Mendelian randomisation is an epidemiological method for estimating causal associations from observational data by using genetic variants as instrumental variables. Typically the genetic variants explain only a small proportion of the variation in the risk factor of interest, and so large sample sizes are required, necessitating data from multiple sources. Meta-analysis based on individual patient data requires synthesis of studies which differ in many aspects. A proposed Bayesian framework is able to estimate a causal effect from each study, and combine these using a hierarchical model. The method is illustrated for data on C-reactive protein and coronary heart disease (CHD) from the C-reactive protein CHD Genetics Collaboration (CCGC). Studies from the CCGC differ in terms of the genetic variants measured, the study design (prospective or retrospective, population-based or case-control), whether C-reactive protein was measured, the time of C-reactive protein measurement (pre- or post-disease), and whether full or tabular data were shared. We show how these data can be combined in an efficient way to give a single estimate of causal association based on the totality of the data available. Compared to a two-stage analysis, the Bayesian method is able to incorporate data on 23% additional participants and 51% more events, leading to a 23–26% gain in efficiency.
Davey SmithGEbrahimS. ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease?Int J Epidemiol2003; 32(1): 1–22.
2.
Davey SmithGEbrahimS. Mendelian randomization: prospects, potentials, and limitations. Int J Epidemiol2004; 33(1): 30–42.
3.
DidelezVSheehanN. Mendelian randomization as an instrumental variable approach to causal inference. Stat Meth Med Res2007; 16(4): 309–330.
4.
GreenlandS. An introduction to instrumental variables for epidemiologists. Int J Epidemiol2000; 29(4): 722–729.
5.
MaldonadoGGreenlandS. Estimating causal effects. Int J Epidemiol2002; 31(2): 422–429.
6.
TaubesGMannCC. Epidemiology faces its limits. Science1995; 269(5221): 164–169.
7.
KhawKTBinghamSWelchA. Relation between plasma ascorbic acid and mortality in men and women in EPIC-Norfolk prospective study: a prospective population study. Lancet2001; 357(9257): 657–663.
8.
Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of antioxidant vitamin supplementation in 20536 high-risk individuals: a randomised placebo-controlled trial. Lancet 2002; 360(9326): 23–33.
9.
AngristJDImbensGWRubinDB. Identification of causal effects using instrumental variables. J Am Stat Assoc1996; 91(434): 444–455.
10.
MartensEPPestmanWRde BoerA. Instrumental variables: application and limitations. Epidemiology2006; 17(3): 260–267.
11.
LawlorDHarbordRSterneJ. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat Med2008; 27(8): 1133–1163.
12.
Davey SmithGEbrahimS. What can Mendelian randomisation tell us about modifiable behavioural and environmental exposures?Br Med J2005; 330(7499): 1076–1079.
13.
KeavneyB. Commentary: Katan's remarkable foresight: genes and causality 18 years on. Int J Epidemiol2004; 33(1): 11.
14.
HernánMARobinsJM. Instruments for causal inference: an epidemiologist's dream?Epidemiology2006; 17(4): 360–372.
15.
WehbyGOhsfeldtRMurrayJ. “Mendelian randomization” equals instrumental variable analysis with genetic instruments. Stat Med2008; 27(15): 2745–2749.
16.
LawlorDWindmeijerFDavey SmithG. Is Mendelian randomization “lost in translation”: Comments on “Mendelian randomization equals instrumental variable analysis with genetic instruments” by Wehby et al. Stat Med2008; 27(15): 2750–2755.
17.
WheatleyKGrayR. Commentary: Mendelian randomization-an update on its use to evaluate allogeneic stem cell transplantation in leukaemia. Int J Epidemiol2004; 33(1): 15–17.
18.
SleimanPGrantSFA. Mendelian randomization in the era of genomewide association studies. Clin Chem2010; 56(5): 723–728.
19.
CRP CHD Genetics Collaboration. Association between C reactive protein and coronary heart disease: Mendelian randomisation analysis based on individual participant data. Br Med J 2011; 342: d548.
