Psychosocial contributors to bone health are not well characterised despite evidence linking mental health and physical disease. This prospective analysis used data from the English Longitudinal Study of Ageing (ELSA) to test whether baseline depressive symptoms predict incident osteoporosis over approximately 20 years. The study also examined whether associations differ by baseline age or sex.
The analytic sample comprised 8,577 community-dwelling adults aged 50 years or older who were free of osteoporosis at baseline (2002–2003). Participants were drawn from ELSA, a large, longitudinal cohort study of older adults in England. Follow-up for incident osteoporosis extended through wave 11 (2023–2024). The authors report that 948 participants (11.1%) developed osteoporosis during follow-up.
Depressive symptoms were assessed at baseline using the 8-item Center for Epidemiologic Studies-Depression Scale (CES-D). The CES-D score was treated continuously in analyses; reported associations describe the change in hazard of incident osteoporosis per one-point increase in baseline CES-D.
The study outcome was incident osteoporosis reported during the follow-up period. Participants were followed from baseline (2002–2003) until wave 11 (2023–2024). The source document reports the number of incident osteoporosis cases (948; 11.1%) but does not provide additional outcome ascertainment details (for example, whether diagnoses were self-reported, clinician-confirmed, or based on bone mineral density testing); these specifics were not reported in the abstract.
Cox proportional hazards regression models were used to estimate associations between baseline CES-D scores and incident osteoporosis. Primary models adjusted for age, sex and wealth. The investigators assessed effect modification by age and sex and performed sequential adjustment guided by a directed acyclic graph for potential confounders including education, wealth, inflammatory conditions and health behaviours. Exact model specifications beyond the covariates named in the abstract and diagnostic checks for proportional hazards were not detailed in the abstract.
Over the 20-year follow-up period, each one-point higher baseline CES-D score was associated with an increased hazard of incident osteoporosis after adjustment for age, sex and wealth (hazard ratio [HR] 1.07, 95% CI 1.04–1.11, p<0.001). This indicates a graded association between depressive symptom burden at baseline and later osteoporosis risk in the overall sample.
The association between baseline depressive symptoms and subsequent osteoporosis was modified by baseline age (interaction p=0.008). In stratified analyses the association was concentrated among participants aged 50–64 years at baseline (HR 1.10, 95% CI 1.06–1.14). No significant association was observed in older baseline age strata. While osteoporosis incidence was higher in women overall, sex did not modify the relationship between depressive symptoms and incident osteoporosis according to the reported findings.
The authors report that the findings were robust to sequential adjustment informed by a directed acyclic graph. After additional adjustment for education, wealth, inflammatory conditions and health behaviours the association remained statistically significant (adjusted HR 1.04, 95% CI 1.01–1.08). The abstract does not provide the full list of health-behaviour variables, nor details on missing data handling, competing risks, or other sensitivity checks; those details were not reported in the abstract.
In this long-term prospective cohort, higher baseline depressive symptoms were associated with greater risk of incident osteoporosis over two decades, with the effect concentrated in adults who were in mid-life (50–64 years) at baseline. The authors suggest that mid-life may represent a window for psychosocial interventions aimed at supporting bone health. Because this report is a preprint, the findings should be interpreted cautiously until peer review and further replication are available.
Ethical approval for the English Longitudinal Study of Ageing was provided by the London Multicentre Research and Ethics Committee (MREC/01/02/91). The authors declare no competing interests. ELSA data are available through the UK Data Service. The article is a preprint and has not been certified by peer review, and the abstract does not include some methodological details such as exact outcome ascertainment methods and certain sensitivity analyses.