Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Nov 1.
Published in final edited form as: Am J Kidney Dis. 2017 Jun 23;70(5):627–637. doi: 10.1053/j.ajkd.2017.05.006

Associations Between Early Kidney Disease With Brain Magnetic Resonance Imaging and Cognitive Function in African Americans With Type 2 Diabetes Mellitus

Barry I Freedman 1, Kaycee M Sink 2, Christina E Hugenschmidt 2, Timothy M Hughes 2, Jeff D Williamson 2, Christopher T Whitlow 3, Nicholette D Palmer 4, Michael E Miller 5, Laura C Lovato 5, Jianzhao Xu 4, S Carrie Smith 4, Lenore J Launer 6, Joshua I Barzilay 7, Robert M Cohen 8, Mark D Sullivan 9, R Nick Bryan 10, Benjamin C Wagner 11, Donald W Bowden 4, Joseph A Maldjian 11, Jasmin Divers 5, on behalf of the AA-DHS MIND and ACCORD MIND Investigators
PMCID: PMC5651190  NIHMSID: NIHMS882031  PMID: 28648301

Abstract

Background

Relationships between early kidney disease, neurocognitive function and brain anatomy are poorly defined in African Americans with type 2 diabetes mellitus (T2DM).

Study Design

Cross-sectional associations were assessed between cerebral anatomy and cognitive performance with estimated glomerular filtration rate (eGFR) and urinary albumin-creatinine ratio (UACR) in African Americans with T2DM.

Setting a& Participants

African Americans with cognitive testing and cerebral magnetic resonance imaging (MRI) in the African American–Diabetes Heart Study Memory in Diabetes (AA-DHS MIND; n=512; 480 with MRI) and Action to Control Cardiovascular Risk in Diabetes (ACCORD) MIND (n=484; 104 with MRI) studies.

Predictors

eGFR (CKD-EPI creatinine equation), spot UACR.

Measurements

MRI-based cerebral white matter volume (WMV), gray matter volume (GMV), and white matter lesion volume (WMLV); cognitive performance (Mini-Mental State Examination, Digit Symbol Coding, Stroop test, and Rey Auditory Verbal Learning Test). Multivariable models adjusted for age, sex, body mass index, scanner, intracranial volume, education, diabetes duration, hemoglobin A1c, low-density lipoprotein cholesterol, smoking, hypertension and cardiovascular disease were used to test for association between kidney phenotypes and the brain in each study; a meta-analysis was performed.

Results

Mean participant age was 60.1±7.9 (SD) years; diabetes duration, 12.1±7.7 years; hemoglobin A1c, 8.3%±1.7%; eGFR, 88.7±21.6 ml/min/1.73 m2; and UACR, 119.2±336.4 mg/g. In the fully-adjusted meta-analysis, higher GMV associated with lower UACR (p<0.05), with a trend toward association with higher eGFR. Higher WMLV was associated with higher UACR (p<0.05) and lower eGFR (p<0.001). WMV was not associated with either kidney parameter. Higher UACR was associated with lower Digit Symbol Coding performance (p<0.001) and a trend toward association with higher Stroop interference; eGFR was not associated with cognitive tests.

Limitations

Cross-sectional; single UACR measurement.

Conclusions

In African Americans with T2DM, mildly high UACR and mildly low eGFR were associated with smaller GMV and increased WMLV. UACR was associated with poorer processing speed and working memory.

Keywords: African American, brain, albuminuria, chronic kidney disease (CKD), type 2 diabetes mellitus (T2DM), cognitive performance, neurocognitive function, estimated glomerular filtration rate (eGFR), urinary albumin-creatinine ratio (UACR), magnetic resonance imaging (MRI)

Compared with the general population, people with type 2 diabetes mellitus (T2DM) and advanced chronic kidney disease (CKD) have higher rates of cognitive dysfunction, more cerebral atrophy, and increased severity of white matter lesions on brain magnetic resonance imaging (MRI) scans.1–7 The associations of mildly reduced estimated glomerular filtration rate (eGFR) and albuminuria (typically assessed as the spot urine albumin-creatinine ratio [UACR]) on brain anatomy and function have been assessed less often, even though these disorders make up a much larger percentage of the population with CKD. Several studies from predominantly European-derived cohorts have associated subtle declines in executive function and possible MRI findings with mild kidney disease in people with diabetes.8, 9 The relationship between mild reductions in eGFR and low level albuminuria with brain structure and function are not well studied in African Americans with T2DM. Although it is logical to assume that these relationships would be the same as in European Americans, differences in susceptibility to T2DM, subclinical cardiovascular disease, kidney disease, and risk of cognitive decline exist between people of African and European descent.10 Brain findings may therefore differ between European Americans and African Americans and variable access to medical care may further confound relationships between brain and kidney disease in the context of diabetes.

