Meta-analysis and Mendelian randomization: A review.

Area of Science:

• Epidemiology
• Genetic Epidemiology
• Biostatistics

Background:

• Mendelian randomization (MR) employs genetic variants as instrumental variables to investigate causal relationships between risk factors and health outcomes.
• Historically, meta-analysis combined MR results from studies with limited genetic variants; now it integrates genome-wide association study (GWAS) summary data for numerous variants.
• Heterogeneity in causal estimates from multiple genetic variants may indicate violations of instrumental variable assumptions.

Purpose of the Study:

• To provide a foundational introduction to Mendelian randomization (MR) and its underlying instrumental variable theory.
• To describe advanced statistical methods for assessing and managing heterogeneity in MR analyses.
• To enhance the reliability and rigor of causal inference in genetic epidemiology.

Main Methods:

• Review of instrumental variable assumptions crucial for Mendelian randomization.
• Description of meta-analysis techniques for combining results from multiple genetic variants.
• Explanation of random effects models, meta-regression, and robust regression for heterogeneity testing and adjustment.

Main Results:

• Heterogeneity among causal estimates can signal potential violations of MR assumptions.
• Advanced statistical methods can identify and correct for heterogeneity, improving causal inference.
• The application of these methods strengthens the validity of MR findings.

Conclusions:

• Mendelian randomization is a powerful tool for causal inference in health research.
• Addressing heterogeneity is critical for robust MR studies.
• Methods like meta-regression enhance the rigor of MR by accounting for data variability.