Letter to the Editor “Association between circulating selenium levels and arterial stiffness: a nationwide cross-sectional study”

Dear Editor,

We read with great interest the article by Zhao et al entitled “Association between circulating selenium levels and arterial stiffness: a nationwide cross-sectional study” published in the International Journal of Surgery^[1]. This letter is compliant with the TITAN Guidelines 2025^[2]. The authors conducted a comprehensive analysis of a nationally representative sample from the National Health and Nutrition Examination Survey, providing valuable insights into the complex relationship between circulating selenium (CSe) and markers of arterial stiffness. Their use of multiple statistical models, including dose–response and mediation analyses, represents a robust framework for clarifying these associations. The identification of a linear relationship with the atherogenic index of plasma (AIP) and a U-shaped relationship with estimated pulse wave velocity is noteworthy and contributes significantly to the field.

While we acknowledge the authors for their comprehensive work, we aim to highlight a methodological consideration regarding the framing of sex as a mediator in the relationship between CSe and arterial stiffness. The mediation analysis reported an indirect effect through sex for high AIP. We contend that this conceptualization may be problematic from a causal inference perspective.

Mediation analysis fundamentally aims to trace a causal pathway where an exposure (CSe) influences an outcome (arterial stiffness) through a mediator. A valid mediator should be a variable on the causal pathway that is potentially modifiable and occurs after the exposure. Biological sex, however, is a fixed, non-modifiable characteristic that may precede and confound the exposure–outcome relationship rather than acting as a mechanism “through” which selenium exerts its effect. It might be more accurately conceptualized as an “effect modifier” or a “baseline confounder,” not a mediating mechanism.

This view is supported by the authors’ own findings, wherein the relationship between CSe and high AIP appears to differ by sex, though the current subgroup comparison cannot substantiate a true interaction effect. This finding implied that sex may act as an “effect modifier,” meaning the relationship between CSe and AIP may differ between males and females. Presenting it additionally as a mediator risks conflating two distinct statistical concepts (modification and mediation) and may lead to an overinterpretation of the results.

To clarify the authors’ methodological rationale and strengthen the interpretation of their findings, we would be grateful if they could address the following points:

1. What was the theoretical or conceptual basis for positioning sex as a mediator within the causal pathway?

2. Was the alternative approach of modeling sex exclusively as an effect modifier by incorporating interaction terms into the regression models considered to complement or validate the mediation analysis?

We believe that addressing these points will enhance the methodological rigor of an otherwise excellent study and provide a more precise interpretation of the complex role that sex plays in the relationship between selenium and cardiovascular health.

Thank you for considering our comments.

Ethical approval

Not applicable.

Consent

Not applicable.

Sources of funding

None.

Author contributions

Ethan M. Ting writes the whole manuscript.

Conflicts of interest disclosure

None.

Research registration unique identifying number (UIN)

Not applicable.

Guarantor

Ethan M. Ting.

Provenance and peer review

Not commissioned.

Data availability statement

Not applicable.

Acknowledgements

Not applicable.