The objective of this study was to examine whether early markers of kidney disease are associated with cognitive impairment and brain MRI changes in African Americans with T2DM. Two African American cohorts with access to medical care were analyzed. Participants had extensive cognitive function testing and brain MRI scans. In contrast to existing reports, these well characterized participants had generally preserved kidney function with low levels of albuminuria, as well as frequent receipt of medications to control blood pressure, blood sugar and serum lipids.

Methods

Setting

The African American–Diabetes Heart Study Memory in Diabetes (AA-DHS MIND) and the Action to Control Cardiovascular Risk in Diabetes (ACCORD) MIND studies evaluated cognitive performance and brain volumes with cerebral MRI in African Americans with T2DM.11–15 AA-DHS MIND is an observational genetic and epidemiologic study that assesses relationships between T2DM, cardiovascular disease, brain anatomy and neurocognitive function. ACCORD was designed to determine if intensive lowering of blood sugar levels, intensive lowering of blood pressure levels, or treatment of blood lipids with a fibrate drug plus a statin drug can reduce the risk of major cardiovascular disease events in patients with T2DM who are at especially high risk of cardiovascular disease.

Study Size and Participants

The sample consisted of all self-reported African Americans with T2DM enrolled in the Wake Forest School of Medicine AA-DHS MIND study and all African Americans with T2DM who were recruited in the ACCORD MIND sub-study. Detailed AA-DHS MIND methods have been reported.13, 14 AA-DHS consists of 717 participants with T2D actively treated with insulin and/or oral agents; those with known serum creatinine concentrations >2 mg/dL were not recruited. A total of 292 AA-DHS participants were re-examined in the AA-DHS MIND with cerebral MRI and cognitive testing and another 220 individuals meeting identical inclusion criteria were later recruited. ACCORD was a randomized, multi-center 2×2 factorial design trial that assessed two levels of hemoglobin A1c (HbA1c) control (standard versus intensive) on adjudicated major cardiovascular disease events; lipid lowering (with a statin with or without a fibrate) and blood pressure control arms (standard versus intensive) were also included. ACCORD MIND was an ancillary study that assessed the effects of blood sugar, blood pressure, and lipid lowering on cognitive function.11, 12 Detailed recruitment strategies have been reported for both studies.11–14 Medication, medical and education histories were collected and body mass index (BMI) and blood pressure recorded in all participants. These studies were approved by their respective Institutional Review Boards and adhered to the principles of the Declaration of Helsinki. All participants provided written informed consent.

Data Sources and Measurement

Biochemical Measures

Measures of HbA1c, fasting glucose and lipids were performed. Albuminuria was quantitated with a spot UACR, and eGFR was computed using the CKD-EPI (CKD Epidemiology Collaboration) equation based on isotope-dilution mass spectrometry–validated serum creatinine concentrations in each study.16

Cognitive Testing

Partially overlapping cognitive batteries were administered to AA-DHS MIND and ACCORD MIND participants. Analyses included the four cognitive tests performed in both studies. The Digit Symbol Coding test (subtest of Wechsler Adult Intelligence Scale) primarily reflects processing speed and, to a lesser extent, working memory.17 Digit Symbol Coding score was defined as the number of correct responses within 2 minutes; a higher score indicates better performance. Response inhibition, a component of executive function, was assessed with a 40-item version of the Stroop task.18 The interference score is reported, computed as [time to completion Stroop 3 + errors Stroop 3] – [time to completion Stroop 2 + errors Stroop 2], where higher scores reflect poorer executive function. The Rey Auditory Verbal Learning Test (RAVLT) is a 15-item word list recall task that tests verbal memory and learning. The RAVLT measures delayed recall, which reflects the number of words correctly recalled after a 25-minute delay ranging from 0 (worst) to 15 (best). Global cognition was assessed in the AA-DHS MIND with the Modified Mini-Mental State Examination (3MS) 19 and Montreal Cognitive Assessment 20 and in ACCORD MIND with the Mini-Mental State Examination (MMSE) 21. Analyses of global cognition in this report were limited to 3MS results from AA-DHS MIND and MMSE from ACCORD MIND.

