Association of Depressive and Anxiety Disorders With Diagnosed Versus Undiagnosed Diabetes: An Epidemiological Study of 90,686 Participants

Objective To compare the odds of depressive and anxiety disorders for participants with diagnosed diabetes, participants with diabetes but unaware of this, and participants without diabetes. Such knowledge might improve etiological insight into psychopathology in diabetes. Methods Data of 90,686 participants (mean age = 45 years; 59% female) from the LifeLines cohort was used. Depressive and anxiety disorders were assessed by the Mini-International Neuropsychiatric Interview. The odds of depression and anxiety were assessed for three groups: a) diagnosed diabetes, diabetes medication use and/or self-reported diabetes; b) undiagnosed diabetes, fasting blood glucose 7.0 mmol/l, but no diabetes medication use and self-reported no diabetes; and c) no diabetes, fasting blood glucose <7.0 mmol/l and self-reported no diabetes. Logistic regression was performed to compare the odds of depression and anxiety in these groups, adjusting for age, sex, diabetes-related diseases, comorbid depressive or anxiety disorders, and glycosylated hemoglobin. Results A total of 3002 (3.3%) participants were diagnosed as having depression and 9018 (9.9%) as having anxiety; 1781 (2.0%) had diagnosed and 786 (0.9%) had undiagnosed diabetes. Both diagnosed (odds ratio [OR] = 1.4:1.1-1.8, p = .006) and undiagnosed (OR = 1.8:1.3-2.6, p = .001) diabetes were independently associated with depression. The odds of depression did not differ between diagnosed and undiagnosed diabetes (OR = 0.7, p = .17). Diagnosed diabetes was independently associated with anxiety (OR = 1.4:1.2-1.7, p < .001), but undiagnosed diabetes was not (OR = 0.8:0.6-1.1, p = .20). The odds of anxiety were significantly higher in diagnosed compared with undiagnosed diabetes (1.68:1.23-2.31, p = .001). Conclusions Depression was more prevalent in participants with diagnosed and undiagnosed diabetes, whereas anxiety was more prevalent only in participants who were aware of their diabetes. Longitudinal research is needed to assess the causal pathways of these associations.