Accelerated Progression from Mild Cognitive Impairment to Dementia in People with Diabetes

Weili Xu; Barbara Caracciolo; Hui-Xin Wang; Bengt Winblad; Lars Bäckman; Chengxuan Qiu; Laura Fratiglioni


Diabetes. 2010;59(11):2928-2935. 

In This Article


Cross-sectional Relationship between Diabetes and MCI

Among the 1,435 dementia-free participants at baseline, 337 were detected as MCI cases, including 131 aMCI and 206 oCIND. Compared with non-MCI subjects, people with MCI were older (OR 0.94, [95% CI 0.89–0.99, P = 0.015]), and had a lower level of education (<8 vs. ≥8 years, OR 2.13 [95% CI 1.35–3.38]) and MMSE score (OR 0.36 [95% CI 0.29–0.46]), and higher proportion of APOE ε4 carriers (OR 1.91 [95% CI 1.15–3.15]). The two groups showed no significant differences in terms of sex, stroke, heart disease, and blood pressure. Logistic regression analyses were performed to assess the cross-sectional relationship between diabetes and MCI based on the dementia-free subjects (n = 1,435). The multi-adjusted ORs related to diabetes or pre-diabetes were 1.32 (95% CI 0.82–1.31) for overall MCI, 1.12 (95% CI 0.67–1.20) for aMCI, and 1.03 (95% CI 0.57–1.23) for oCIND.

Nonparticipants of the Follow-up

Of the 1,435 individuals who were initially identified as free of dementia at baseline, 170 subjects were lost to the first follow-up (Fig. 1). Compared with participants, dropouts were younger, but the two groups had no significant differences in other features. Multiple logistic regression analysis showed that being a dropout had ORs of 0.94 (95% CI 0.91–0.98) for old age, 1.26 (95% CI 0.84–1.87) for female sex, 1.24 (95% CI 0.89–1.73) for low education, 1.02 (95% CI 0.94–1.09) for a lower score on MMSE (<24 vs. ≥24), 0.35 (95% CI 0.28–1.06) for diabetes, 0.69 (95% CI 0.47–1.09) for heart disease, and 0.66 (95% CI 0.45–1.08) for stroke.

Characteristics of the Two Cohorts

Of the 1,265 nondemented participants, 963 were cognitively intact and 302 had MCI, which constituted the two cohorts for longitudinal analyses. In the cognitively intact cohort, 56 subjects (5.8%) had diabetes, and 30 (3.1%) had pre-diabetes (Table 1). In the MCI cohort, 18 (6.0%) subjects had diabetes, and 16 (5.3%) were ascertained as having pre-diabetes. Among the 34 MCI subjects with diabetes or pre-diabetes, 27 (79.4%) were classified as oCIND, and 7 (20.6%) had aMCI (Table 2).

Diabetes and MCI in the Cognitively Intact Cohort

In the cognitively intact cohort, during the 9-year follow-up (4,656 person-years; mean per person = 4.8 years; maximum = 10.5 years), 182 participants developed MCI, including 42 aMCI and 140 oCIND. In addition, 212 were diagnosed with dementia, including 150 AD and 26 VaD. Table 3 shows the incidence rate standardized by age, sex, and education, and multi-adjusted HRs of MCI related to diabetes and pre-diabetes. We found no statistically significant association of either diabetes or pre-diabetes with incident MCI and its subgroups, whereas diabetes and pre-diabetes were significantly related to an increased risk of incident dementia. This association was present even when adjusting for several potential confounders, including APOE and vascular disorders. No interaction with APOE genotype or other variables was obtained.

Diabetes and Dementia in the MCI Cohort

In the MCI cohort, during the same period (1,092 person-years; mean per person = 3.6 years; maximum = 10.3 years), 155 subjects developed dementia, including 125 AD and 4 VaD. Using either standard or time-dependent Cox models, we found that diabetes and pre-diabetes were associated with an increased risk of developing dementia and AD in people with MCI after controlling for possible confounders, including APOE and vascular disorders (Table 4). Of the 125 AD subjects, 57 had aMCI and 68 had oCIND at baseline. We further performed stratified analysis according to MCI subgroups. Diabetes and pre-diabetes substantially increased the risk of progression from oCIND to dementia (HR 4.31 [95% CI 1.97–9.42]), but not from aMCI (HR 2.24 [95% CI 0.75–6.71]). However, the statistical power was limited for these stratified analyses because of the small number of people in each subgroup of MCI.

Kaplan-Meier survival analysis in people with MCI showed that the median time from baseline to dementia occurrence was 1.83 years (95% CI 2.44–4.24) in people with diabetes or pre-diabetes, and 5.01 years (95% CI 5.15–6.19) in people without such conditions. Thus, diabetes or pre-diabetes accelerated the progression from MCI to dementia by an average of 3.18 years (Fig. 2).

Figure 2.

Cumulative hazard for the progression from MCI to dementia by diabetes status in the MCI cohort (adjusted for age, sex, and education).

Supplementary Analyses

Similar results were obtained when we used only the incident MCI cohort that was identified at the first follow-up examination (n = 94, of whom 36 developed dementia during the next 6-year follow-up), and when we repeated the analyses among participants who survived until the time when dementia status was determined (n = 170, 136 dementia and 115 AD). Further, we performed the analyses by leaving out the participants with missing values of APOE genotype or BMI, which produced results that were the same as those from the initial analysis. Finally, in the MCI cohort, among the 34 people with either diabetes or pre-diabetes, 18 (52.9%) developed dementia and 16 (47.1%) died during the 9-year follow-up period. Diabetes and pre-diabetes were significantly associated with elevated mortality (HR 1.75 [95% CI 1.17–2.60]) after adjustment for age, sex, education, baseline MMSE score, BMI, APOE genotype, and vascular disorders.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.