Depressive symptoms in the association between herpes zoster and heart disease in middle-aged and older adults: A longitudinal cohort study.
BACKGROUND Herpes zoster (HZ) has been linked to elevated cardiovascular risk, but the mechanisms remain unclear. Depressive symptoms, particularly specific domains of symptoms, may show statistical mediation in this association. This study examined the longitudinal association between HZ and heart disease, assessed whether depressive symptoms showed statistical mediation, and conducted exploratory internal prediction analyses using machine-learning survival models. METHODS Based on the data from the Health and Retirement Study (HRS), we used Cox proportional hazards models to estimate the association between HZ and heart disease and mediation analyses to evaluate the role of total, cognitive-affective, and somatic depressive symptoms in middle-aged and older adults. In addition, five survival prediction models, including Cox, RSF, GBM, LASSO-Cox, and XGBoost, were evaluated using the concordance index (C-index). RESULTS Among 16,195 participants, HZ was associated with an increased risk of heart disease (adjusted HR 1.22, 95% CI 1.09-1.38, P = 0.001). Significant statistical indirect effects were observed for total (P = 0.003) and somatic depressive symptoms (P = 0.001), whereas no significant indirect effect was observed for cognitive-affective symptoms. Among the evaluated models, Cox and LASSO-Cox showed the highest discrimination, with C-indices of approximately 0.67. CONCLUSION HZ was associated with an increased risk of incident heart disease. Statistical indirect effects were observed through depressive symptoms, particularly somatic symptoms, under the assumed temporal ordering. HZ and depressive symptoms provided modest additional predictive information in exploratory internal analyses.