types