Abstract
Background:
Carotid intima–media thickness (cIMT) is considered a risk factor for and predictor of cerebrovascular disease. In this study, we explored the contribution of cIMT to cognitive impairment (CI) in dialysis patients and the role of cerebral small vascular disease (CSVD) and brain atrophy in this process.
Methods:
Cognitive function was assessed using a comprehensive cognitive test battery. CSVD and brain volume were assessed by magnetic resonance imaging, and cIMT was measured by ultrasonography. Multivariable analysis and mediation were used to explore the relevant relationships among cIMT, CI, CSVD and brain volume.
Results:
Seventy-three dialysis patients were enrolled. Approximately 54.8% were diagnosed with increased cIMT. The increased cIMT group was older and had lower serum albumin and creatinine levels than the normal cIMT group. There was no difference in the CSVD prevalence between the different cIMT groups. Patients in the normal, unilaterally and bilaterally increased cIMT subgroups demonstrated a gradual decrease in brain-matter volume and degenerate cognitive function. cIMT was related to cognitive function and gray-/white-matter volume. Increased cIMT was associated with a significantly increased risk of a reduced Mini Mental State Examination/Montreal Cognitive Assessment score and Trail A/B time delay. Mediation analysis showed that CI was mediated by brain-matter volume but not by CSVD.
Conclusion:
Increased cIMT was an independent risk factor for impairment of global cognitive function, memory, and executive function. The impact of cIMT on cognition was not induced by CSVD but by brain atrophy. cIMT may be a useful tool for screening patients at high risk of CI in the dialysis population.
Keywords: carotid intima–media thickness, cerebral small vessel diseases, cognition, dialysis, gray matter, white matter
Introduction
Dialysis patients (end-stage renal disease, ESRD-5D) suffer from a higher prevalence and worse outcome of cerebrovascular disease than the general population.1,2 In addition to stroke, dialysis patients also have a higher prevalence of cerebral small vessel disease and cognitive decline.3,4
Cerebrovascular disease and carotid-artery disease are both risk factors for cognitive impairment (CI) in the general population.5,6 Carotid intima–media thickness (cIMT) is widely used as a noninvasive index to assess carotid atherosclerosis. The results from the Cardiovascular Health Study showed that in patients without clinically evident cerebrovascular disease, cognitive decline was associated with severe stenosis of the left internal carotid artery.6 In an animal model, it was shown that relief of bilateral carotid stenosis could improve cognitive function.7 However, the mechanisms of cognitive impairment in patients with asymptomatic carotid stenosis are still not well understood, with one possible explanation being the absence of a good model. Chronic kidney disease (CKD) is considered an independent risk factor for carotid atherosclerosis.8 A recent study demonstrated decreased cerebral blood flow (CBF) during the dialysis procedure,9 which may magnify the influence of carotid atherosclerosis on brain lesions in ESRD patients. Thus, the ESRD population would be an optimal model for discovering the relationships among carotid atherosclerosis, cognitive function, and CSVD.
Therefore, we hypothesized that carotid atherosclerosis is partly responsible for CI in dialysis patients and that subclinical CSVD and brain atrophy may be involved in the pathogenesis of this process.
Materials and methods
Study design
This is a retrospective study. The present study was approved by the Ethics Committee of Peking Union Medical College Hospital (PUMCH; IRB number S564). Written informed consent was obtained from each participant.
Participants
In the 2013–2014 CSVD-CI study, 416 maintenance dialysis patients from our dialysis center were screened from July 2013 to July 201410 with the following criteria: age 18 years or older and regular dialysis for at least 1 month. Patients with the following criteria were excluded from the study: systemic infection, acute cardiovascular or stroke events, cancer, surgery, or trauma in the month prior to the study; metabolic encephalopathy, mental disorders, emotional disorders, epilepsy requiring long-term treatments, or diagnosis of dementia; history of nonatherosclerotic arterial disease, such as Anti-neutrophil cytoplasmic antibody (ANCA) vasculitis; metal prosthesis; and inability to complete the cognitive assessment.
We recruited patients from the original study who had at least one carotid-artery ultrasound record with cIMT measurement results within a year before or after the brain magnetic resonance imaging (MRI) procedure based on the Hospital Information System at PUMCH.
According to cIMT measurements, patients were divided into a normal cIMT group and an increased cIMT group. In the increased cIMT group, patients were further categorized by dividing them into unilaterally and bilaterally increased cIMT subgroups.
Clinical and laboratory data
Demographics, comorbid conditions, dialysis parameters and physical examination results including dialysis modality, dialysis vintage, smoking status, history of hypertension, diabetes mellitus, hyperlipidemia, self-reported stroke, ischemic heart disease, and related medications were recorded. Laboratory data from blood examinations conducted during the same period as the brain MRI evaluation were evaluated. For hemodialysis (HD) patients, the blood sample was taken at the first HD session in a week.
Imaging and analyses
Carotid-artery ultrasound
Philips iU22 (Bothell Everett Highway, Bothell, USA) color Doppler ultrasonic diagnostic instruments with a probe frequency of 7.5 MHz were used for carotid ultrasonography. Using the carotid bulb as an anatomic reference, measurements were performed 10 mm below the carotid bifurcation in the common carotid arteries, at the bifurcation level, and 10 mm after the carotid bifurcation in the internal carotid arteries. The intima–media thickness of the common carotid artery was defined as the mean of the maximum intima–media thicknesses of the near and far walls. The greater value of the right or left common cIMT (CAA-IMT) was used for analysis. Values of cIMT ⩾1 mm were defined as increased cIMT;11 otherwise, they were defined as normal cIMT. The sonographers were blinded to the clinical information.