Cerebral Magnetic Resonance Imaging

Detailed MRI imaging methods for the AA-DHS MIND have been reported.22, 23 Of 512 participants, 480 underwent MRI scanning: 61 had 1.5-T and 419 had 3.0-T MRI scans; cognitive testing was performed and kidney phenotypes determined on the day of the MRI. Fifteen individuals underwent both 1.5-T and 3.0-T scans to account for between-scanner variation, and adjustments were made in structural brain volumes based on any systematic differences in volumetric measures, as described.22, 23 In AA-DHS MIND, structural analysis of the T1-weighted images (for total intracranial volume [TICV] comprised of gray matter volume [GMV], white matter volume [WMV] and cerebrospinal fluid volume; GMV; and WMV) was performed using the SPM824 new segment procedure as implemented in the VBM8 toolbox.25 WMLV was computed from the fluid-attenuated inversion recovery (FLAIR) images using the lesion segmentation toolbox toolbox,26 for SPM8 and an optimized threshold (k) of 0.25 as previously described.22 ACCORD MIND MRI analyses (described in next paragraph) excluded the cerebellum, while the VBM8 pipeline uses a whole brain template. In order to make the values more comparable, the VBM8 results were masked to exclude the cerebellum.

ACCORD MIND 1.5-T MRI scans were performed on participants from the study sites at Columbia University, Wake Forest School of Medicine, and the University of Minnesota. MRI scans were offered to all participants until the pre-specified sample of 640 scans was complete. Of these 640 MIND participants, 104 were African American. Scans were performed a median of 49 days after cognitive testing and kidney phenotypes had been measured. ACCORD MIND MRI included a three-dimensional fast spoiled gradient-echo T1-weighted (repetition time, 21 ms; flip angle, 30°; echo time, 8 ms), two-dimensional axial fast spin-echo FLAIR (repletion time, 8000 ms; inversion time, 2000 ms; echo time, 100 ms), and proton-density/T2-weighted (repetition time, 3200 ms; echo time1,2=27 ms and 120 ms) sequences. Image data was analyzed at the ACCORD-MIND MRI Center at the Department of Radiology, University of Pennsylvania School of Medicine (R.N.B.). Processing of MRI sequences involved template registration and semi-automated segmentation.27–30 Using a previously validated but conservative technique called white matter lesion segmentation, abnormality maps based on structural MRI (proton density, T1-weighted, T2-weighted, FLAIR) were used to identify WML.27–30 White matter lesion segmentation is a support vector machine classifier trained on expert human manual segmentation of WML on FLAIR images that strongly correlates with human observer definition of WML. Both studies provided measures of GMV, WMV and WMLV.

Statistical Methods

General linear models were fitted separately for each study to test for association between measures of kidney function with MRI and cognitive test outcomes. Box-Cox31 transformations were applied where appropriate to ensure that the distributional assumptions of conditional normality and homogeneity of variance of the residuals were satisfied. Because different transformations were applied within AA-DHS-MIND and ACCORD-MIND, the resulting model coefficients are on different transformed scales, and are difficult to compare directly. To aid interpretation, the inverse of the Box-Cox transformation was applied separately in each dataset to provide association parameters that are interpretable on the original measurement scale.32 Standardized coefficients (β̂/σ̂) are presented, where β̂ represents the parameter estimates associated with the kidney function measure and σ̂ denotes the estimated residual standard deviation from the Box-Cox transformed regression model. In AA-DHS-MIND, a minimally adjusted model was fitted including age, sex, TICV and MRI scanner (1.5 T versus 3 T) as covariates for MRI outcomes. The minimally adjusted models fitted with the ACCORD-MIND data were adjusted for age, sex, total brain volume (comprised of WMV plus GMV), and further adjusted for study site. Fully adjusted models were fitted in each dataset separately. These models, in addition to the covariates included in the minimally-adjusted model, also adjusted for education level, BMI, HbA1c and hypertension. Diagnostic tests performed on the model residuals showed that the assumptions underlying the linear models were met. Similar models were used for testing for association between the kidney function measures and cognitive tests. The minimally-adjusted models included age, sex and education level as covariates in AA-DHS-MIND, and age, sex, education level and study site in ACCORD-MIND. The fully-adjusted models were further adjusted for BMI, HbA1c, hypertension (yes/no), and low-density lipoprotein (LDL) cholesterol levels. Inverse variance meta-analysis of the results observed separately in the two studies was performed using the standardized coefficients. The approach we employed controlled for minor differences in administration or test composition between studies (e.g, 3MS and MMSE). We also provide estimated I2 values, a measure of the heterogeneity between studies.

