Evaluation of the glymphatic system using the DTI-ALPS index in type 2 diabetes mellitus-induced cognitive impairment

INTRODUCTION

Type 2 diabetes mellitus (T2DM) is a significant global health challenge, with a rising prevalence rate of 11.2% in China.^[1] Among T2DM patients, approximately 30–60% are affected by cognitive impairments, manifesting as memory decline, attention deficits, and slowed processing speed.^[2] These cognitive deficits not only exacerbate the course of diabetes but also elevate healthcare costs and reduce the quality of life for affected individuals. Notably, T2DM patients are at twice the risk of developing Alzheimer’s disease (AD) compared to non-diabetic individuals,^[3,4] underscoring the urgent need for early detection and intervention for diabetes-related cognitive impairment.

The mechanisms underlying cognitive impairment in T2DM are multifactorial, with pathological changes in the central nervous system driven by a complex interplay of metabolic, vascular, and inflammatory factors.^[5-7] Emerging evidence points to glymphatic dysfunction as a key contributor to cognitive decline in diabetes.^[1] The glymphatic system is a recently discovered waste clearance pathway that facilitates the removal of metabolic waste, including amyloid-beta (Aβ), from the brain. Disruptions to this system in diabetes may impair the clearance of toxic proteins, contributing to neurodegeneration and cognitive impairment.^[1]

The glymphatic system’s function hinges on the movement of cerebrospinal fluid (CSF) through perivascular spaces (PVSs), facilitated by the water channel protein aquaporin-4 (AQP4) on astrocytic foot processes.^[8,9] A key player in this intricate mechanism is the water channel protein AQP4, located on the astrocytic foot processes, facilitating the transportation of CSF into the brain parenchyma.^[10,11] Research has linked glymphatic dysfunction to the pathogenesis of several neurological disorders, including AD, epilepsy, stroke, and normal-pressure hydrocephalus.^[12-15] In particular, studies have shown that T2DM is associated with compromised glymphatic function. For example, whole-brain magnetic resonance imaging (MRI) studies on diabetic rats have demonstrated reduced clearance of interstitial fluid in the hippocampus and hypothalamus – regions critical for cognitive function and vulnerable to Alzheimer’s pathology.^[16]

In addition, both animal and human studies support the notion that T2DM impairs glymphatic activity,^[17] suggesting a broader role for glymphatic dysfunction in diabetes-related cognitive decline.

MRI has become an increasingly valuable tool for noninvasive evaluation of the glymphatic system.^[18] One innovative method, PVS diffusion tensor imaging analysis (DTI-ALPS), was introduced by Taoka et al. to assess glymphatic function in the human brain.^[14] This technique has since been applied to conditions such as dementia, idiopathic rapid eye movement sleep disorder, and Parkinson’s disease.^[19-21] However, few studies have applied DTI-ALPS to evaluate glymphatic dysfunction in T2DM patients,^[1] despite the clear need for non-invasive biomarkers in this population.

The objective of the present study is to utilize this advanced DTI-based approach to evaluate glymphatic function in T2DM patients with cognitive impairment by calculating the ALPS index. Furthermore, this study aims to explore the correlation between the ALPS index and clinical markers in T2DM, such as glycated hemoglobin (HbA1C), vitamin D (VitD), Montreal Cognitive Assessment (MoCA) scores, and high-density lipoprotein cholesterol (HDL-C). By employing this state-of-the-art imaging technique, we aim to deepen our understanding of the relationship between glymphatic dysfunction and cognitive decline in T2DM, potentially offering novel insights for both clinical management and therapeutic intervention.

MATERIAL AND METHODS

Image post-processing and evaluation

DTI Studio software was used in this study to measure DTI metrics (https://www.mristudio.org/). Briefly, the steps were as follows: (i) raw DTI data of a single individual were imported into the software; (ii) automatic image registration was performed; and (iii) the diffusion tensor was calculated, including a color-coded fractional anisotropy (FA) map and diffusivity in the directions of the x-axis, y-axis, and z-axis. Since, in the lateral ventricular body plane, the direction of the PVS is perpendicular to the ventricular wall (mostly in the left-right direction/X-axis direction), which is also perpendicular to the direction of projection fibers (mainly in the Z-axis direction) and association fibers (mainly in the Y-axis direction), the diffusivity along the X-axis at least partially represents the diffusivity along the PVS direction. On the FA pseudo-color map of the lateral ventricular body plane [Figure 1a], regions of interests (ROIs) with a diameter of 5 mm were outlined for the projection fibers, association fibers areas in the left hemisphere (all subjects were right-handed), obtaining the diffusivity in the X, Y, and Z directions for each fiber [Figure 1b].