Brain MRI and CSVD
The participants underwent brain magnetic resonance (MR) examinations on a 3.0 T MR scanner (DISCOVERY MR750; General Electric, Milwaukee, WI, US) using an eight-channel head coil. Imaging sequences included T1-weighted and T2-weighted sequences, fluid-attenuated inversion recovery and susceptibility-weighted imaging MR sequences. One radiologist and one neurologist blinded to the clinical information independently assessed the presence and location of different signs of recent lacunar infarcts, lacunes (similar to the definition for chronic lacunar infarcts), white-matter hyperintensities (WMHs), and cerebral microbleeds (CMBs).10,12–14 Discrepancies were resolved by consensus after discussion.
Brain volume measurement
Structural three-dimensional (3D) T1-weighted MRI images were acquired using a 3.0-T Siemens Vision scanner. VBM8, a toolbox of SPM8 software (https://www.fil.ion.ucl.ac.uk/spm/software/spm8/), was used for the morphometric analysis, which was applied with the DARTEL algorithm. First, the structural images were segmented into gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) using a standard unified segmentation model in SPM8. The sum of GM and WM was defined as brain matter (BM), while the whole-brain volume was the total volume of BM and CSF. The percentage of the GM volume in the whole brain was defined as the GM fraction; the percentage of WM volume in the whole brain was defined as the WM fraction; and the percentage of the BM volume in the whole brain was defined as the BM fraction. After that, the DARTEL approach was performed for registration, normalization, and modulation in the DARTEL space. In detail, a new DARTEL template was constructed based on the deformation fields from the segmentation procedure, and all the individual deformation field maps were registered to the new template. Then, the GM images were normalized to a study-specific template in the Montreal Neurological Institute space. All the brain region volumes were measured by calculating the total number of voxels in each image’s data, with resampling voxel size of 1 cm isotropic and spatially smoothed with 8 mm full width at half-maximum Gaussian kernel.15
Cognitive assessments
Medical staff members from our dialysis center were trained by neurologists to administer the cognitive tests. Cognitive assessments were performed on an intradialysis day for HD patients and during clinical visits for peritoneal dialysis patients.
In the cognitive test battery, global cognitive function was assessed with the Chinese versions of the Mini Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA).4 The specific domains of cognitive function were evaluated by the Chinese version of the 12-word Philadelphia Verbal Learning Test (czP(r)VLT-12)16 to assess immediate, short-delayed and long-delayed verbal memory, the Boston Naming Test17 to assess language ability, and Trail-Making Tests A and B18 to assess executive function (Trail B is more complicated than Trail A). The scores of czP(r)VLT-12 ranged from 0 to 12, with higher scores denoting better performance. Higher scores on the Boston Naming Test represent higher ability. The time taken to complete the Trail A/B test performance was recorded.
Statistical analysis
Continuous data with a normal distribution are presented as the means ± standard deviation and were analyzed by Student’s t-test and one-way analysis of variance. The data with high skew are presented as the median and interquartile range and were analyzed by the Mann–Whitney test and the Kruskal–Wallis test. Categorical data are presented as proportions and were analyzed by the chi-square test and Fisher’s exact test. Correlations between variables were assessed by Pearson’s coefficient test. The differences in various cognitive scores and brain volumes in the normal, unilateral and bilateral cIMT increased subgroups were compared with the trend test. Variables with p < 0.05 in the univariate analyses were selected as candidate variables for the multivariable regression analyses, which were performed to investigate whether cIMT was associated with impaired cognitive function. The SPSS 21.0 statistical package (IBM Corporation, Armonk, NY, US) was used for all analyses, and a value of p < 0.05 was considered statistically significant. In addition, the fit curve of age and cIMT with various cognitive functions was performed with Origin (Originlab Corporation, Northampton, MA, US), R (http://www.R-project.org), and EmpowerStats software (www.empowerstats.com, X&Y solutions, Inc. Boston, MA, US).
Mediation of the association of cIMT with cognitive function by CSVD and brain volume
Mediation analysis was performed using the template described by Baron and Kenny and the method described by Vanderweele and Vansteelandt.19 Figure 1 illustrates the hypothetical causal model. To perform mediation analysis, it is necessary to test three pathways: step 1, the association of cIMT with CSVD or BM (a); step 2, the association of CSVD or BM with cognitive function (b); and step 3, the association of cIMT with cognitive function (c’). If all three associations are confirmed, mediation can be established in a fourth step through pathway (c). According to Baron and Kenny, the mediation of the relationship is full when (c) is not significant and partial when (c) is significant. Mediation is absent when not all steps are satisfied. All pathways were tested using univariable and multivariable regression analysis.
Figure 1.
Model of the hypothetical causal pathway from cIMT to cognitive function in dialysis patients.
Mediation of the association of cIMT with cognitive function by CSVD and brain volume: total effect (c) = direct effect (c’) + indirect effect (ab).
BM, brain matter; cIMT, carotid intima–media thickness; CSVD, cerebral small vascular disease; GM, gray matter; WM, white matter.
Results
From our dialysis-CSVD-CI study cohort, 73 dialysis patients had undergone carotid-artery ultrasonography within 1 year before or after the brain MRI scan and thus were recruited in this study.
In our cohort, 54.8% (40/73) of patients were diagnosed with increased cIMT, and 13/73 patients had unilateral cIMT increments (10 were left carotid arteries, 3 were right carotid arteries). All our patients were right handed. A total of 37.0% (27/73) were male, and the average age was 57.9 years. Furthermore, 67.1% (49/73) of them were on HD, and the rest were on peritoneal dialysis. The median dialysis vintage was 56.0 months. Of the patients, 94.5% were hypertensive, 28.8% were diabetic, 45.2% had hyperlipidemia, and 8.2% had a history of stroke (Table 1).
