Association of diffusion tensor imaging along the perivascular space index with cognitive impairment in type 2 diabetes mellitus.

Abstract

Background: Type 2 diabetes mellitus (T2DM) often coexists with cognitive impairment, but the exact mechanisms underlying this relationship remain unclear. Glymphatic dysfunction may be a significant factor contributing to cognitive impairment. This study aimed to assess the changes in diffusion tensor imaging along the perivascular space (DTI-ALPS) index in patients with T2DM and to determine its association with cognitive impairment and whole-brain white-matter damage.

Methods: We prospectively recruited patients with T2DM and healthy controls (HCs) for a case-control study in which all participants underwent cognitive assessments and underwent magnetic resonance imaging (MRI) scans. We calculated the DTI-ALPS index of HCs and patients with T2DM based on the DTI. Furthermore, we calculated whole-brain fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), radial diffusivity (RD), and structural connectivity parameters, including the average clustering coefficient (Cp), characteristic path length (Lp), global efficiency (Eglob), and small-world index (Sigma) to assess brain white-matter damage. A general linear model was used to analyze the differences between the groups, which was followed by multiple linear regression analysis to examine the relationship among the DTI-ALPS index, MRI parameters, and cognitive function.

Results: This study enrolled 70 patients with T2DM (age 66.37±9.85 years; 48 males) and 35 HCs (age 64.03±11.26 years; 21 males). There were 35 patients with T2DM with mild cognitive impairment (T2DM-MCI). The DTI-ALPS index and whole-brain FA of patients with T2DM were reduced, while the whole-brain MD, AD, and RD were increased, especially in patients with T2DM-MCI. The Eglob of structural connectivity was decreased and Lp increased in patients with T2DM. No significant differences were found in the Cp or the Sigma between the T2DM group and HCs. We found that in the patients with T2DM, the DTI-ALPS index was associated with Montreal Cognitive Assessment [β =0.577; 95% confidence interval (CI): 0.360 to 0.793; P false discovery rate (P_FDR)<0.001], FA (β =0.340; 95% CI: 0.069 to 0.611; P_FDR=0.043), MD (β =-0.338; 95% CI: -0.610 to -0.065; P_FDR=0.043), AD (β =-0.440; 95% CI: -0.712 to -0.168; P_FDR=0.016), and RD values (β =-0.313; 95% CI: -0.581 to -0.044; P_FDR=0.046). No association was found between the DTI-ALPS index and the structural connectivity parameters. In addition, mediation analysis showed that FA partially mediated the association of decreased DTI-ALPS index with cognitive impairment in patients with T2DM (β =0.086; mediation effect 14.86%).

Conclusions: Our study demonstrated that the DTI-ALPS index is decreased in patients with T2DM, especially in those with T2DM-MCI. A decrease in DTI-ALPS index is associated with cognitive impairment and whole-brain white-matter damage in patients with T2DM.