Chronic kidney disease (CKD) is a condition characterized by a gradual loss of kidney function. Estimated glomerular filtration rate (eGFR) is an objective indicator for evaluating glomerular filtration capacity and is usually used to reflect renal function. End-stage renal disease (ESRD) is the final stage of chronic kidney disease (CKD), which is defined as an eGFR of less than 15 ml/min/1.73 m². ESRD patients require renal replacement therapy via maintenance dialysis or kidney transplantation to maintain life (National Kidney Foundation, 2002; Romagnani et al., 2017). CKD induces a series of pathophysiological processes which in turn affect the structural and functional integrity of the brain, which is more pronounced in ESRD patients (Lu et al., 2015). In addition, ESRD patients are often accompanied by cognitive decline. A study has confirmed that ESRD is an independent risk factor for cognitive impairment (O’Lone et al., 2016).

Neuroimaging research has greatly promoted our understanding of the ESRD-related neurological changes. Most of the previous quantitative analyses focused on the abnormal structure of gray or white matter (Zhang et al., 2013; Murea et al., 2015; Meurs et al., 2016). In recent years, more and more studies have confirmed that the integrity of white matter microstructure in ESRD patients has also changed using various analytical methods based on diffusion tensor imaging (DTI), including region of interest (ROI), voxel-based analysis (VBA), and tract-based spatial statistics (TBSS) (Chou et al., 2013, 2019; Kong et al., 2014; Chen et al., 2015; Yin et al., 2018; Liu et al., 2020). However, the location and range of decrease of white matter integrity varied differently across studies. For example, a ROI-based study found that compared with the control group, FA values in all regions were reduced (Hsieh et al., 2009). Results from VBA studies showed FA reduction within the middle cerebellar peduncle, bilateral sagittal stratum, and the genu and splenium of the corpus callosum (Chou et al., 2013). TBSS results have shown that lower FA values within the bilateral corona radiata (Kong et al., 2014; Yin et al., 2018), bilateral inferior fronto-occipital fasciculus (IFOF) (Eldehni et al., 2019), bilateral superior longitudinal fasciculus (SLF) (Zhang et al., 2015; Yin et al., 2018), left anterior thalamic radiation (ATR) (Zhang et al., 2015), the body of the corpus callosum (Drew et al., 2017), and bilateral inferior longitudinal fasciculus (ILF) (Eldehni et al., 2019). In addition, ROI-based analysis lacks consistent standard and largely dependent on the personal evaluation. Because of the differences in the shape of the white matter fiber bundles between subjects, VBA is not accurate enough at the individual level (Wassermann et al., 2011). Although TBSS can be more accurate than VBA, it still fails to provide specific properties for positioning the white matter integrity along each fiber bundle (Bach et al., 2014). Automatic fiber quantification (AFQ) is a new algorithm that can automatically identify the main fiber bundles, allowing researchers to quantify diffusion indicators at anatomically equivalent locations along the fiber tracts (Yeatman et al., 2012). This tract-oriented automatic quantization approach provides more detailed diffusion parameters when mean measurement along each tract is not obvious.

To our best knowledge, this is the first study of applying the AFQ technique to explore potential regional white matter fiber alterations in ESRD patients. We hypothesize that abnormal white matter integrity existed in specific regions of tracts in ESRD patients. We also explored the relationships of DTI metrics with clinical and laboratory characteristics to further investigate the locations of brain structure changes that are associated with certain clinical manifestations.

In this study, we first investigate the specific alterations of white matter microstructure in patients with ESRD by AFQ and further explored the relationships of white matter integrity with MoCA scores and biochemical indicators. We found the following points: (1) ESRD patients had extensive microstructure fragility of white matter fiber bundles, especially on the left side of the hemisphere; (2) changes of white matter microstructure might not only occur along the entire white matter fiber bundle, but also at some specific fiber segments; and (3) the point-wise level analysis was more sensitive than fiber tract level analysis and white matter fiber damage might be associated with clinical symptoms and cognitive function.