Table 1.
Clinical characteristics of patients with normal and elevated cIMT.
| Total (n = 73) | Normal cIMT group (n = 33) | Increased cIMT group (n = 40) | p value | |
|---|---|---|---|---|
| cIMT, mm | 1.0 (0.8–1.2) | 0.8 (0.7–0.9) | 1.2 (1.0–1.4) | <0.001 |
| Male, n (%) | 27 (37.0%) | 11 (33.3%) | 16 (40.0%) | 0.557 |
| Age, years | 57.9 ± 11.5 | 52.5 ± 11.3 | 62.4 ± 9.6 | <0.001 |
| Level of education, n (%) | 0.673 | |||
| Elementary school | 3 (4.1%) | 1 (3.0%) | 2 (5.0%) | |
| Middle school | 15 (20.5%) | 5 (15.2%) | 10 (25.0%) | |
| High school | 24 (32.9%) | 11 (33.3%) | 13 (32.5%) | |
| >High school | 31 (42.5%) | 16 (48.5%) | 15 (37.5%) | |
| Smoking (n, %) | 16 (21.9%) | 6 (18.2%) | 10 (25.0%) | 0.483 |
| Hypertension (n, %) | 69 (94.5%) | 32 (97.0%) | 37 (92.5%) | 0.622 |
| Diabetes mellitus, n (%) | 21 (28.8%) | 9 (27.3%) | 12 (30.0%) | 0.798 |
| Hyperlipidemia, n (%) | 33 (45.2%) | 14 (42.4%) | 19 (47.5%) | 0.665 |
| Self-reported stroke history, n (%) | 6 (8.2%) | 2 (6.1%) | 4 (10.0%) | 0.683 |
| Med_PLT, n (%) | 25 (34.2%) | 8 (24.2%) | 17 (42.5%) | 0.102 |
| Med_heparin, n (%) | 49 (67.1%) | 22 (66.7%) | 27 (67.5%) | 0.940 |
| Med_RAASI, n (%) | 35 (47.9%) | 16 (48.5%) | 19 (47.5%) | 0.933 |
| Dialysis modality, HD, n (%) | 49 (67.1%) | 22 (66.7%) | 27 (67.5%) | 0.940 |
| Dialysis vintage, months | 56.0 (16.0–109.0) | 63.0 (17.5–104.5) | 48.0 (14.3–109.0) | 0.546 |
| BMI, kg/m2 | 23.0 ± 3.4 | 23.0 ± 3.2 | 22.9 ± 3.6 | 0.926 |
| SBP, mmHg | 139.4 ± 16.2 | 138.4 ± 14.1 | 140.3 ± 17.8 | 0.614 |
| DBP, mmHg | 78.1 ± 11.1 | 80.3 ± 11.0 | 76.2 ± 11.0 | 0.116 |
| Hemoglobin, g/l | 113.6 ± 9.5 | 112.4 ± 10.3 | 114.6 ± 8.8 | 0.333 |
| Serum albumin, g/l | 38.3 ± 4.2 | 39.4 ± 3.8 | 37.4 ± 4.4 | 0.046 |
| Serum phosphorus, mmol/l | 1.7 ± 0.4 | 1.7 ± 0.4 | 1.7 ± 0.5 | 0.832 |
| Serum PTH, pg/ml | 252.0 (144.5–486.0) | 361.0 (196.5–723.5) | 215.0 (114.5–385.5) | 0.052 |
| Triglycerides, mmol/l | 1.6 ± 0.9 | 1.6 ± 0.7 | 1.6 ± 1.0 | 0.716 |
| Total cholesterol, mmol/l | 4.5 ± 1.1 | 4.7 ± 1.2 | 4.4 ± 1.0 | 0.281 |
| HDL-C, mmol/l | 1.2 ± 0.4 | 1.2 ± 0.4 | 1.2 ± 0.3 | 0.992 |
| LDL-C, mmol/l | 2.5 ± 0.8 | 2.7 ± 0.9 | 2.4 ± 0.8 | 0.172 |
| Hs-CRP, mg/l | 2.3 (0.8–6.2) | 2.0 (0.8–4.3) | 2.6 (1.0–8.3) | 0.333 |
| Log triglycerides | 0.1 ± 0.2 | 0.2 ± 0.2 | 0.1 ± 0.3 | 0.313 |
| Log total cholesterol | 0.6 ± 0.1 | 0.7 ± 0.1 | 0.6 ± 0.1 | 0.338 |
| Log HDL-C | 0.0 ± 0.1 | 0.0 ± 0.1 | 0.0 ± 0.1 | 0.939 |
| Log LDL-C | 0.4 ± 0.1 | 0.4 ± 0.1 | 0.4 ± 0.1 | 0.170 |
| Log hs-CRP | 0.4 ± 0.5 | 0.3 ± 0.5 | 0.4 ± 0.6 | 0.364 |
| Serum creatinine, μmol/l | 927.2 ± 247.3 | 1007.8 ± 204.4 | 860.7 ± 261.9 | 0.010 |
The p value is for the comparison between the normal and increased cIMT groups.
Bolded numerals indicate statistical significance.
BMI, body mass index; cIMT, carotid intima–media thickness; DBP, diastolic blood pressure; HD, haemodialysis; HDL-C, high-density lipoprotein cholesterol; hs-CRP, high-sensitivity C-reactive protein; LDL-C, low-density lipoprotein cholesterol; MAP, mean arterial pressure; Med_RAASI, renin–angiotensin–aldosterone-system-inhibitors medication use; Med_PLT, antiplatelet medication use; PTH, parathyroid hormone; SBP, systolic blood pressure.
