Depressive Symptom Clusters Differentially Predict Cardiovascular Hospitalization in People With Type 2 Diabetes

BackgroundDepression has been associated with the development of cardiovascular disease in people with type 2 diabetes.ObjectiveWe examined whether symptoms related to the 2 core features of depression—dysphoria and anhedonia—and anxiety were differentially associated with cardiovascular hospitalization and whether there were symptom-specific mechanisms (alcohol, smoking, physical activity, body mass index, glucose, cholesterol, and blood pressure) in play.MethodA total of 1465 people in Dutch primary care completed the Edinburgh Depression Scale in 2005 and were followed up until first cardiovascular hospitalization during follow-up (event) or December 31, 2010 (study end). Cox regression analyses examined (1) differences in time to hospitalization for a cardiovascular event between people with a low vs a high baseline dysphoria/anhedonia/anxiety score (adjusting for demographic and clinical confounders) and (2) mediating mechanisms.ResultsA total of 191 people were hospitalized for a cardiovascular event. In univariable analysis, dysphoria predicted a shorter time to cardiovascular hospitalization (hazard ratio = 1.49, 95% CI: 1.02–2.17). After adjustment for confounders, neither dysphoria (hazard ratio = 1.55, 95% CI: 0.91–2.64) nor anhedonia (hazard ratio = 0.83, 95% CI: 0.47–1.48) was significantly associated with time to cardiovascular hospitalization. Anxiety was associated with a longer time to cardiovascular hospitalization (adjusted hazard ratio = 0.49, 95% CI: 0.27–0.89). However, none of the selected factors qualified as a mediator for the (adjusted) association between anxiety and time to cardiovascular hospitalization.DiscussionDysphoria was associated with a shorter time to cardiovascular hospitalization in unadjusted analyses only, whereas anxiety predicted later hospitalization after confounder adjustment. Anhedonia did not show a significant association. Mechanistic pathways remain unclear.