Type 2 diabetes mellitus and risk of dementia: A population-based 14-year follow-up study in Germany

Abstract

Introduction In the face of population ageing, dementia is an urgent public health problem. Dementia has been reported to be a comorbidity of type 2 diabetes mellitus in older people, but the pathological mechanisms linking both conditions remains unclear. The major gene associated with pathology of dementia, especially Alzheimer's disease, is the apolipoprotein E (APOE) gene, which has also been implicated in predisposition to diabetes. Given the scarcity of epidemiological studies on possible interactions so far, we examined the association of type 2 diabetes alone and combined with APOE genotype with incidence of dementia in a cohort of older Germans followed for approximately 14 years while accounting for competing risk of death. Methods We used data from an on-going population-based prospective cohort study (ESTHER) in the state of Saarland in the south-west of Germany. In total, 9949 elderly participants aged 50–74 years were recruited from July 2000 to December 2002 by their general practitioners followed by 5 waves of interviews up to now. Data were collected using standardized self-administered questionnaires including health-related information and basic sociodemographics. Self-reported medical diagnoses of diabetes were validated by participants’ general practitioners at each interview. Dementia-related information was collected retrospectively among all ESTHER participants, including those who dropped out or deceased over the follow-up. APOE genotype was measured using TaqMan SNP genotyping assays and participants were divided into e4 allele carriers and non-carriers. Standard Cox proportional hazard regression models were used to analyze the association of diabetes, as well as diabetes combined with APOE e4 carrier status, with incidence of all-cause dementia and its main subtypes (vascular dementia and Alzheimer's disease). Fine and Gray competing risk models were used to analyze the association of diabetes and dementia after adjusting for competing risk of death. Results Of 5648 eligible participants with complete information, 793 had type 2 diabetes (mean age 63.8 ± 6.4 years) at baseline, and 4855 did not have type 2 diabetes (mean age 61.4 ± 6.5 years). Among the overall sample, 304 developed dementia, including 90 with Alzheimer's disease and 113 with vascular dementia. In the fully-adjusted Cox model, the risk of all-cause dementia was 61% higher (hazard ratio [HR]: 1.61; 95% confidence interval [CI]: 1.22–2.12) among type 2 diabetes patients than among non-diabetes patients. Adjusted HRs for Alzheimer's disease and vascular dementia were 2.30 (95% CI: 1.38–3.86) and 1.41 (95% CI: 0.89–2.25), respectively. After adjusting for competing risk of death, type 2 diabetes was associated with an adjusted HR of 1.47 (95% CI: 1.10–1.96) for all-cause dementia, 2.18 (95% CI: 1.27–3.75) for Alzheimer's disease and 1.27 (95% CI: 0.78–2.02) for vascular dementia. Compared with those who were neither type 2 diabetes patient nor APOE e4 carrier, those with both risk factors had significantly higher risk of Alzheimer's disease (HR: 4.10; 95% CI: 1.67–10.07), vascular dementia (HR: 2.41; 95% CI: 1.14–5.09) and all-cause dementia (HR: 2.18; 95% CI: 1.37–3.46), and the HRs were higher than that of each factor alone. Conclusions Our data indicates that type 2 diabetes is associated with increased incidence of dementia in older population, and particularly with Alzheimer's disease. The association between diabetes and dementia was attenuated after accounting for competing risk of mortality but remained strong. Carrying the APOE e4 genotype further increases the risk of dementia in type 2 diabetes patients, especially for Alzheimer's disease.