Close

Figure 1:

Schematic diagram of the alignment of three types of fibers at the level of the lateral ventricle. (a) Fraction of anisotropy pseudo-color map shows the direction of the three types of fibers, projection fibers in the Z-axis direction (blue), united fibers (green) in the Y-axis direction, and subcortical fibers (red) in the X-axis direction, and the diffusion coefficients of these fibers were measured by placing a 5-mm diameter regions of interest in each of the three types of fibers. (b) The direction of perivascular spaces (PVSs) (gray columns) in relation to the direction of the three types of nerve fibers, with the direction of PVS perpendicular to the direction of projection fibers and united fibers.

Export to PPT

The activity of the glymphatic system was evaluated using the DTI-ALPS index.^[22] The DTI-ALPS index is defined as follows: ALPS index = mean (Dxxproj, Dxxassoc)/mean (Dyyproj, Dzzassoc). Here, Dxxproj and Dxxassoc are the Dx values (Dx represents diffusivity in the X-axis direction) for projection fibers and association fibers, Dyyproj is the diffusivity in the Y-axis direction for projection fibers, and Dzzassoc is the diffusivity in the Z-axis direction for association fibers.

RESULTS

Comparison of DTI-ALPS index between T2DM patients with cognitive impairment and the HC group

The comparison of the DTI-ALPS index between the T2DM with cognitive impairment group and the HC group is summarized in Table 2. Independent samples t-tests were performed to assess the differences in ALPS, DxxProj, Dxxass, DyyProj, Dyyass, DzzProj, and Dzzass between the two groups. The analysis revealed no statistically significant difference in DyyProj (P > 0.05) [Figure 2a]. However, the T2DM group with cognitive impairment demonstrated significantly lower values in ALPS, DxxProj, Dxxass, Dyyass, DzzProj, and Dzzass compared to the HC group, with statistically significant differences observed (P < 0.05) [Figure 2a and b].

Table 2: Comparison of DTI-ALPS index between two groups of patients.

Indicator	T2DM with cognitive impairment group (n=40)	Healthy control group (n=40)	t/Z/χ2	P-value
ALPS	1.21±0.12	1.55±0.2	9.142	<0.001
DxxProj	0.59±0.07	0.67±0.08	4.542	<0.001
Dxxass	0.62±0.09	0.78±0.13	6.745	<0.001
DyyProj	0.51±0.09	0.53±0.09	1.104	0.273
Dyyass	1.01±0.15	1.12±0.15	3.366	0.001
DzzProj	0.97±0.14	1.04±0.14	2.039	0.045
Dzzass	0.39±0.09	0.47±0.1	3.986	<0.001

T2DM: Type 2 diabetes mellitus, ALPS: Along the perivascular space, DTI: Diffusion Tensor Imaging

Close

Figure 2:

(a and b) Comparison of Diffusion Tensor Imaging - Along the Perivascular Space (DTI-ALPS) index between two groups of patients. *Comparison of the type 2 diabetes mellitus with cognitive impairment group and the healthy control group, P < 0.05. The x-axis scale for part (a) includes the categories: DdxProj, Ddxass, DyProj, Dyass, DdzProj, and Ddzal.

The x-axis scale for part (b) includes the categories: Experimental group and Control group.

Export to PPT

Correlation analysis between ALPS index and clinical data in patients with T2DM

The correlation between the ALPS index and various clinical parameters is summarized in Table 3. Spearman correlation analysis revealed no statistically significant associations between the ALPS index and gender, IL-6, FINS, or C-P levels (P > 0.05).

Table 3: The correlation between ALPS index and clinical data.

Correlation	ALPS
	r	P-value
Gender	−0.228	0.157
Age	−0.224	0.165
BMIb	−0.157	0.334
Diastolic blood pressure	−0.235	0.145
Systolic blood pressure	−0.149	0.358
IL-6a	0.177	0.274
HbA1Cb	−0.557	<0.001
TGb	0.200	0.216
TCb	0.173	0.286
LDL-Cb	−0.104	0.522
FPGb	−0.181	0.263
2hPGb	0.027	0.871
INS fastinga	−0.255	0.112
C-P fastinga	−0.243	0.131
VitDb	0.646	<0.001
MoCAb	0.676	<0.001
HDL-Cb	0.385	0.014
ASTb	0.202	0.212
ALTb	0.281	0.079
eGFRb	0.090	0.579
CysCb	−0.078	0.632
UREAb	0.240	0.136
CREAb	0.028	0.864

