A novel Mendelian randomization method with binary risk factor and outcome.

Area of Science:

• Epidemiology
• Genetic Epidemiology
• Statistical Genetics

Background:

• Mendelian randomization (MR) utilizes genetic variants as instrumental variables (IVs) in observational studies.
• While MR performance with binary outcomes is studied, few investigations address binary risk factors alongside binary outcomes.

Purpose of the Study:

• Develop a novel MR estimator for binary risk factors and outcomes.
• Compare the new estimator's performance against existing MR methods through simulations and real-world data.

Main Methods:

• Adapted a bivariate Bernoulli distribution for the IV setting.
• Evaluated empirical bias and asymptotic coverage probabilities using simulations.
• Compared the proposed method with Wald, 2SPS, 2SRI, and GMM estimators.
• Applied the method to CLEAR study data to assess smoking's causal effect on rheumatoid arthritis risk.

Main Results:

• The proposed MR method demonstrated low bias, comparable to 2SPS.
• Adequate coverage probabilities were observed for the proposed method, 2SPS, and 2SRI.
• Wald and GMM methods exhibited poor coverage in several simulation scenarios.
• No statistically significant causal effect of smoking on rheumatoid arthritis risk was found.

Conclusions:

• The developed MR method is robust for binary risk factors and outcomes, performing similarly to 2SPS and 2SRI regarding bias.
• The proposed method provides a more suitable framework for hypothesis testing than 2SPS or 2SRI, which require adjustments.