Abstract
BACKGROUND: Our previous multicenter case-control study showed that aging, up-regulation of platelet glycogen synthase kinase-3β (GSK-3β), impaired olfactory function, and ApoE ϵ4 genotype were associated with cognitive decline in type 2 diabetes mellitus (T2DM) patients. However, the causal relationship between these biomarkers and the development of cognitive decline in T2DM patients remains unclear. METHODS: To further investigate this potential relationship, we designed a 6-year follow-up study in 273 T2DM patients with normal cognitive in our previous study. Baseline characteristics of the study population were compared between T2DM patients with and without incident mild cognitive impairment (MCI). We utilized Cox proportional hazard regression models to assess the risk of cognitive impairment associated with various baseline biomarkers. Receiver operating characteristic curves (ROC) were performed to evaluate the diagnostic accuracy of these biomarkers in predicting cognitive impairment. RESULTS: During a median follow-up time of 6 years (with a range of 4 to 9 years), 40 patients (16.13%) with T2DM developed MCI. Participants who developed incident MCI were more likely to be older, have a lower education level, have more diabetic complications, a higher percentage of ApoE ϵ4 allele and a higher level of platelet GSK-3β activity (rGSK-3β) at baseline (P<0.05). In the longitudinal follow-up, individuals with higher levels of rGSK-3β were more likely to develop incident MCI, with an adjusted hazard ratio (HR) of 1.60 (95% confidence interval [CI] 1.05, 2.46), even after controlling for potential confounders. The AUC of the combination of age, rGSK-3β and ApoEϵ4 allele predicted for incident MCI was 0.71. CONCLUSION: Platelet GSK-3β activity could be a useful biomarker to predict cognitive decline, suggesting the feasibility of identifying vulnerable population and implementing early prevention for dementia.