Non-linear Mendelian randomization: detection of biases using negative controls with a focus on BMI, Vitamin D and LDL cholesterol

Fergus W. Hamilton,David A. Hughes,Wes Spiller,Kate Tilling,George Davey Smith

DOI: https://doi.org/10.1007/s10654-024-01113-9

2024-05-26

European Journal of Epidemiology

Abstract:Mendelian randomisation (MR) is an established technique in epidemiological investigation, using the principle of random allocation of genetic variants at conception to estimate the causal linear effect of an exposure on an outcome. Extensions to this technique include non-linear approaches that allow for differential effects of the exposure on the outcome depending on the level of the exposure. A widely used non-linear method is the residual approach, which estimates the causal effect within different strata of the non-genetically predicted exposure (i.e. the "residual" exposure). These "local" causal estimates are then used to make inferences about non-linear effects. Recent work has identified that this method can lead to estimates that are seriously biased, and a new method—the doubly-ranked method—has been introduced as a possibly more robust approach. In this paper, we perform negative control outcome analyses in the MR context. These are analyses with outcomes onto which the exposure should have no predicted causal effect. Using both methods we find clearly biased estimates in certain situations. We additionally examined a situation for which there are robust randomised controlled trial estimates of effects—that of low-density lipoprotein cholesterol (LDL-C) reduction onto myocardial infarction, where randomised trials have provided strong evidence of the shape of the relationship. The doubly-ranked method did not identify the same shape as the trial data, and for LDL-C and other lipids they generated some highly implausible findings. Therefore, we suggest there should be extensive simulation and empirical methodological examination of performance of both methods for NLMR under different conditions before further use of these methods. In the interim, use of NLMR methods needs justification, and a number of sanity checks (such as analysis of negative and positive control outcomes, sensitivity analyses excluding removal of strata at the extremes of the distribution, examination of biological plausibility and triangulation of results) should be performed.

public, environmental & occupational health

What problem does this paper attempt to address?

The paper aims to address the bias issues present in nonlinear Mendelian Randomization (MR) methods and evaluate the performance of these methods under different circumstances. Specifically, the paper focuses on the residual approach and the doubly-ranked method, two nonlinear MR methods.

### Main Research Questions:
1. **Evaluate Two Nonlinear MR Methods**: Assess whether the residual approach and the doubly-ranked method exhibit bias in estimating causal effects through negative control outcome analysis.
2. **Detect Bias**: Use negative control outcomes (such as age, gender, etc.) where no causal relationship is known to exist, to detect whether these two methods produce non-zero estimates.
3. **Validate Effectiveness**: Validate the performance of these two methods by comparing the results of randomized controlled trials on the impact of lowering low-density lipoprotein cholesterol (LDL-C) on myocardial infarction.

### Specific Objectives:
- Conduct negative control outcome analysis using data from the UK Biobank.
- Detect the impact of exposure factors such as vitamin D, body mass index (BMI), LDL-C, and triglycerides (TG) on negative control outcomes.
- Compare the results obtained from nonlinear MR methods with those from traditional MR methods.
- Compare the performance of these two methods across different strata to evaluate their reliability.

### Key Findings:
- The estimates obtained from traditional MR methods are close to zero, while the two nonlinear MR methods produce significant differences across different strata.
- Both the residual approach and the doubly-ranked method exhibit noticeable bias in certain situations.
- The doubly-ranked method produces estimates closer to zero in some analyses but still shows bias in other cases.

### Conclusion:
The paper suggests that extensive simulation and empirical studies are needed before further application of nonlinear MR methods to ensure their reliability and accuracy under different conditions. Additionally, multiple checks, such as negative and positive control outcome analyses and sensitivity analyses, should be conducted when using nonlinear MR methods to ensure the reasonableness of the results.