20.
CRP CHD Genetics Collaboration. Collaborative pooled analysis of data on C-reactive protein gene variants and coronary disease: judging causality by Mendelian randomisation. Eur J Epidemiol 2008; 23(8): 531–540.
21.
BochudMChioleroAElstonRC. A cautionary note on the use of Mendelian randomization to infer causation in observational epidemiology. Int J Epidemiol2008; 37(2): 414–416.
22.
Davey SmithGLawlorDHarbordR. Clustered environments and randomized genes: a fundamental distinction between conventional and genetic epidemiology. PLoS Med2007; 4(12): e352.
23.
RidkerPMPaynterNPDanikJS. Interpretation of Mendelian randomization studies and the search for causal pathways in atherothrombosis: the need for caution. Metab Syndr Relat Disord2010; 8(6): 465–469.
24.
BurgessSThompsonSG. CRP CHD Genetics Collaboration.Bayesian methods for meta-analysis of causal relationships estimated using genetic instrumental variables. Stat Med2010; 29(12): 1298–1311.
25.
Emerging Risk Factors Collaboration. C-reactive protein concentration and risk of coronary heart disease, stroke, and mortality: an individual participant meta-analysis. Lancet 2010; 375(9709): 132–140.
26.
CasasJShahTCooperJ. Insight into the nature of the CRP-coronary event association using Mendelian randomization. Int J Epidemiol2006; 35(4): 922–931.
27.
CasasJShahTHingoraniA. C-reactive protein and coronary heart disease: a critical review. J Intern Med2008; 264(4): 295–314.
28.
DaneshJPepysMB. C-reactive protein and coronary disease: is there a causal link?Circulation2009; 120(21): 2036–2039.
29.
ThomasDCContiDV. Commentary: the concept of ‘Mendelian Randomization’. Int J Epidemiol2004; 33(1): 21–25.
30.
BowdenJVansteelandtS. Mendelian randomisation analysis of case-control data using structural mean models. Stat Med2011; 30(6): 678–694.
31.
JohnsonADHandsakerREPulitSL. SNAP: a web-based tool for identification and annotation of proxy SNPs using HapMap. Bioinformatics2008; 24(24): 2938–2939.
32.
BoundJJaegerDBakerR. Problems with instrumental variables estimation when the correlation between the instruments and the endogenous explanatory variable is weak. J Am Stat Assoc1995; 90(430): 443–450.
33.
StaigerDStockJ. Instrumental variables regression with weak instruments. Econometrica1997; 65(3): 557–586.
34.
BurgessSThompsonSG. Bias in causal estimates from Mendelian randomization studies with weak instruments. Stat Med2011; 30(11): 1312–1323.
35.
BurgessSThompsonSGCRP CHD Genetics Collaboration. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol2011; 40(3): 755–764.
36.
PalmerTMLawlorDAHarbordRM. Using multiple genetic variants as instrumental variables for modifiable risk factors. Stat Meth Med Res2012.
37.
BaumCSchafferMStillmanS. Instrumental variables and GMM: Estimation and testing. Stata Journal2003; 3(1): 1–31.
38.
RassenJASchneeweissSGlynnRJ. Instrumental variable analysis for estimation of treatment effects with dichotomous outcomes. Am J Epidemiol2009; 169(3): 273–284.
39.
FosterE. Instrumental variables for logistic regression: an illustration. Soc Sci Res1997; 26(4): 487–504.
40.
TerzaJVBasuARathouzPJ. Two-stage residual inclusion estimation: addressing endogeneity in health econometric modeling. J Health Econom2008; 27(3): 531–543.
41.
CaiBSmallDSTen HaveTR. Two-stage instrumental variable methods for estimating the causal odds ratio: Analysis of bias. Stat Med2011; 30(15): 1809–1824.
42.
DidelezVMengSSheehanNA. Assumptions of IV methods for observational epidemiology. Stat Sci2010; 25(1): 22–40.
43.
