Highlights

Volume 35, Issue 1 (2025)

Special Article

Identifying a sufficient set of confounders is crucial for estimating the causal effect of exposure on outcome.
We summarize approaches to identifying a sufficient set of confounders from both epidemiological and statistical perspectives.
We also describe some pitfalls and considerations in confounder selection.
E-value and robustness value are introduced as approaches to assess the sensitivity to unmeasured confounders.
Integrating these principles and approaches will enhance our understanding of confounder selection in future epidemiological studies.

A comprehensive side-by-side comparison of different instrumental variable (IV) methods in binary outcome and continuous exposures with many IVs was performed.
The findings indicated that the bias associated with these methods could not be fully addressed.
The use of multiple IV methods (e.g., one method for primary analysis and another for sensitivity analysis) may be helpful.

We employ disease-mediating mortality separating disease-specific mortality into disease (the mediator) and mortality (the outcome) to characterize the natural course of hepatitis C virus (HCV).
We identified three liver diseases (eg, liver cirrhosis, liver cancer) and nine non-liver diseases (eg, septicemia, renal diseases, endocrine diseases, hypertension) that mediate HCV-induced mortality and provide the proportion of mediation to reflect their relative contributions.
For developing countries with limited access to direct-acting antiviral agents, our study may guide health policies in screening for particular diseases (except liver diseases) to reduce the preventable deaths of the HCV-infected population.

The association between income and colorectal cancer (CRC) risk varies across countries.
The role of income dynamics in CRC risk is unknown in individuals with type 2 diabetes mellitus (T2DM) who are at increased risk of colorectal cancer.
Individuals with T2DM who had persistent low-income status over 5 years or experienced income declines had an elevated CRC risk.
In contrast, constant high income over 5 years was associated with reduced risk of colorectal cancer.

This study investigated the association between the patterns of various metabolites and cognitive decline in Japanese community-dwelling older adults.
Amino acids were protectively associated with cognitive decline, whereas ketone metabolites were associated with higher odds of cognitive decline.
Metabolome monitoring can be useful for predicting and preventing future cognitive decline.

Short Communication

We compared self-reported participation in cancer screenings and health checkups with municipality records.
Self-reported participation in colorectal, stomach, breast, and cervical cancer screenings was moderately valid.
Self-reported participation in lung cancer screening was underreported and participation in health checkup was overreported.
Self-reported screening participation should be carefully interpreted to assess the performance of preventive measures.