In contrast to AA-DHS MIND where 480 of 512 (93.8%) of participants had an MRI, 104 of 484 (21.5%) of African American ACCORD MIND participants had an MRI; this rate was similar to that in all ACCORD MIND participants where 640 of 2972 (21.5%) had an MRI.

Results

A total of 996 self-reported African Americans with T2DM were evaluated with cognitive testing; 512 participants from AA-DHS MIND (480 of whom had an MRI) and 484 participants from ACCORD MIND (104 of whom had an MRI). Approximately 62% of participants were female, with mean participant age of 60.1 ±7.9 (standard deviation) years; T2DM duration, 12.1 ±7.7 years; HbA1c, 8.3%±1.7%; eGFR, 88.7 ±21.6 ml/min/1.73 m2; and UACR, 119.2 ±336.4 mg/g. Demographic and laboratory characteristics of AA-DHS MIND and ACCORD MIND participants, respectively, are presented in Tables 1 and 2. The demographic and laboratory results of the African American ACCORD MIND subset with an MRI (Table S1, available as online supplementary material) were similar to those of the full sample of African Americans in ACCORD MIND (Table 2).

Table 1.

Demographic and laboratory characteristics of African American AA-DHS MIND participants

Variable at MIND Visit All Kidney disease absent* Kidney disease present* P
N Mean±SD; Median or % N Mean±SD; Median or % N Mean±SD; Median or %
Age (years) 510 58.6 (9.6) 57.6 317 58.3 (9.4) 57.0 193 59.0 (9.8) 58.3 0.4
Female sex 511 60.9% 318 61.6% 193 59.6% 0.6
BMI (kg/m2) 511 35.2 (8.4) 33.7 318 35.0 (8.3) 33.5 193 35.5 (8.6) 34.1 0.4
Education 0.1
 Some High school or less 511 [12.7%] 318 [10.7%] 193 [16.1%]
 High school graduate 511 [26.4%] 318 [27.7%] 193 [24.4%]
 Technical school/Associate degree 511 [43.4%] 318 [41.5%] 193 [46.6%]
 College graduate/Post-graduate 511 [17.4%] 318 [20.1%] 193 [13.0%]
Diabetes duration (years) 508 13.0 (7.7) 11.6 317 12.2 (7.5) 10.5 191 14.5 (7.8) 13.5 <0.001
Current or past smoker 508 [54.7%] 315 [54.3%] 193 [55.4%] 0.8
Systolic blood pressure (mm Hg) 511 131.5 (17.9) 131.0 318 128.7 (16.0) 130.0 193 136.2 (20.0) 133.0 <0.001
Diastolic blood pressure (mm Hg) 511 77.0 (11.3) 77.0 318 75.9 (10.4) 76.0 193 78.7 (12.5) 78.0 0.03
Hypertension 501 [85.8%] 310 [81.3%] 191 [93.2%] <0.001
History of prior CVD 510 [22.5%] 317 [19.2%] 193 [28.0%] 0.02
eGFR (ml/min/1.73m2) 511 88.2 (23.4) 89.0 318 94.3 (17.6) 93.0 193 78.0 (28.0) 78.0 <0.001