Between the normal and increased cIMT groups (Table 1), sex, education level, combined diseases, dialysis modality and vintage, average blood pressure, RAASI (renin–angiotensin–aldosterone-system inhibitors) medication use, serum phosphorus levels, parathyroid hormone, and lipid profiles were comparable. However, those in the increased cIMT group were older (62.4 ± 9.6 years versus 52.5 ± 11.3 years) and had lower serum albumin (37.4 ± 4.4 g/l versus 39.4 ± 3.8 g/l) and lower serum creatinine levels (860.7 ± 261.9 μmol/l versus 1007.8 ± 204.4 μmol/l) than the normal cIMT patients.
CSVD and brain volume in different cIMT groups
There was only a tendency for more CSVD lesions (lacunes, WMH, and CMB) in the increased cIMT group than in the normal cIMT group, but none of the three types of CSVD reached statistical significance (Table 2).
Table 2.
CSVD and brain volume of patients in different cIMT groups.
| Total (n = 73) | Normal cIMT group (n = 33) |
Increased cIMT group (n = 40) | p value | Unilateral increased cIMT group (n = 13) | Bilateral increased cIMT group (n = 27) |
p for trend* | |
|---|---|---|---|---|---|---|---|
| Lacune, n (%) | 14 (19.2%) | 4 (12.1%) | 10 (25.0%) | 0.164 | 3 (23.1%) | 7 (25.9%) | 0.175 |
| CMB, n (%) | 24 (32.9%) | 7 (21.2%) | 17 (42.5%) | 0.054 | 4 (30.8%) | 13 (48.1%) | 0.029 |
| WMH, n (%) | 25 (34.2%) | 9 (27.3%) | 16 (40.0%) | 0.254 | 3 (23.1%) | 13 (48.1%) | 0.100 |
| Whole-brain volume (cm3) | 1473.6 ± 133.6 | 1476.5 ± 118.0 | 1471.3 ± 146.7 | 0.870 | 1519.2 ± 146.3 | 1447.3 ± 143.6 | 0.408 |
| Gray-matter volume (cm3) |
572.8 ± 60.1 | 591.2 ± 65.1 | 557.8 ± 51.7 | 0.018 | 580.7 ± 57.0 | 546.3 ± 45.7 | 0.004 |
| White-matter volume (cm3) | 451.0 ± 55.9 | 466.1 ± 56.8 | 438.652.5 | 0.038 | 460.1 ± 50.0 | 427.8 ± 51.3 | 0.009 |
| Brain-matter volume (cm3) | 1023.9 ± 108.9 | 1057.4 ± 114.4 | 996.4 ± 97.2 | 0.018 | 1040.8 ± 100.4 | 974.1 ± 89.3 | 0.003 |
| Gray-matter fraction (%) | 39.0 ± 3.5 | 40.1 ± 3.6 | 38.1 ± 3.2 | 0.015 | 38.3 ± 2.8 | 38.0 ± 3.5 | 0.021 |
| White-matter fraction (%) | 30.6 ± 2.7 | 31.6 ± 2.8 | 29.8 ± 2.4 | 0.007 | 30.3 ± 2.2 | 29.6 ± 2.5 | 0.005 |
| Brain-matter fraction (%) | 69.6 ± 5.5 | 71.6 ± 5.7 | 67.9 ± 4.9 | 0.004 | 68.6 ± 4.3 | 67.6 ± 5.2 | 0.004 |
n indicates the number of patients with the relevant characteristics.
The p value is for the comparison between normal and increased cIMT groups.
p for trend, comparison of normal, unilateral and bilateral cIMT increased subgroups.
Bolded numerals indicate statistical significance.
cIMT, carotid intima–media thickness; CMB, cerebral microbleed; CSVD, cerebral small vascular disease; WMH, white-matter hyperintensity.
Between the normal and increased cIMT groups, there was no difference in whole-brain volume, but the increased cIMT group showed significant decreases in GM, WM, and BM volume and corresponding fractions (Table 2). Moreover, there was a significant gradually declining trend in the normal, unilateral and bilateral increased cIMT subgroups.
We also analyzed the relationship between cIMT and different brain volumes. The cIMT was negative relative to the GM volume (r = −0.412, p < 0.001), WM volume (r = −0.290, p = 0.014), BM volume (r = −0.376, p = 0.001), and corresponding fractions but not relative to whole-brain volume (Supplemental Table S1).
Cognitive function and cIMT
Cognitive function in different cIMT groups
The patients with increased cIMT showed poorer global cognitive performance on the MMSE and MoCA than the normal cIMT participants, and more patients were diagnosed with CI (Table 3). The increased cIMT group also had poorer memory function on the czP(r)VLT-12, especially for immediate- and long-delayed memory; they also showed worse executive function in Trail B. The two different cIMT groups exhibited a similar language ability regarding the Boston Naming score and executive function in Trail A. We further compared patient cognitive function in the normal, unilaterally and bilaterally increased cIMT subgroups, showing a trend of more carotid-artery involvement and more severe CI in global cognitive function, memory, and executive function (Figure 2).
Table 3.
