# Association of severe depression with bone mineral content: a mediation analysis in PERSIAN cohort study

## TL;DR

This study explores how severe depression is linked to lower bone mineral content, using mediators like sleep and physical activity in a large Iranian cohort.

## Contribution

The study introduces a mediation analysis framework to uncover direct and indirect effects of severe depression on bone health through lifestyle and sleep factors.

## Key findings

- Severe depression shows significant direct and indirect associations with reduced bone mineral content.
- Falling asleep duration and physical activity (METs) act as significant mediators in the depression-BMC relationship.
- Confounding variables strongly influence the observed associations, with many effects becoming non-significant after adjustment.

## Abstract

Depression may lead to lower Bone Mineral Content (BMC), as an indicator of bone health. In this study, we aim to estimate the direct, indirect, and total associations of depression with BMC through various mediators to explain the mechanism behind the reduction in BMC.

This is a cross-sectional study within the PERSIAN perspective cohort study at Mashhad University of Medical Sciences (MUMS) in northeastern Iran, involving 3179 personnel in 2019–2020. Depression was assessed using the standard DASS-21 questionnaire. Smoking, taking sleeping pills, falling asleep duration, and metabolic equivalents (METs) served as mediators, based on the literature review. We employed mediation analysis packages, with a counterfactual approach, to estimate the various associations between depression, the mediators, and BMC using Stata.

The mean age of participants with moderate and severe depression levels was 43.26 and 42.59 years, respectively. Most of the participants with depression were women. In multivariable analysis, the reference interaction (coef = 0.012), mediated interaction (coef = 0.004), and controlled direct effect (coef = -0.024) of severe depression through the falling asleep duration, as well as the pure indirect effect of severe depression through METs (coef = 0.001), remained significant (p < 0.05).

Severe depression level demonstrates both direct and indirect associations with BMC, along with interactions with smoking and the falling asleep duration as mediators. All the mediators exhibit a pure direct association with BMC. Nevertheless, it appears that our confounding variables exert a strong influence on these relationships, as most of the associations were not significant after adjustment.

The online version contains supplementary material available at 10.1186/s12889-026-26261-3.