Table 2—

Comparative prospective studies of type 2 diabetes and incident depression

Author (year)Follow-upIncident n/Total nSample sourceSample composition*Depression assessmentDiabetes assessmentSelected estimate (95% CI)Statistical adjustment of selected estimate
Palinkas et al. (28)8 years118/971Population-basedAge: ≥50 yearsBDI75-g oral glucose tolerance test, FPG, non-FPG0.73 (0.41–1.30)Age, sex, physical activity, and BMI
57% F
100% W
Polsky et al. (37)6 years571/8,387Population-basedAge: 51–61 yearsCES-DSelf-report1.17 (0.98–1.41)Age, sex, race, marital status, education, wealth, income, self-rated health, disability, baseline CES-D, and chronic conditions
52% F
87% W
Brown et al. (38)12 years2,534/88,776Population-based health care registryAge: ≥20 years1) Rx for antidepressant and 2) depression-related claim (ICD-9 code 296, 309, or 311) within ±3 months of Rx1) Two or more diabetes-related claims within a 2-year period (ICD-9 code 250), or 2) diabetes-related hospitalization, or 3) hypoglycemic prescription1.04 (0.94–1.15)Age, sex, number of MD visits, comorbid arthritis, cancer and vascular disease, and insulin use
53% F
de Jonge et al. (30)5 years231/4,803Population-basedAge: ≥55 yearsGMS-AGECATSelf-report1.42 (1.04–1.93)Age and sex
58% F
Kim et al. (29)2 years63/521Population-basedAge: ≥65 yearsGMS-AGECATSelf-report1.0 (0.4–2.5)Unadjusted
55% F
100% A
Engum (25)10 years2,303/7,291Population-basedAge: ≥30 yearsHospital Anxiety and Depression Scale–Depression subscaleSelf-report, confirmed with FPG1.24 (0.78–1.98)Age, sex, education, and marital status
55% F
Maraldi et al. (39)5.9 years594/2,522Population-basedAge: 70–79 yearsCES-D or self-reported antidepressant useSelf-report diagnosis/ hypoglycemic medication use or FPG1.31 (1.07–1.61)Adjusted for age, sex, race, study site, and baseline CES-D
48% F
56% W
  • *

    * Racial/ethnic composition not provided for some studies: A, Asian; B, African American; F, female; W, non-Hispanic white. BDI, Beck Depression Inventory; CES-D, Centers for Epidemiologic Studies Depression Scale; FPG, fasting plasma glucose; GMS-AGECAT, Geriatric Mental State B3 diagnostic schedule with application of the Automated Geriatric Examination for Computer-Assisted Taxonomy algorithm.

  • Selected estimate refers to estimate that is most closely adjusted for only demographic characteristics (age, sex, race/ethnicity, socioeconomic indicators, and marital status) and is the estimate used in the pooled analyses and is depicted in Figure 1.