aSpearman correlation analysis was used, ^bPearson correlation analysis was used. BMI: Body mass index, IL-6: Interleukin-6, HbA1C: Glycated hemoglobin, TG: Triglycerides, TC: Total cholesterol, LDL-C: Low-density lipoprotein cholesterol, FPG: Fasting plasma glucose, 2hPG: 2-h postprandial glucose, INS fasting: Insulin fasting, C-P fasting: C-peptide fasting, VitD: vitamin D, MoCA: Montreal cognitive assessment, HDL-C: High-density lipoprotein cholesterol, AST: Aspartate aminotransferase, ALT: Alanine aminotransferase, eGFR: Estimated glomerular filtration rate, CysC: Cystatin C, CREA: Creatinine, ALPS: Along the Perivascular Space

Pearson correlation analysis was used to explore the relationship between the ALPS index and a range of clinical variables, including age, BMI, systolic blood pressure, diastolic blood pressure, HbA1C, TG, TC, LDL-C, FPG, 2-h plasma glucose (2hPG), VitD, MoCA scores, HDL-C, AST, ALT, eGFR, CysC, urea, and CREA.

Analysis showed that the ALPS index was significantly positively correlated with HDL-C, MoCA score, and VitD levels (r > 0, P < 0.05) [Figure 3a-c] and significantly negatively correlated with HbA1C levels (r < 0, P < 0.05) [Figure 3d]. No statistically significant correlations were found between the ALPS index and age, BMI, systolic or diastolic blood pressure, TG, TC, LDL-C, FPG, 2hPG, AST, ALT, eGFR, CysC, urea, or CREA (P > 0.05).

Close

Figure 3:

(a-d) Scatterplot of correlation between ALPS index and clinical data. ALPS: Along the Perivascular Space, HDL: High-Density Lipoprotein, MoCA: Montreal Cognitive Assessment, HbA1C: Hemoglobin A1C.

Export to PPT

DISCUSSION

The DTI-ALPS index offers a promising metric for evaluating cognitive impairment, particularly in individuals with elevated blood glucose levels. In cases of diabetes-related cognitive dysfunction, alterations in the DTI-ALPS index reflect disruptions in the glymphatic system. This study utilized the DTI-ALPS method to assess diffusion imaging data, demonstrating its utility as a quick and efficient tool for monitoring glymphatic activity over time. Specifically, we compared diffusivity in the PVS at the lateral ventricle body level to the projection and association fibers. This comparison provides key insights into subtle changes in glymphatic function, enhancing our understanding of how these processes are altered in diabetes-related cognitive decline.

T2DM is associated with an increased risk of cognitive impairment and AD, with diabetes implicated as a significant risk factor for these conditions. A critical neuropathological mechanism in diabetic cognitive impairment is the accumulation of protein breakdown products, notably Aβ plaques, which adversely affect cognitive function.^[23,24] Previous studies have highlighted the glymphatic system’s role in Aβ clearance, linking impaired interstitial fluid (ISF) clearance in diabetes mellitus to cognitive decline.^[25,26]

In diabetes, inflammation in the PVS can damage neurovascular units, leading to increased PVS and features characteristic of diabetic brains and other dementia-related conditions.^[27,28] The interplay between Aβ deposition and neurovascular damage following glymphatic dysfunction may create a positive feedback loop, worsening glymphatic impairment in the diabetic brain. Our study demonstrates that T2DM patients with cognitive impairment have a significantly lower DTI-ALPS index compared to HCs. This suggests that glymphatic dysfunction may play a key role in the pathophysiology of cognitive impairment and cerebrovascular diseases in this population. These findings highlight the intricate relationship between diabetes, glymphatic function, and cognitive health, emphasizing the importance of glymphatic dysfunction in diabetes-related cognitive decline.

Previous clinical studies have consistently linked elevated FPG with cortical thinning and cognitive decline.^[29-35] Impaired fasting glucose and glucose tolerance have also been associated with cognitive deficits,^[36,37] indicating that brain damage may precede the onset of diabetes by several years. These findings underscore the need for early detection and management of chronically elevated FPG in individuals at risk for diabetes. Furthermore, elevated HbA1c and FPG levels are linked to increased mortality and the risk of cardiovascular and cerebrovascular events, irrespective of diabetic status.^[37-39] Compelling evidence also suggests that elevated blood glucose or HbA1c levels significantly increase the risk of cognitive decline and dementia.^[40] These insights highlight the broader implications of chronic hyperglycemia, stressing the importance of glucose management to mitigate cognitive and cardiovascular risks.