Cognitive function of patients with normal and elevated cIMT.
| Total (n = 73) | Normal cIMT group (n = 33) | Increased cIMT group (n = 40) | p value | |
|---|---|---|---|---|
| MMSE score | 29.0 (27.0–30.0) | 29.0 (28.0–30.0) | 29.0 (26.0–30.0) | 0.219 |
| CI-MMSE, n (%) | 13 (17.8%) | 1 (3.0%) | 12 (30.0%) | 0.003 |
| MoCA score | 24.0 (20.5–26.0) | 25.0 (23.0–27.0) | 22.0 (19.0–25.8) | 0.007 |
| CI-MoCA, n (%) | 47 (64.4%) | 17 (51.5%) | 30 (75.0%) | 0.037 |
| czP(r)VLT-12 | ||||
| Immediate memory score | 7.4 (5.9–8.6) | 8.2 (6.8–9.2) | 7.0 (5.1–8.0) | 0.006 |
| Short-delayed memory score | 8.0 (7.0–10.0) | 9.0 (7.0–10.5) | 8.0 (6.0–9.0) | 0.065 |
| Long-delayed memory score | 8.0 (6.0–10.0) | 9.0 (7.0–10.0) | 7.0 (5.0–9.0) | 0.029 |
| Boston Naming score | 34.0 (30.0–36.5) | 34.0 (31.0–36.0) | 32.5 (30.0–37.0) | 0.480 |
| Trail A, s | 60.0 (38.5–81.5) | 50.0 (34.0–78.0) | 60.5 (42.8–86.0) | 0.183 |
| Trail B, s | 112.0 (75.0–180.0) | 83.0 (64.0–149.0) | 138.5 (90.3–225.8) | 0.006 |
Data are presented as n, number of patients included (and percentage) in the prevalence section; data are presented as median (and interquartile range) in non-prevalence setion.
The p value is for the comparison between normal and increased cIMT groups.
Bolded numerals indicate statistical significance.
CI-MMSE, cognitive impairment evaluated by MMSE; CI-MoCA, cognitive impairment evaluated by MoCA; cIMT, carotid intima–media thickness; czP(r)VLT-12, Philadelphia Verbal Learning Test; MMSE, Mini Mental State Examination; MoCA, Montreal Cognitive Assessment.
Figure 2.
Performances of various cognitive functions by cIMT groups.
*p for trend, comparison of normal, unilateral and bilateral cIMT increased subgroups. The x axis represents the cIMT group (dots: normal cIMT group; squares: unilateral increased cIMT group; and triangles: bilateral increased cIMT group), and the y axis shows various cognitive functions.
cIMT, carotid intima–media thickness; MMSE, Mini Mental State Examination; MoCA, Montreal Cognitive Assessment.
As people age, cognitive function declines. The patients in the increased cIMT group demonstrated a steeper slope of deterioration in all aspects of cognitive function with age than those in the normal cIMT group (Figure 3).
Figure 3.
Linear fit chart of age with various cognitive function grouping by normal and increased cIMT.
The x axis represents age, and the y axis represents various cognitive functions.
Dots: normal cIMT group; squares: unilateral increased cIMT group.
β1, linear regression coefficient of the normal cIMT group; β2, linear regression coefficient of the increased cIMT group.
cIMT, Carotid intima–media thickness; MMSE, Mini Mental State Examination; MoCA, Montreal Cognitive Assessment.
Effect of cIMT on cognitive function
As cIMT increased, a declining tendency was seen in global cognitive function in the whole cohort and the domains of memory and executive function, as shown in Figure 4.
Figure 4.
Fit curve of cIMT with various cognitive functions adjusted for age.
To demonstrate the influence of cIMT on cognitive function, the correlations between cIMT and various cognitive functions were assessed by partial correlation analysis, adjusted for age confounders. ‘r’ represents the correlation coefficient between cIMT and cognitive function.
The x axis represents cIMT, and the y axis represents various cognitive functions.
cIMT, carotid intima–media thickness.
To investigate the effect of carotid-artery stenosis on global and different domains of cognitive function, we further built several models to analyze the relationship between carotid-artery stenosis and cognitive testing scores in the whole cohort (Table 4). In the crude model, both the global cognitive assessment score and each specific aspect of cognitive function showed negative relevance to the thickness of the cIMT. In model 1, after adjusting for age, education level, antiplatelet medication, dialysis modality, and serum albumin and creatinine levels, cIMT still showed an effect on global cognitive function [for the MMSE, β = −1.28 (95% confidence interval −2.23 to −0.33); for the MoCA, β = −2.29 (95% confidence interval −4.17 to −0.40)] and executive function [for Trail A, β = 24.14 (95% confidence interval 6.73–41.55); for Trail B, β = 60.91 (95% confidence interval 26.31–95.52)]. In model 2, we adjusted for three types of CSVD based on model 1; the influence of cIMT on cognitive function remained the same as in model 2. Based on model 2, we further adjusted for BM volume in model 3. In this model, the influence of cIMT on different domains of cognition was decreased compared with that in model 2.
Table 4.