eGFR <60 ml/min/1.73m2 511 [12.3%] 318 [0.0%] 193 [32.6%] <0.001
eGFR <90 ml/min/1.73m2 511 [50.1%] 318 [41.5%] 193 [64.2%] <0.001
UACR (mg/g) 511 134.9 (380.6) 10.0 318 8.1 (7.2) 5.4 193 343.9 (560.6) 82.8 <0.001
UACR at MIND visit
 10–30 mg/g 511 [18.6%] 318 [27.0%] 193 [4.7%] <0.001
 >30 mg/g 511 [31.3%] 318 [0.0%] 193 [82.9%] <0.001
 30–300 mg/g 511 [21.1%] 318 [0.0%] 193 [56.0%] <0.001
 >300 mg/g 511 [10.2%] 318 [0.0%] 193 [26.9%] <0.001
Serum glucose (mg/dL) 511 150.3 (63.8) 134.0 318 141.5 (53.0) 127.5 193 164.8 (76.4) 149.0 0.001
Hemoglobin A1c (%) 510 8.1 (2.0) 7.4 317 7.8 (1.9) 7.2 193 8.5 (2.1) 7.9 <0.001
LDL cholesterol (mg/dl) 296 110.1 (35.8) 107.0 188 104.6 (33.6) 101.0 108 119.6 (37.6) 115.0 <0.001
HDL cholesterol (mg/dl) 297 48.0 (13.1) 46.0 189 48.1 (12.4) 46.0 108 47.8 (14.4) 45.0 0.5
Use of statins 298 [50.3%] 190 [47.4%] 108 [55.6%] 0.2
Use of insulin 300 [37.3%] 192 [29.7%] 108 [50.9%] <0.001
GMV (cm3) 477 479.7 (55.3) 477.1 301 483.2 (51.1) 483.4 176 473.8 (61.5) 466.2 0.03
WMV (cm3) 477 440.7 (58.0) 434.4 301 440.2 (56.8) 434.1 176 441.5 (60.2) 435.3 0.8
Total intracranial volume (cm3) 477 1141.1 (123.5) 1134.6 301 1139.2 (117.6) 1129.7 176 1144.3 (133.3) 1142.3 0.7
WMLV, automated (cm3) 474 7.3 (11.1) 2.6 300 6.8 (11.1) 2.0 174 8.3 (10.9) 3.9 <0.001
3MS, 0–100 511 85.9 (8.5) 87.0 318 86.6 (8.2) 88.0 193 84.7 (9.0) 86.0 0.03
RAVLT delayed recall, 0–15 504 6.5 (2.4) 6.2 315 6.5 (2.4) 6.2 189 6.5 (2.3) 6.6 0.9
Digit Symbol Coding, 0–133 503 50.0 (16.4) 49.0 315 51.6 (15.9) 50.0 188 47.3 (16.9) 47.0 0.003
Stroop Interference 501 34.4 (17.4) 31.0 315 33.5 (17.9) 30.0 186 36.0 (16.6) 32.5 0.07
Open in a new tab

Note: Conversion factors for units: cholesterol in mg/dL to mmol/L, ×0.02586; glucose in mg/dL to mmol/L, ×0.05551 3MS, Modified Mini-Mental State Examination; BMI, body mass index; CVD, cardiovascular disease; eGFR, estimated glomerular filtration rate; GMV, gray matter volume HDL, high-density lipoprotein; LDL, low-density lipoprotein; MIND,; UACR, urine albumin-creatinine ratio; RAVLT, Rey Auditory Verbal Learning Test; WMV, white matter volume; SD, standard deviation; AA-DHS, African American–Diabetes Heart Study; MIND, Memory in Diabetes; ACCORD, Action to Control Cardiovascular Risk in Diabetes; WMLV, white matter lesion volume.

*

Presence of kidney disease defined as eGFR <60 ml/min/1.73m2 and/or UACR >30 mg/g

Table 2.

