Abstract
Background: Diabetes, particularly in midlife, is a risk factor for dementia. However, how cumulative fasting blood glucose (FBG) levels contribute to dementia risk is unclear.
Objective: To determine associations between cumulative FBG levels and cognitive decline.
Methods: We pooled data from 15,294 individuals aged 18 or older without dementia or stroke (55% women; mean [SD] age at first cognitive assessment 52 [21] years) from 6 cohorts (1971 to 2024): Atherosclerosis Risk in Communities Study (ARIC), Coronary Artery Risk Development in Young Adults (CARDIA) Study, Cardiovascular Health Study (CHS), Framingham Offspring Study (FOS), Multi-Ethnic Study of Atherosclerosis (MESA), and Northern Manhattan Study (NOMAS). Controlling for other known risk factors for dementia, we used linear mixed effects models to examine the association between time-dependent cumulative mean FBG level and trajectories of harmonized global cognition (primary outcome), memory, and executive function. Cognitive outcomes were standardized as T-scores (mean [SD], 50 [10]); a 1-point difference represents a 0.1-SD difference in cognition. Higher scores indicate better performance. Median (IQR) follow-up was 23 (9, 36) years. The median (IQR) number of FBG measurements was 5 (2, 7) and the median (IQR) time between first and last FBG measurement was 9 (5, 30) years.
Results: Higher time-dependent cumulative mean FBG level was associated with significantly faster declines in global cognition (-0.004 points per year faster per each 10 mg/dL increase [95% CI, -0.007 to -0.001]; P=0.006), memory (-0.014 points per year faster per each 10 mg/dL increase [95% CI, -0.021 to -0.008]; P<0.001), and executive function (-0.008 points per year faster per each 10 mg/dL increase [95% CI, -0.011 to -0.005]; P<0.001) (Table 1, Model A; Figure 1). Further adjustment for diabetes treatment at first cognitive assessment did not change the associations between FBG and cognitive decline, although diabetes treatment was associated with slower executive function decline (0.05 points per year slower [95% CI, 0.006-0.094]; P=0.03) (Table 1, Model B).
Conclusion: These results suggest that higher cumulative mean FBG levels may contribute to later-life cognitive decline. If causal, the declines in global cognition, executive function, and memory due to long-term mean FBG levels we observed, equivalent to 0.5 to 2.0 years of cognitive aging per 10 mg/dL higher mean FBG, would be clinically significant.