Temporal Association Between Diabetes and Depression and Their Impact on Cardiovascular Disease in Chinese Older Adults

ABSTRACT Previous studies on the direction of the association between diabetes and depression were inconsistent, leaving a knowledge gap in understanding the temporal sequence of the association and its impact on the risk of cardiovascular disease (CVD). This study aimed to examine the temporal relationship between diabetes and depression and their impact on the risk of CVD. We included 7124 participants aged ≥ 45 years with repeated measurement of glycated hemoglobin (HbA1c) and depression (measured by Center for Epidemiological Studies‐Depression scale, CES‐D) from the China Health and Retirement Longitudinal study. The cross‐lagged panel and mediation analyses were performed to assess the temporal relationship between HbA1c and depression, and their impact on CVD. The adjusted cross‐lagged path coefficient from baseline HbA1c to follow‐up CES‐D score ( β 1 , 0.181; 95% confidence interval [CI], 0.033–0.329; p = 0.0165) was significantly greater than the path coefficient from baseline CES‐D to follow‐up HbA1c ( β 2 , 0.001; 95% CI, −0.005 to 0.003; p = 0.6442) ( p difference = 0.0160), indicating a unidirectional relationship between HbA1c and CES‐D. Similar results were observed for the unidirectional relationship between baseline HbA1c and annual changes in CES‐D ( p difference = 0.0453). The annual changes in CES‐D during the follow‐up period significantly increased with higher quartiles of baseline HbA1c ( p = 0.005 for trend). Additionally, 14.45% of total association between HbA1c and CVD was mediated through follow‐up depression. Among Chinese middle‐aged and elder adults, diabetes was associated with a subsequent increase in depression, and depression partially mediated the pathways from diabetes to incident CVD.