VitD deficiency has been associated with cognitive impairment progression.^[41-44] The lower VitD levels observed in diabetic individuals with cognitive impairment may contribute to elevated DTI-ALPS index. VitD plays a critical role in promoting neuronal growth and differentiation, vital for normal brain function.^[45] In addition, it regulates neurotransmitter synthesis and release, affecting neuronal communication.^[46] Consequently, VitD deficiency may lead to structural and functional neuronal abnormalities, exacerbating cognitive decline. Our study revealed a positive correlation between VitD levels and the DTI-ALPS index, suggesting that VitD could be explored as a novel indicator for assessing glymphatic activity. This notion is supported by previous findings that DTI-ALPS effectively measures glymphatic function.^[14]

However, further investigation is needed to validate VitD as an indicator of glymphatic activity and its potential to enhance assessments of glymphatic dynamics and cognitive health.

This study represents the first investigation of glymphatic function in individuals with both T2DM and cognitive impairment, revealing significant glymphatic dysfunction in this population. These findings advance our understanding of the neurological mechanisms underlying cognitive dysfunction in T2DM. However, several limitations must be acknowledged. The relatively small sample size limits the generalizability of our findings. Furthermore, the current method only computes diffusion rates along the Dx, Dy, and Dz axes, restricting diffusivity measurements to regions along the lateral ventricle outside the lateral ventricular body plane. This regionally confined analysis may overlook important glymphatic changes in other brain areas, such as the hippocampus, prefrontal cortex, or cerebellum, which are also relevant to cognitive function. Future work should consider integrating whole-brain imaging techniques or multi-shell DTI acquisitions to provide a more comprehensive assessment of glymphatic activity. This method does not allow for independent evaluation of diffusion along the PVS in areas not aligned with the primary fiber bundles. In addition, the manual placement of ROIs introduces a degree of subjectivity into the measurements, despite efforts to ensure objectivity. Finally, this single-institution study may limit the external validity of the findings. Future research should involve larger, more diverse cohorts to further evaluate glymphatic dysfunction across different populations. In addition to the aforementioned methodological constraints, several potential confounding variables may have influenced our findings. Although we excluded participants with known neurological, psychiatric, or systemic comorbidities, lifestyle-related factors – such as physical activity level, sleep duration and quality, dietary habits, and psychosocial stress – were not formally evaluated or adjusted for in this study. These factors are known to impact glymphatic function through modulation of cerebrovascular tone, astrocytic AQP4 polarization, and interstitial fluid dynamics. For instance, sleep disturbances have been shown to impair glymphatic clearance efficiency, particularly during non-REM sleep when CSF flow is enhanced. Physical inactivity may similarly contribute to vascular stiffness and reduced perivascular pulsatility, both of which are critical to CSF-ISF exchange. Furthermore, although none of the T2DM patients were on medication at the time of enrollment, prior exposure to antihypertensive or lipid-lowering agents could have long-term effects on glymphatic circulation, which were not accounted for in our cross-sectional analysis.

These unmeasured confounding factors may partly explain the variability in ALPS index among individuals with similar glycemic profiles, highlighting the complexity of glymphatic dysfunction in diabetes. Future studies should incorporate standardized questionnaires and objective measures (e.g., actigraphy, polysomnography, dietary logs) to capture these behavioral and physiological parameters. Multivariate regression models or stratified analyses could then be employed to isolate the independent effect of diabetes on glymphatic impairment, thereby improving causal inference. In this context, longitudinal designs would also allow for temporal evaluation of how lifestyle interventions or pharmacological treatments modulate glymphatic activity in T2DM patients over time. Despite these limitations, this study offers valuable clinical insights into the neurological abnormalities associated with T2DM and cognitive impairment.

CONCLUSION

The findings of this study confirm the presence of glymphatic dysfunction in individuals with T2DM and cognitive impairment, as evidenced by the DTI-ALPS index. This research not only reinforces the link between glymphatic system dysfunction and T2DM-related cognitive decline but also establishes the DTI-ALPS index as a promising and reliable tool for evaluating glymphatic activity in this population. Given its demonstrated effectiveness, the DTI-ALPS methodology offers significant potential for future research aimed at further elucidating the role of the glymphatic system in T2DM and its associated cognitive impairments.