There are several mechanisms for the development of cognitive dysfunction related to white matter microstructure fragility in ESRD patients. First, the accumulation of neurotoxins caused by dialysis or uremia can cause central nervous system complications (Smogorzewski, 2001). Second, asymptomatic cerebral infarction, white matter lesions, and microhemorrhage caused by degenerative microvascular disease can cause vascular dementia (Arismendi-Morillo and Fernández-Abreu, 2010). Third, hemodialysis leads to changes in cerebral blood flow, leading to brain tissue and metabolic disorders (Kanai et al., 2001). Fourth, increased oxidative stress and chronic inflammation in ESRD may lead to endothelial and neuronal damage (Bugnicourt et al., 2013).

In our research, we applied four diffusion metrics to evaluate the alterations in the microstructure of white matter fiber bundles. FA and MD reflect the overall directionality and magnitude of water diffusion, respectively (Le Bihan et al., 2001). RD and AD denote the extent of diffusion perpendicular and parallel to fiber orientations, respectively (Assaf and Pasternak, 2008). All these metrics are purported to be sensitive to myelin and axon damage. Therefore, by using MD, AD, and RD values together with the FA value, white matter changes in ESRD patients can be better characterized. In comparison with HCs at the overall fiber level, ESRD patients exhibited widespread damaged tract profiles, with significant FA reduction in 43.75% (7/16) and MD, AD, and RD increase in 75% (12/16), 31.25% (5/16), and 75% (12/16) of the fiber tracts examined, respectively. And point-wise level analysis allowed us to obtain more accurate damaged segments, especially in MD and RD. These damage segments were mainly observed on fibers included projection fiber (CST), commissural fibers (forceps minor, forceps major), and association fibers (IFOF, ILF, SLF), especially for the association fibers, which have changed in all the four diffusion metrics. IFOF, as the longest associative bundle, is an important link connecting the frontal, temporal, and occipital lobes (Caverzasi et al., 2014) and plays a major role in the neuromotor function and processing of vison and hearing. Studies have found that the loss of WM integrity in IFOF have been considered a risk factor for Alzheimer’s disease, and Alzheimer’s disease patients show extensive damage along the IFOF bundle, including the anterior, posterior, and central region associated with memory impairment (Fu et al., 2014; Chen H. F. et al., 2020). Our findings indicated that ESRD patients have extensive damage to the left IFOF, including the anterior, posterior, and central regions, and mainly characterized in the form of decreased FA values and increased MD values, indicating that the microstructure changes of IFOF are mainly myelin damage. ILF is a critical pathway that primarily connects the occipital cortex and temporal lobe. Impaired integrity of ILF and SLF is associated with language comprehension and attention-deficit disorders (Wolfers et al., 2015; Shin et al., 2019). These results are similar to previous TBSS studies on CKD (Yin et al., 2018; Liu et al., 2020), while more detailed localization information of damaged segments of fibers can be determined in our current study based on AFQ analyses. Meanwhile, these results indicated that MD and RD values are more sensitive than FA in discovering the extent of white matter microstructure vulnerability, which consistent with previous studies (Amlien and Fjell, 2014; Jin et al., 2017; Chen H. F. et al., 2020). In addition, this study found that almost all segments of forceps minor had abnormal white matter microstructures, and with the accumulation of urea, the microstructure of nodes 16–26 in the forceps minor tended to damage. The corpus callosum is composed of white matter tracts that connect the left and right cerebral hemispheres and is the main commissural region of the brain. And the forceps minor is the fiber that connects the lateral and medial regions of the bilateral frontal lobes, which is involved in processing speed (Kerchner et al., 2012). Patients with ESRD can experience cognitive changes in different aspects, and the decline in executive function is the most obvious (Berger et al., 2016). Based on the point of view, this study found that with the decline of renal function, the microstructure of the frontal part of the forceps minor may be damaged, which may lead to impaired cognitive function. Unfortunately, our study did not carry out tests related to executive function, and further research is needed.

