Depression has been associated with the development of cardiovascular disease in people with type 2 diabetes. We examined whether symptoms related to the two core features of depression –dysphoria and anhedonia– and anxiety were differentially associated with cardiovascular hospitalisation and whether there were symptom-specific behavioural or pathophysiological mechanisms in play. A total of 1,465 people completed the Edinburgh Depression Scale (including subscales measuring dysphoria, anhedonia and anxiety) in 2005 and were followed until first cardiovascular hospitalisation during follow-up, death, or the end of the study period (December 31, 2010). Cox regression analyses were used to determine whether there was a difference in time to hospitalisation for a cardiovascular event between people with a low versus a high dysphoria/anhedonia/anxiety score at baseline (adjusting for demographic and clinical confounders) and to identify mediating mechanisms. At the end of follow-up, 191 people had experienced a hospitalisation for a cardiovascular event. There was no difference in time to first cardiovascular hospitalisation during follow-up between people with and without anhedonia or between people with and without anxiety, but individuals reporting high dysphoria did have a shorter mean time to event than those without high dysphoria (1,758 [95% CI 1,687–1,830] vs. 1,831 [95% CI 1,809–1,853] days). After adjustment for confounding factors, neither anhedonia (HR=0.83, 95% CI 0.47–1.48, p=0.53) nor dysphoria (HR=1.55, 95% CI 0.91–2.64, p=0.11) were significantly associated with time to cardiovascular hospitalisation. Anxiety was associated with a longer time to cardiovascular hospitalisation (HR=0.49, 95% CI 0.27–0.89, p=0.02). However, none of the selected pathophysiological and behavioural factors qualified as a mediator for the (adjusted) association between anxiety and time to cardiovascular hospitalisation. In conclusion, symptom-clusters of negative emotions predicted time to first hospitalisation for a cardiovascular event during follow-up, while symptoms of anhedonia did not. Mechanistic pathways should be explored further.