Multiple linear regression analysis of cIMT associated with impaired cognitive function.
| Crude |
Model 1 |
Model 2 |
Model 3 |
|||||
|---|---|---|---|---|---|---|---|---|
| β (95% CI) | p value | β (95% CI) | p value | β (95% CI) | p value | β (95% CI) | p value | |
| MMSE score | −1.80 (−2.76 to −0.83) |
<0.001 | −1.28 (−2.23 to −0.33) | 0.009 | −1.31 (−2.24 to −0.38) | 0.007 | −1.15 (−2.13 to −0.18) | 0.021 |
| MoCA score | −4.30 (−6.12 to −2.48) |
<0.001 | −2.29 (−4.17 to −0.40) | 0.018 | −2.26 (−4.07 to −0.45) | 0.015 | −1.86 (−3.76 to 0.04) | 0.055 |
| czP(r)VLT-12 | ||||||||
| Immediate memory score | −1.49 (−2.30 to −0.67) |
0.001 | −0.72 (−1.58 to 0.14) | 0.098 | −0.76 (−1.63 to 0.12) | 0.089 | −0.81 (−1.72 to 0.09) | 0.078 |
| Short-delayed memory score | −1.94 (−3.20 to −0.67) |
0.003 | −0.52 (−1.80 to 0.75) | 0.413 | −0.61 (−0.88 to 0.66) | 0.338 | −0.82 (−2.18 to 0.53) | 0.229 |
| Long-delayed memory score | −2.17 (−3.33 to −1.00) |
<0.001 | −0.89 (−2.05 to 0.29) | 0.136 | −0.96 (−2.15 to 0.23) | 0.113 | −1.13 (−2.38 to 0.13) | 0.079 |
| Boston Naming score | −2.48 (−4.45 to −0.51) |
0.014 | −0.68 (−2.76 to 1.41) | 0.518 | −0.67 (−2.74 to 1.41) | 0.522 | 0.23 (−1.87 to 2.34) | 0.824 |
| Trail A, s | 37.96 (22.02–53.90) |
<0.001 | 24.14 (6.73–41.55) | 0.007 | 23.67 (6.21–41.12) | 0.009 | 26.44 (7.77–45.12) | 0.006 |
| Trail B, s | 96.01 (62.43–129.59) |
<0.001 | 60.91 (26.31–95.52) | 0.001 | 59.36 (23.89–94.82) | 0.001 | 56.27 (19.04–93.50) | 0.004 |
Model 1: adjusted for age, education level, antiplatelet medication, dialysis modality, serum albumin and serum creatinine.
Model 2: adjusted for the variables entered into model 1 and lacunes, cerebral microbleeds, and white-matter hyperintensity.
Model 3: adjusted for the variables entered into model 2, gray-matter volume, and white-matter volume.
CI, confidence interval; cIMT, carotid intima–media thickness; czP(r)VLT-12, Philadelphia Verbal Learning Test; MMSE, Mini Mental State Examination; MoCA, Montreal Cognitive Assessment.
To provide deeper insight into the possible pathway from cIMT to CI, we also analyzed the mediation effect of the association between cIMT and cognitive function by CSVD and brain volume (Supplemental Table S2). Considering brain structure, cIMT could generate global cognitive and memory, language, and executive function impairment by full or partial mediation through GM, WM, and BM, but none of these cognitive impairments was mediated through CSVD.
Discussion
In this retrospective study, we focused on a specific population of dialysis patients who had a high burden and increased comorbidity of cardiovascular and cerebrovascular disease. CKD was shown to be an independent risk factor for carotid atherosclerosis,8 and increased cIMT is associated with early-stage20 and advanced CKD.21 cIMT is considered a surrogate for cardiovascular disease and is a variable predictor of stroke and cardiovascular events.22–24 Our original cohort also showed a high prevalence of cognitive function decline and CSVD in dialysis patients.4,10 Thus, ESRD-5D patients would be an appropriate population in which to explore the relationship between vascular factors (e.g. cIMT, carotid stenosis) and brain lesions.
In our cohort, patients with increased cIMT had worse performance in global cognitive function, memory, and executive function. The severity of increased cIMT was relative to cognitive function. Figure 2 shows more carotid-artery involvement and worsened cognitive performance. This trend was statistically significant in global function and the domains of memory and executive function. Figure 3 shows that patients with cIMT increases had a faster rate of decline in all domains of cognitive function. After adjusting for age, the thickness was still relative to the MMSE/MoCA, memory, and executive function.
To further illustrate the possible mechanism of the relationship between cIMT and cognitive function from the perspective of brain lesions, we introduced cerebrovascular disease and brain volume for analysis. Our results showed that the increased cIMT group had a significantly decreased BM volume and that these BM volumes were negative relative to cIMT. However, there was no difference in the three typical CSVD features between the two different cIMT groups.
In the cognitive function comparison, greater cIMT was significantly associated with worse performance on all cognitive tests in the unadjusted regression model. After adjustment for demographics, clinical risk factors and CSVD, cIMT remained negatively associated with global function and executive function. These findings were similar to those of several longitudinal studies.25–27 There is no doubt that CSVD is related to cognitive function. CSVD may be caused by microembolisms due to carotid plaque rupture. However, Frazier et al.26 found that the relationship between executive function decline and cIMT was independent of WM injury and cerebral infarct. Beyond their findings, we found that lacune and microbleed CSVDs did not affect the impact of cIMT on cognitive function. Carotid atherosclerosis might lead to brain hypoperfusion and atrophy. Our finding that the BM volumes were negative relative to cIMT just provided evidence for it. In the adjusted models, after adjusting for BM volumes, the relationship between cIMT and cognitive tests shrank or disappeared. These results indicated that brain atrophy may be an important mediator from cIMT to cognition. We performed a further mediation analysis. These models proved that GM, WM, and total BM mediated the association of cIMT with cognitive function in all domains, while CSVD was not involved in this process.
A study on ESRD showed that CBF was increased in patients with ESRD compared with healthy controls; however, the CBF of dialysis patients was lower than that of nondialysis ESRD patients.28 In contrast, a recent study in HD patients showed decreased CBF during the HD procedure and demonstrated a possible link between the dialysis-induced CBF decrease and cognitive decline, in which CSVD WMHs and brain atrophy may be involved.9 It seems to be controversial in CBF changes in ESRD patients. In fact, at the early stage of carotid stenosis, blood flow may accelerate, while at the late stage, blood flow decreases. In the general population, the autoregulation of CBF could compensate for the change in blood supply. However, in dialysis patients, regardless of the stage of carotid stenosis, there might be dysfunction of CBF autoregulation, which will decrease the total brain blood supply and eventually result in brain atrophy. Our findings about the role of CAA-IMT in brain volume and cognitive function specifically support the long-term influence of CBF changes in brain lesions in dialysis patients.
