Depressive symptoms after incident cardiovascular disease are largely explained by pre-existing symptoms in middle-aged and older adults

Background: Depressive symptoms are common following cardiovascular disease (CVD); however, causality and mechanisms underlying this relationship remain unclear. Loneliness and social isolation are risk factors for both CVD and depressive symptoms, yet their roles as mediators are poorly understood. This study investigated whether loneliness or social isolation mediate any causal effect of incident CVD on subsequent depressive symptoms.Methods: We analysed data on adults aged ≥50 years from the English Longitudinal Study of Ageing. We measured incident CVD at wave 7 (2014-2015; exposure), loneliness and social isolation at wave 8 (2016-2017; mediators), and depressive symptoms at wave 9 (2018-2019; outcome). We employed causal mediation analysis to estimate total effects (TE) and natural indirect effects (NIE) in two separate single- mediator models adjusting for depressive symptoms at wave 6 (2012-13).Results: There were 3,727 participants in the loneliness mediator model and 2,506 in the social isolation model. In both, CVD did not have a significant total effect on subsequent depressive symptoms after full adjustment (loneliness: TE = 0.18 [-0.05, 0.41]; social isolation: TE = 0.25 [-0.01, 0.52]). While NIE estimates were also non- significant in both mediator models, loneliness was estimated to significantly predict subsequent depressive symptoms (conceptual path b = 0.21 [0.17, 0.25]).Conclusions: Adjustment for pre-existing depressive symptoms substantially attenuated the observed effects, suggesting that depressive symptoms before incident CVD explain much of the total and indirect causal effects. However, loneliness independently predicted subsequent depressive symptoms, regardless of incident CVD, highlighting its relevance as a possible modifiable target for psychological intervention.