PalmerTMSterneJACHarbordRM. Instrumental variable estimation of causal risk ratios and causal odds ratios in Mendelian randomization analyses. Am J Epidemiol2011; 173(12): 1392–1403.
44.
StockJYogoM. Testing for weak instruments in linear IV regression [Working Paper Series]. SSRNeLibrary2002; 11: T0284.
45.
AngristJDPischkeJS. Instrumental variables in action: sometimes you get what you need. In: Mostly harmless econometrics: an empiricist's companion, Princeton, NJ: Princeton University Press, 2009, pp. 113–220.
46.
WooldridgeJM. Econometric analysis of cross section and panel data, Massachusetts: The MIT Press, 2002.
47.
GreenlandS. Interpretation and choice of effect measures in epidemiologic analyses. Am J Epidemiol1987; 125(5): 761–768.
48.
BurgessSThompsonSG. Improvement of bias and coverage in instrumental variable analysis with weak instruments for continuous and binary outcomes. Stat Med2012.
49.
BurgessS. Collapsibility for IV analyses of binary outcomes. In: Statistical issues in Mendelian randomization: use of genetic instrumental variables for assessing causal associations, Cambridge: University of Cambridge, 2012, pp. 73–98.
50.
NagelkerkeNFidlerVBernsenR. Estimating treatment effects in randomized clinical trials in the presence of non-compliance. Stat Med2000; 19(14): 1849–1864.
51.
PalmerTThompsonJTobinM. Adjusting for bias and unmeasured confounding in Mendelian randomization studies with binary responses. Int J Epidemiol2008; 37(5): 1161–1168.
52.
DerSimonianRLairdN. Meta-analysis in clinical trials. Control Clin Trial1986; 7(3): 177–188.
53.
FordINorrieJAhmadiS. Model inconsistency, illustrated by the Cox proportional hazards model. Stat Med1995; 14(8): 735–746.
54.
SpiegelhalterDThomasABestN. WinBUGS version 1.4 user manual, Cambridge: MRC Biostatistics Unit, 2003.
55.
GelmanARubinDB. Inference from iterative simulation using multiple sequences. Stat Sci1992; 7(4): 457–472.
56.
BreslowNDayN. Conditional logistic regression for matched sets. In: The analysis of case-control studies, Lyon: International Agency for Research on Cancer, 1980, pp. 248–279.
57.
WhittemoreAS. Collapsibility of multidimensional contingency tables. J Roy Stat Soc: Series B (Methodological)1978; 40(3): 328–340.
58.
DavidsonRMacKinnonJG. Simultaneous equation models. In: Estimation and inference in econometrics, New York: Oxford University Press, 1993, pp. 622–668.
59.
FrostCThompsonSG. Correcting for regression dilution bias: comparison of methods for a single predictor variable. J Roy Stat Soc: Ser A2000; 163(2): 173–189.
60.
SuttonAJAbramsKR. Bayesian methods in meta-analysis and evidence synthesis. Stat Meth Med Res2001; 10(4): 277–303.
61.
AngristJDKruegerAB. The effect of age at school entry on educational attainment: an application of instrumental variables with moments from two samples. J Am Stat Assoc1992; 87(418): 328–336.
62.
ThompsonJMinelliCAbramsK. Meta-analysis of genetic studies using Mendelian randomization - a multivariate approach. Stat Med2005; 24(14): 2241–2254.
63.
HigginsJThompsonSGSpiegelhalterDJ. A re-evaluation of random-effects meta-analysis. J Roy Stat Soc: Ser A2009; 172(1): 137–159.
64.
VansteelandtSBowdenJBabanezhadM. On instrumental variables estimation of causal odds ratios. Stat Sci2011; 26(3): 403–422.
65.
GreenlandSRobinsJPearlJ. Confounding and collapsibility in causal inference. Stat Sci1999; 14(1): 29–46.
66.
MinelliCThompsonJTobinM. An integrated approach to the meta-analysis of genetic association studies using Mendelian randomization. Am J Epidemiol2004; 160(5): 445–452.
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