The treatment of carotid stenosis in dialysis is conservative. Carotid revascularization in ESRD showed a low risk–benefit ratio with higher rates of in-hospital mortality and moderate-to-severe disability.29 However, our results show a relationship between cIMT and BM volume/cognitive decline. From this point of view, monitoring the increase in cIMT and preventing it from developing carotid stenosis would be more appropriate in ESRD patients. Control of CAA-IMT might contribute to preventing cognitive decline in dialysis patients. Currently, strict control of hypertension and hyperglycemia, smoking cessation, and prevention of CKD progression are central therapies for carotid atherosclerotic disease23,30 and related brain lesions.
To the best of our knowledge, this is the first study about cIMT and cognitive function that focuses on brain volume and CSVD at the same time. However, this study still has some limitations. Our findings are based on a retrospective observational study, and further follow up on cognitive function is needed to make the results more convincing. Second, although very few studies have focused on cIMT and brain imaging changes in dialysis patients, our sample size was still relatively small compared with studies in the general population. Finally, we assessed only CAA-IMT, and assessment of the internal carotid artery may provide more information.
Conclusion
Our study supports a role for CAA-IMT in cognitive function impairment in dialysis patients, especially in global cognitive function, memory and executive function. In the relationship between cIMT and cognitive decline, BM atrophy demonstrated a more important effect than CSVD. Furthermore, dialysis patients with increased cIMT need to be more closely monitored for cognitive decline.
Supplemental Material
Supplemental material, Supp_Material_cIMT_and_CSVD_cognitive_function20200623 for Carotid intima–media thickness relative to cognitive impairment in dialysis patients, and their relationship with brain volume and cerebral small vessel disease by Ke Zheng, Yujun Qian, Tiaye Lin, Fei Han, Hui You, Xixi Tao, Bo Hou, Jing Yuan, Haiyun Wang, Dingding Zhang, Ke Lv, Feng Feng, Yicheng Zhu and Xuemei Li in Therapeutic Advances in Chronic Disease
Footnotes
Conflict of interest statement: The authors declare that there is no conflict of interest.
Funding: The authors disclose receipt of the following financial support for the research, authorship, and/or publication of this article: the contributing study was funded by the 12th National Science and Technology Support Program (2011BAI10B02) and CAMS Innovation Fund for Medical Sciences (CIFMS) (2017-I2M-2-001).
ORCID iDs: Ke Zheng 
https://orcid.org/0000-0002-4624-0881
Xuemei Li
https://orcid.org/0000-0002-5966-6240
Supplemental material: Supplemental material for this article is available online.
Contributor Information
Ke Zheng, Department of Nephrology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Yujun Qian, Department of Nephrology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Tiaye Lin, Department of Radiology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Fei Han, Department of Neurology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Hui You, Department of Radiology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Xixi Tao, Department of Diagnostic Ultrasound, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Bo Hou, Department of Radiology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Jing Yuan, Department of Neurology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Haiyun Wang, Department of Nephrology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Dingding Zhang, Central Research Laboratory, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Ke Lv, Department of Diagnostic Ultrasound, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Feng Feng, Department of Radiology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Yicheng Zhu, Department of Neurology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Xuemei Li, Department of Nephrology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, No. 1 Shuaifuyuan, Dongcheng District, Dongcheng-qu, Beijing 100730, China.
References
- 1. Hill C. Death registrations in England and Wales, 2008, causes. Health Stat Q 2009; 43: 68–76. [DOI] [PubMed] [Google Scholar]
- 2. Saran R, Robinson B, Abbott KC, et al. US renal data system 2017 annual data report: epidemiology of kidney disease in the United States. Am J Kidney Dis 2018; 71(3 Suppl. 1): A7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Zheng K, Li X, Feng F, et al. Brain magnetic resonance and cognitive function changes in maintenance hemodialysis patients. Nephrol Dial Transplant 2015; 30(Suppl. 3): iii557. [Google Scholar]
- 4. Zheng K, Wang H, Hou B, et al. Malnutrition-inflammation is a risk factor for cerebral small vessel diseases and cognitive decline in peritoneal dialysis patients: a cross-sectional observational study. BMC Nephrol 2017; 18: 366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Gottesman RF, Hillis AE. Predictors and assessment of cognitive dysfunction resulting from ischaemic stroke. Lancet Neurol 2010; 9: 895–905. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Johnston SC, O’Meara ES, Manolio TA, et al. Cognitive impairment and decline are associated with carotid artery disease in patients without clinically evident cerebrovascular disease. Ann Intern Med 2004; 140: 237–247. [DOI] [PubMed] [Google Scholar]
- 7. Duan W, Chun-Qing Z, Zheng J, et al. Relief of carotid stenosis improves impaired cognition in a rat model of chronic cerebral hypoperfusion. Acta Neurobiol Exp (Wars) 2011; 71: 233–243. [DOI] [PubMed] [Google Scholar]
