Table 1

List and features of the prospective cohort studies included in quantitative analyses

Source [reference]Cohort and features (nation)Follow-up (years)Age at inclusion (years)Age at diagnosis of PD (years)Risk of PD (95% CI)Adjustment variablesFindingsPotential estimation bias
Hu et al. 2007 [6]51,552 subjects without PD, and type 1 diabetes (Finland)18.0 [mean]25–74 [range]Women: 65.8 Men: 64.3 [mean]1.83 (1.21–2.76)Age, sex, smoking, BMI, alcohol use, coffee and tea consumption, education, physical activity, systolic blood pressure, and total cholesterol (baseline)Type 2 diabetes was associated with an increased risk of PD. In sensitivity analysis (exclusion of those who had vascular diseases at baseline or who developed stroke during follow-up [N = 47,353]) risk of PD was 1.94 [1.21–3.11].Only baseline self-reported diabetes and confounders were included in risk analyses. Surveillance bias might account for higher rates in diabetes. Finally, as a result of case ascertainment procedure, it could not be excluded that few cases (mild untreated PD patients) were lost to identification.
Simon et al. 2007 [5]171,879 subjects without prevalent stroke and PD at baseline (121,046 women [Nurses' Health Study] and 50,833 men [Health Professionals Follow-up Study]); participants developing stroke before PD onset were censored throughout the follow-up (U.S.)Women: 22.9 Men: 12.6 [mean]Women: 30–55Men: 40–75 [range]Women: 63.5 Men: 69.7 [mean]1.04 (0.74–1.46)Age, sex, and smoking. The adjustment for multiple updated (every 2 years) covariates [BMI, physical activity, alcohol, caffeine and energy intake, and comorbidities] produced similar results (data not shown)Preceding diabetes (both type 1 and type 2) was not related to increased risk of PD. The association of baseline diabetes was also nonsignificant (RR = 1.12 [0.69–1.81]).Data on diabetes were self-reported.
Driver et al. 2008 [12]21,841 male subjects free of cancer, vascular disease, dementia, and PD enrolled in the Physicians' Health Study (U.S.)23.1 [median]40–84 [range]73.1 [median]1.34 (1.01–1.77)Age, smoking, alcohol use, BMI, physical activity, hypertension, and high serum cholesterol (updated yearly)Updated history type 2 diabetes was associated with increased risk of PD. Sensitivity analyses showed that PD was more associated with short duration and uncomplicated diabetes, and low BMI. In sensitivity analysis (exclusion of those developing vascular disease during the follow-up [N = 16,423]) risk of PD was 1.46 [1.00–2.13].Data on diabetes and PD were self-reported. The increased risk for those with shorter duration of diabetes could be explained by detection bias from increased medical surveillance.
Xu et al. 2011 [13]288,662 subjects without prevalent PD enrolled in the National Institutes of Health-AARP Diet and Health Study (U.S.)15 [mean]50–71 [range]66.7 [7.3] (mean [SD])1.41 (1.20–1.66)Age, sex, race, BMI, physical activity, smoking, coffee intake, and educationPreceding diabetes (both type 1 and type 2) was associated with an increased risk of PD. In sensitivity analysis (exclusion of those with stroke, heart disease, cancer and poor/fair health [N = 215,723]) risk of PD was 1.34 [1.06–1.69].Data on diabetes were self-reported. Only baseline diabetes and confounders were included in risk analyses. The increased risk of PD could be partly explained by detection bias from increased medical surveillance in diabetic participants.