Yongsheng Tian, Chuan Lin, Tianyu Zhao, Rui La, Shasha Zhang, Xu Wu, Dinghua Jiang, Lixin Huang, Qian Wu, Jiangping Zeng, Shenghao Wang
Among older adults, lower femoral neck T-score was associated with higher all-cause mortality. The relation was nonlinear, with an estimated turning point at a less negative T-score in men than in women. Depressive symptoms contributed a modest but statistically significant fraction of this association in both sexes, supporting the value of including psychological status in the clinical assessment of osteoporosis.
Background Impaired bone health is recognized as a contributor to mortality in later life, yet the pathways connecting these conditions are incompletely understood. Existing work has emphasized fractures and functional decline, whereas the contribution of depressive symptoms, and whether that contribution differs between men and women, has received less attention. Methods Participants were 5823 adults aged 60 years or older from five NHANES cycles (2005–2018), including 2773 women and 3050 men. We evaluated femoral neck T-score as a categorical and continuous exposure using Cox models-incorporating strict proportional hazards verifications and corrections-and formally tested sex-exposure interactions. Survival probabilities were compared via Kaplan-Meier curves. We also examined departures from linearity with restricted cubic splines and threshold analyses, and tested whether depressive symptoms carried part of the exposure-mortality association in sex-specific mediation models. Finally, random forest imputation-based sensitivity analyses were conducted to ensure the robustness of our findings against missing covariates. Results After full covariate adjustment, higher femoral neck T-score was associated with lower all-cause mortality in women (HR 0.88, 95% CI 0.80–0.96) and men (HR 0.92, 95% CI 0.86–0.99). The osteoporosis category showed a larger point estimate in men (HR 1.73, 95% CI 1.34–2.22) than in women (HR 1.35, 95% CI 1.03–1.78), although this difference was numerical rather than confirmed by a significant interaction. Survival declined in a stepwise manner from normal bone mass through osteopenia to osteoporosis in both sexes. The relation was nonlinear, with an estimated inflection point at T = −2.903 in women and T = −1.810 in men. Depressive symptoms accounted for a small but statistically detectable part of the association in women (3.2%, 95% CI 0.3–14.6) and men (5.1%, 95% CI 0.4–22.7). Conclusions Among older adults, lower femoral neck T-score was associated with higher all-cause mortality. The relation was nonlinear, with an estimated turning point at a less negative T-score in men than in women. Depressive symptoms contributed a modest but statistically significant fraction of this association in both sexes, supporting the value of including psychological status in the clinical assessment of osteoporosis.