Abstract
BACKGROUND AND PURPOSE: The prevalence of dementia is rapidly increasing. Attempts to further understand modifiable risk factors such as diabetes mellitus (DM) are urgently needed to inform public health policies for prevention. Thus, the objective of the current study was to assess the relationship between DM and risk of dementia and non-dementia mortality amongst women in the California Teachers Study prospective cohort. METHODS: Women (n = 124,509) aged 22-104 years at baseline were included. DM was ascertained from self-reported questionnaires and hospital-linked records. Dementia-related deaths were ascertained from state and national records. Competing risk regression models were used to estimate cause-specific hazard ratios and 95% confidence intervals for the association of DM with dementia- and non-dementia-related mortality. RESULTS: There were 10,511 total DM cases and 3625 deaths due to dementia over a mean of 21.3 years of follow-up. Fully adjusted cause-specific hazard ratios of the association with DM were 2.26 (2.01, 2.55) for dementia-related and 1.97 (1.89, 2.05) for the competing risk of non-dementia-related mortality. This association was strongest amongst participants with incident DM, younger age at baseline and higher alcohol consumption or who were overweight. CONCLUSIONS: In the California Teachers Study, women with DM had increased risk of mortality due to both dementia and non-dementia causes; however, the risk of mortality due to dementia was elevated compared to non-dementia causes only amongst participants with incident DM.