High-dimensional mediation analysis for longitudinal mediators and survival outcomes

Area of Science:

• Biostatistics
• Epidemiology
• Genomics

Background:

• Identifying epigenetic pathways linking environmental exposures to health outcomes is vital.
• Existing high-dimensional mediation analysis methods are underdeveloped for longitudinal mediators and survival outcomes.

Purpose of the Study:

• To develop and validate a novel method for high-dimensional mediation analysis with time-varying longitudinal mediators and a survival outcome.
• To capture and assess mediation effects over time in complex biological systems.

Main Methods:

• Utilized a longitudinal mixed effects model to analyze the exposure-mediator relationship.
• Employed a Cox proportional hazards model with time-varying mediators to link the mediating process to survival outcomes.
• Implemented mediation-based sure independence screening and Lasso inference for dimension reduction and identification of significant mediators, alongside a joint significance test for hypothesis testing.

Main Results:

• The proposed method effectively handles high-dimensional, longitudinal, time-varying mediators in mediation analysis.
• Demonstrated utility and validity through simulation studies and application to the Coronary Artery Risk Development in Young Adults Study.

Conclusions:

• The developed method advances high-dimensional mediation analysis for longitudinal data, offering a robust approach to investigate complex exposure-health outcome relationships.
• This work provides a valuable tool for researchers studying epigenetic mechanisms and their impact on survival.