Caution against examining the role of reverse causality in Mendelian Randomization

Area of Science:

• Genetic Epidemiology
• Statistical Genetics
• Causal Inference

Background:

• Mendelian Randomization (MR) is increasingly used to investigate causal relationships between phenotypes in genetic association studies.
• The MR Steiger approach is an extension of MR designed to infer the causal direction between two phenotypes in prospective studies.

Purpose of the Study:

• To examine and quantify the ability of the MR Steiger approach to correctly determine the direction of causality between two phenotypes.
• To assess the impact of pleiotropy on the accuracy of the MR Steiger approach.
• To evaluate the MR Steiger method's application in a real-world dataset (COPDGene study) concerning smoking and lung function.

Main Methods:

• Conducted simulation studies to evaluate the MR Steiger approach's performance in determining causal direction.
• Applied the MR Steiger method to the COPDGene study data to investigate the relationship between smoking and lung function.
• Developed an R package ('reverseDirection') to simulate user-specified scenarios for assessing the MR Steiger approach's reliability.

Main Results:

• Simulation results indicate that the MR Steiger approach may fail to correctly identify the direction of causality, especially when pleiotropy is present.
• Application to the COPDGene study provided insights into the role of smoking in lung function using the MR Steiger method.

Conclusions:

• The MR Steiger approach requires cautious application for inferring causal direction due to potential inaccuracies, particularly in the presence of pleiotropy.
• The 'reverseDirection' R package offers a tool for researchers to test the MR Steiger method's validity under various conditions.