What is Mendelian Randomization, and how is it used to infer causality in epidemiology?

Question

Studies of human traits and diseases are inherently confined to observational studies, known as epidemiological studies. This can make it very hard to determine what actually causes a particular disease or trait, and what is just a by-product or biomarker of the disease. **Mendelian randomization** is a statistical technique used by genetic epidemiologists to determine *causal* effects, for instance whether a biomarker actually influences disease risk, or whether it is just a statistical association. From basics, how can genetics be used to determine causal associations between traits (such as blood markers) and increased risk of disease, and on what assumptions is this based?

Answer 1

Mendelian Randomization (MR) is an analytical approach used in epidemiology to infer causality between an exposure (like a risk factor or behavioral trait) and an outcome (such as a disease) by utilizing genetic variants as instrumental variables. The concept is based on Mendel’s laws of inheritance and leverages genetic information to overcome confounding and reverse causation that often complicate observational studies.

Key Features of Mendelian Randomization

1. Genetic Variants as Instruments:
2. MR typically employs single nucleotide polymorphisms (SNPs) as instrumental variables. These genetic markers are associated with the exposure of interest (e.g., cholesterol levels, smoking status).
3. The use of inherited genetic variants helps to assume that they are randomly distributed in the population, akin to random allocation in a randomized controlled trial.
4. Three Core Assumptions of MR:
5. Relevance: The genetic variant is associated with the exposure (the risk factor).
6. Independence: The genetic variant is not associated with any confounding factors that affect both the exposure and outcome.
7. Exclusion Restriction: The genetic variant affects the outcome only through the exposure and not through any other pathways.

How Mendelian Randomization is Used to Infer Causality

1. Hypothesis Testing:
2. Researchers formulate a hypothesis about the causal relationship between an exposure and an outcome (e.g., whether increased cholesterol causes cardiovascular disease).
3. Identifying Genetic Instruments:
4. Large genome-wide association studies (GWAS) identify SNPs that are significantly associated with the exposure.
5. Data Analysis:
6. MR uses these SNPs to perform statistical analyses that assess whether variations in the genetic instruments lead to differences in the outcome.
7. Techniques like two-sample MR can be employed, where data from two different studies (one for the exposure and one for the outcome) are used to determine the effect.
8. Causal Inference:
9. If the genetic variants associated with the exposure are also associated with changes in the outcome, this supports a causal relationship. For example, if genotypes linked to higher LDL cholesterol are associated with increased risk of heart disease, it suggests that LDL cholesterol plays a role in the etiology of heart disease.
10. Assessing Potential Issues:
11. MR can also help to identify biases such as horizontal pleiotropy, where a genetic variant affects the outcome through pathways unrelated to the exposure. Researchers need to use sensitivity analyses to check for such biases.

Advantages of Mendelian Randomization

1. Minimizing Confounding: Since genetic variants are assigned at conception, they are not influenced by environmental factors or lifestyle choices that may confound observational relationships.
2. Temporal Clarity: MR can help establish a temporal relationship, reducing the risk of reverse causation.
3. Cost-Effectiveness: Utilizing existing genetic data can be less costly and faster than conducting comprehensive long-term studies or trials.