- 8. Yu FP, Zhao YC, Gu B, et al. Chronic kidney disease and carotid atherosclerosis in patients with acute stroke. Neurologist 2015; 20: 23–26. [DOI] [PubMed] [Google Scholar]
- 9. Findlay MD, Dawson J, Dickie DA, et al. Investigating the relationship between cerebral blood flow and cognitive function in hemodialysis patients. J Am Soc Nephrol 2019; 30: 147–158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Qian Y, Zheng K, Wang H, et al. Cerebral microbleeds and their influence on cognitive impairment in dialysis patients. Brain Imaging Behav. Epub ahead of print 2 January 2020. DOI: 10.1007/s11682-019-00235-z. [DOI] [PubMed] [Google Scholar]
- 11. Jung KW, Shon YM, Yang DW, et al. Coexisting carotid atherosclerosis in patients with intracranial small- or large-vessel disease. J Clin Neurol 2012; 8: 104–108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Wardlaw JM, Smith EE, Biessels GJ, et al. Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol 2013; 12: 822–838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Pasi M, Charidimou A, Boulouis G, et al. Mixed-location cerebral hemorrhage/microbleeds: underlying microangiopathy and recurrence risk. Neurology 2018; 90: e119–e126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Yakushiji Y, Noguchi T, Hara M, et al. Distributional impact of brain microbleeds on global cognitive function in adults without neurological disorder. Stroke 2012; 43: 1800–1805. [DOI] [PubMed] [Google Scholar]
- 15. Zhang YD, Zhou JS, Lu FM, et al. Reduced gray matter volume in male adolescent violent offenders. PeerJ 2019; 7: e7349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Bezdicek O, Libon DJ, Stepankova H, et al. Development, validity, and normative data study for the 12-word Philadelphia verbal learning test [czP(r)VLT-12] among older and very old Czech adults. Clin Neuropsychol 2014; 28: 1162–1181. [DOI] [PubMed] [Google Scholar]
- 17. Chen TB, Lin CY, Lin KN, et al. Culture qualitatively but not quantitatively influences performance in the Boston naming test in a Chinese-speaking population. Dement Geriatr Cogn Dis Extra 2014; 4: 86–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Saarikivi KA, Huotilainen M, Tervaniemi M, et al. Selectively enhanced development of working memory in musically trained children and adolescents. Front Integr Neurosci 2019; 13: 62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Boers AMM, Jansen IGH, Brown S, et al. Mediation of the relationship between endovascular therapy and functional outcome by follow-up infarct volume in patients with acute ischemic stroke. JAMA Neurol 2019; 76: 194–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Wu Y, Hou J, Li J, et al. Correlation between carotid intima-media thickness and early-stage chronic kidney disease: results from asymptomatic polyvascular abnormalities in community study. J Stroke Cerebrovasc Dis 2016; 25: 259–265. [DOI] [PubMed] [Google Scholar]
- 21. Asp AM, Wallquist C, Rickenlund A, et al. Aspects of carotid structure and function in health and different stages of chronic kidney disease. Clin Physiol Funct Imaging 2018; 38: 402–408. [DOI] [PubMed] [Google Scholar]
- 22. Polak JF, O’Leary DH. Carotid intima-media thickness as surrogate for and predictor of CVD. Glob Heart 2016; 11: 295–312.e3. [DOI] [PubMed] [Google Scholar]
- 23. Lee EJ, Kim HJ, Bae JM, et al. Relevance of common carotid intima-media thickness and carotid plaque as risk factors for ischemic stroke in patients with type 2 diabetes mellitus. AJNR Am J Neuroradiol 2007; 28: 916–919. [PMC free article] [PubMed] [Google Scholar]
- 24. Borras M, Cambray S, Crespo-Masip M, et al. Peritoneal dialysis is an independent factor associated to lower intima media thickness in dialysis patients free from previous cardiovascular disease. Front Physiol 2018; 9: 1743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Zeki Al, Hazzouri A, Vittinghoff E, Sidney S, et al. Intima-media thickness and cognitive function in stroke-free middle-aged adults: findings from the coronary artery risk development in young adults study. Stroke 2015; 46: 2190–2196. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Frazier DT, Seider T, Bettcher BM, et al. The role of carotid intima-media thickness in predicting longitudinal cognitive function in an older adult cohort. Cerebrovasc Dis 2014; 38: 441–447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Moon JH, Lim S, Han JW, et al. Carotid intima-media thickness is associated with the progression of cognitive impairment in older adults. Stroke 2015; 46: 1024–1030. [DOI] [PubMed] [Google Scholar]
- 28. Jiang XL, Wen JQ, Zhang LJ, et al. Cerebral blood flow changes in hemodialysis and peritoneal dialysis patients: an arterial-spin labeling MR imaging. Metab Brain Dis 2016; 31: 929–936. [DOI] [PubMed] [Google Scholar]
- 29. Adil MM, Saeed F, Chaudhary SA, et al. Comparative outcomes of carotid artery stent placement and carotid endarterectomy in patients with chronic kidney disease and end-stage renal disease. J Stroke Cerebrovasc Dis 2016; 25: 1721–1727. [DOI] [PubMed] [Google Scholar]
- 30. Tagelsir S, Gameraddin MB, Babiker MS, et al. Doppler sonographic assessment of carotid arteries in Sudanese stroke patients. Brain Circ 2017; 3: 114–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental material, Supp_Material_cIMT_and_CSVD_cognitive_function20200623 for Carotid intima–media thickness relative to cognitive impairment in dialysis patients, and their relationship with brain volume and cerebral small vessel disease by Ke Zheng, Yujun Qian, Tiaye Lin, Fei Han, Hui You, Xixi Tao, Bo Hou, Jing Yuan, Haiyun Wang, Dingding Zhang, Ke Lv, Feng Feng, Yicheng Zhu and Xuemei Li in Therapeutic Advances in Chronic Disease