Besides, our study found that most of the damaged fiber bundles were in the left hemisphere, suggesting that the white matter vulnerability may have hemispheric heterogeneity. Previous study on microstructure damage in Alzheimer’s disease has also found heterogeneity in brain structure changes between hemispheres (Chen H. et al., 2020). At present, studies have shown that patients with CKD are more prone to result in cognitive decline and Alzheimer’s disease (Etgen et al., 2012; Zhang et al., 2020). Therefore, we speculate that these changes in brain white matter structure of ESRD patients may be related to the occurrence and development of Alzheimer’s disease.

In addition, this study found that changes in the diffusion metrics of ATR were related to the decline in cognitive level, especially visuospatial/executive. The ATR, which connects the mediodorsal and anterior thalamic nuclei (ATN) with the frontal cortex and the anterior cingulate cortex, can process incoming information from the hippocampus (related to cognitive functions) (Setiadi et al., 2021). Therefore, the decreased integrity of ATR may be associated with decreased cognitive functions. Different from the hemispheric heterogeneity reflected by the comparison between groups, we found that the cognitive-related ATR white matter damage were mainly located in the right hemisphere, which may be related to the priority of damage first appearing in the dominant hemisphere. Besides, a study related to reduced attention of young bilingual adults showed that the integrity of the white matter fibers connecting the anterior cingulate cortex and the frontal regions was decreased, and this manifestation only exists in the right hemisphere (Mamiya et al., 2018). Similarly, Huang et al. (2020) also reported that the right ATR was associated with executive function in cerebral small vessel disease patients.

Excitingly, this work found that specific white matter segments were closely associated with specific laboratory test results, such as urea, UA, and Cre. In particular, through multiple linear regression analysis, these specific segments corresponding to specific renal function indexes were determined. Put simply, renal dysfunction in ESRD patients may mediate the corresponding specific white matter segment abnormalities. In this study, we found that with the increase of urea level, mean AD value of the left ATR and nodes of 66–100 in the left CST tended to be damaged. Besides, our study showed positive correlations of urea level with the increased MD in the left ATR. Our interpretation of this result is that impaired renal function can result in the continuous accumulation of metabolites, which in turn leads to demyelination and axonal damage (Mogi and Horiuchi, 2011; Lu et al., 2015). Previous TBSS studies on ESRD have also demonstrated that serum creatinine is related to white matter integrity (Kim et al., 2011; Yin et al., 2018).

Despite the advantages of this method, we should not ignore some limitations of this study. First, our prospective study was conducted on a relatively small sample, and further investigations on relatively large population are still needed. Second, due to the threshold setting in fiber tracking, some fibers such as the CH and AF were adjacent to the gray matter, which makes them impossible to be tracked. Third, in this study, we only focused on patients with ESRD, and in the future, patients with early CKD should be enrolled to explore the correlation between disease progression and white matter microstructure changes. Fourth, due to participant compliance, not all participants had completed the cognitive assessments, which may have a certain impact on the relevant results.

In conclusion, this study identified specific abnormalities in major white matter fibers in ESRD using a tract profile approach. Besides, the results of our study indicates that the accumulation of serum urea and uric acid may lead to vulnerable white matter microstructure. We believe that the findings of this work will enhance our understanding of white matter abnormalities in ESRD.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

This study was approved by the Ethics Committee of the First Affiliated Hospital of Dalian Medical University. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

YJ and YL: guarantor of integrity of entire study. YJ, YL, BG, YC, JJ, and PC: literature research. YJ, YL, BG, YC, JJ, PC, QS, LL, NW, WW, and YM: clinical studies. YJ, YL and YM: experimental studies. YJ, YL, and JJ: statistical analysis. YJ, YL, LL, and YM: manuscript editing. All authors: study concepts, study design, data acquisition, data analysis, interpretation, manuscript drafting, manuscript revision for important intellectual content, approval of final version of submitted manuscript, agrees to ensure any questions related to the work are appropriately resolved.

LL was employed by the company Philips Healthcare China. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.