Sensitivity Analyses for Robust Causal Inference from Mendelian Randomization Analyses with Multiple Genetic Variants.
Burgess S, Bowden J, Fall T, Ingelsson E, Thompson SG. · 2017
Read Sensitivity Analyses for Robust Causal Inference from Mendelian Randomization Analyses with Multiple Genetic Variants. by Burgess S, Bowden J, Fall T, Ingelsson E, Thompson SG. (2017) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc5133381.txt.
Introduction
Stephen Burgess, Jack Bowden, Tove Fall, Erik Ingelsson, and Simon G. Thompson set out robustness checks for Mendelian randomization in Epidemiology, with the hosted text dated 2017. The method uses gene variants as natural experiments to test whether a risk factor causes a disease, but it leans on strict assumptions about those variants.
When many variants are used at once, some may cheat the assumptions, for instance by affecting the outcome through a side path. The paper explains the logic of instrumental variables and develops sensitivity analyses that probe whether a causal claim survives when some instruments are faulty. Such checks, the authors argue, let researchers spot results driven by bad instruments and report findings that hold up under weaker assumptions.
Catalog record
| CREATOR | Burgess S, Bowden J, Fall T, Ingelsson E, Thompson SG. |
|---|---|
| YEAR | 2017 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Sensitivity Analyses for Robust Causal Inference from Mendelian Randomization Analyses with Multiple Genetic Variants By Stephen Burgess, Jack Bowden, Tove Fall, Erik Ingelsson, Simon G. Thompson Published in: Epidemiology (Cambridge, Mass.) DOI: 10.1097/EDE.0000000000000559 ABSTRACT Supplemental Digital Content is available in the text. An instrumental variable in an observational study behaves similarly to random treatment assignment in an experimental setting.1 It provides a natural experiment, whereby individuals with different levels of the instrumental variable differ on average with respect to the putative risk factor, but not with respect to any confounders of the risk factor–outcome association.2 Mendelian randomization is the use of a genetic variant as a proxy for a modifiable risk factor.3,4 If a genetic variant satisfies the assumptions of an instrumental variable for the risk factor, then whether there is an association between the genetic variant and the outcome is a test of whether the risk factor is a cause of the outcome.5 The instrumental variable assumptions are satisfied for a genetic variant if - (i) the genetic variant is associated with the risk factor; - (ii) the genetic variant is not associated with confounders of the risk factor–outcome relationship; and - (iii) the genetic variant is not associated with the outcome conditional on the risk factor and confounders of the risk factor–outcome relationship.6 These assumptions imply that the only causal pathway from the genetic variant to the outcome is via the risk factor, and there is no other causal pathway either directly to the outcome or via a confounder.7 A diagram corresponding to these assumptions is presented in Figure 1. [Figure: Diagram of instrumental variable assumptions for Mendelian randomization. The three assumptions (i, ii, iii) are illustrated by the presence of an arrow, indicating the effect of one variable on the other (assumption i), or by a dashed line with a cross, indicating that there is no direct effect of one variable on the other (assumptions ii and iii).] We further assume that all valid instrumental variables identify the same causal parameter; we return to this assumption in the discussion. For this interpretation to hold, it is necessary for certain parametric assumptions to hold. In this article, we assume that the effects of (i) the instrumental variables on the risk factor, (ii) the instrumental variables on the outcome, (iii) the risk factor on the outcome are linear without effect modification; and (iv) the association of the genetic variant with the risk factor is homogeneous in the population.5 These assumptions are not necessary for the identification of a causal effect, but they ensure that the estimate from each instrumental variable targets the same average causal effect.8 Weaker assumptions can identify a local average causal effect;9 however, the local average causal effect is likely to differ for each instrumental variable. Although these assumptions are strict, the causal estimate from an instrumental variable analysis is a valid test statistic for the causal null hypothesis without requiring the assumptions of linearity, homogeneity, or monotonicity.10 In any case, the causal effect of intervention on a risk factor is likely to depend on several aspects of the intervention (e.g., its magnitude, duration, and pathway), and therefore will not precisely correspond to the estimate from a Mendelian randomization analysis.11 Hence, we would urge practitioners to view the assessment of causality as the primary result of a Mendelian randomization, and not to interpret any causal estimate too literally.12 We also assume that the genetic variants are mutually independent in their distributions, although extensions are available for most of the analysis methods in the case of correlated variants, provided that the correlation structure is known.13 Genetic variants are particularly suitable candidate instrumental variables, as they are fixed at conception, and hence cannot be affected by environmental factors that could otherwise lead to confounding or reverse causation.14 However, there are many well-documented ways in which the instrumental variable assumptions may be violated for any particular genetic variant, such as pleiotropy, linkage disequilibrium, and population stratification.3,15 For risk factors that are soluble protein biomarkers, there is often a gene region that encodes the protein (for example, the CRP gene region for C-reactive protein16), or a regulator or inhibitor of the protein (e.g., the IL6R gene region for interleukin-617). Using one or more variants from such a gene region as instrumental variables would be ideal for a Mendelian randomization analysis, as these genetic variants would be the most likely to satisfy the instrumental variable assumptions, and the most informative proxies for intervention on the risk factor.18 However, such genetic variants do not exist for many risk factors. The approach of using multiple genetic variants in different gene regions is particularly suitable for complex risk factors that are multifactorial and polygenic, such as body mass index,19 height,20 or blood pressure.21 Summarized data (in particular, beta-coefficients and standard errors) on genetic associations with the risk factor can be combined with summarized data on genetic associations with the outcome (that are often publicly available for download) to provide causal effect estimates, under the assumption that the genetic variants are all instrumental variables.22,23 Using multiple genetic variants increases the power of a Mendelian randomization investigation compared with an analysis based on a single variant.24 However, even if only one of the genetic variants is not a valid instrumental variable, the causal estimate based on all the variants from a conventional Mendelian randomization analysis will be biased and type 1 (false positive) error rates will be inflated.25,26 In this article, we describe a range of sensitivity analyses that either support or question the validity of causal inference from a Mendelian randomization analysis with multiple genetic variants. These sensitivity analyses will be useful for judging whether a causal conclusion from such an analysis is plausible or not. We focus on those sensitivity analyses that can be implemented using summarized data only. We consider approaches under two broad categories: methods for assessing the instrumental variable assumptions, and robust analysis methods that rely on a less stringent set of assumptions than a conventional Mendelian randomization analysis. We illustrate the approaches using the example of estimating the causal effect of C-reactive protein (CRP) on coronary artery disease (CAD) risk using four genetic variants in the CRP gene region,16 and using 17 genetic variants (eTable A1; http://links.lww.com/EDE/B114) that have been shown to be associated with CRP at a genome-wide level of significance in a large meta-analysis—see eFigure in Ref. 27—beta-coefficients represent per allele associations with log-transformed CRP concentrations. Genetic associations with CAD risk were taken from the CARDIoGRAM consortium;28 beta-coefficients represent per allele log odds ratios for CAD risk. Ethical approval for the analyses using four genetic variants in the CRP gene region was granted by the Cambridgeshire ethics review committee; for the analyses using 17 genetic variants associated with CRP concentrations and with CAD risk, ethical approval was granted to the constituent studies by local institutional review boards. For reference, the causal estimate based on the genetic variants in the CRP gene region is null (odds ratio: 1.00, 95% confidence interval: 0.90, 1.13 per 1-SD increase in CRP concentrations [equal to a 1.05-unit increase in log-transformed CRP or a 2.86-fold increase]), whereas the “causal” estimate using an inverse-variance weighted method based on the genome-wide significant variants (a less reliable