Demographic and laboratory characteristics of African American ACCORD MIND participants

Variable at MIND Visit All Kidney disease absent* Kidney disease present* P
N Mean±SD; Median or % N Mean±SD; Median or % N Mean±SD; Median or %
Age (years) 484 61.7 (5.7) 61.0 297 61.3 (5.4) 60 187 62.4 (6.0) 62 0.04
Female sex 484 [62.6%] 297 [63.0%] 187 [62.0%] 0.8
BMI (kg/m2) 484 32.7 (5.5) 32.1 297 32.4 (5.3) 31.6 187 33.2 (5.7) 32.7 0.2
Education 0.02
 Some high school or less 484 [19.4%] 297 [15.5%] 187 [25.7%]
 High school graduate 484 [30.6%] 297 [34.3%] 187 [24.6%]
 Technical school/Associate degree 484 [30.8%] 297 [32.0%] 187 [28.9%]
 College graduate/Post-graduate 484 [19.2%] 297 [18.2%] 187 [20.9%]
Diabetes duration (years) 478 11.0 (7.7) 10.0 293 9.7 (7.1) 8 185 13.2 (8.2) 12 <0.001
Current or past Smoker 483 [52.8%] 296 [50.3%] 187 [56.7%] 0.2
Systolic blood pressure (mm Hg) 478 138.7 (18.7) 136.0 292 135.3 (17.0) 133.0 186 144.1 (19.9) 142.0 <0.001
Diastolic blood pressure (mm Hg) 478 77.2 (10.8) 77.0 292 76.7 (10.4) 77.0 186 78.0 (11.4) 77.0 0.5
Hypertension 484 [92.4%] 297 [89.9%] 187 [96.3%] 0.01
History of prior CVD 484 [20.9%] 297 [18.9%] 187 [24.1%] 0.2
eGFR (ml/min/1.73m2) 484 89.2 (19.5) 91.3 297 94.1 (15.9) 94.6 187 81.3 (21.9) 80.9 <0.001
eGFR <60 ml/min/1.73m2 484 [8.1%] 297 [0.0%] 187 [20.9%] <0.001
eGFR <90 ml/min/1.73m2 484 [47.7%] 297 [37.7%] 187 [63.6%] <0.001
UACR (mg/g) 484 102.8 (282.7) 13.6 297 9.9 (6.7) 7.5 187 250.2 (414.5) 68.5 <0.001
UACR at MIND visit
 10–30 mg/g 484 [23.1%] 297 [35.0%] 187 [4.3%] <0.001
 >30 mg/g 484 [34.5%] 297 [0.0%] 187 [89.3%] <0.001
 30–300 mg/g 484 [25.2%] 297 [0.0%] 187 [65.2%] <0.001
 >300 mg/g 484 [9.3%] 297 [0.0%] 187 [24.1%] <0.001
Serum glucose (mg/dL) 484 168.1 (63.5) 159.5 297 162.2 (58.3) 155 187 177.4 (70.2) 165 0.02
Hemoglobin A1c 483 8.6 (1.1) 8.3 296 8.5 (1.1) 8.3 187 8.7 (1.2) 8.4 0.1
LDL cholesterol (mg/dl) 484 110.1 (33.2) 105.0 297 110.3 (33.7) 104 187 109.8 (32.5) 107 0.7
HDL cholesterol (mg/dl) 484 48.5 (13.1) 47.0 297 48.7 (13.4) 47 187 48.1 (12.8) 47 0.7
Use of statins 484 [62.6%] 297 [64.6%] 187 [59.4%] 0.2
Use of insulin 484 [40.7%] 297 [33.7%] 187 [51.9%] <0.001
GMV (cm3) 104 439.6 (47.2) 427.2 69 441.1 (46.0) 425.4 35 436.8 (50.0) 428.1 0.5
WMV (cm3) 104 441.1 (53.8) 434.5 69 440.7 (53.2) 435.9 35 441.7 (55.8) 431.0 0.9
Total brain volume (cm3) 104 880.7 (87.9) 867.5 69 881.8 (88.0) 867.4 35 878.5 (88.8) 870.4 0.8
WMLV, automated (cm3) 104 1.9 (3.4) 0.8 69 1.6 (2.8) 0.6 35 2.6 (4.4) 1.4 0.05
MMSE 484 26.2 (2.9) 27.0 297 26.3 (2.8) 27 187 26.1 (3.0) 27 0.6
RAVLTest delayed recall, 0–15 484 7.1 (2.5) 7.0 297 7.2 (2.5) 7.2 187 6.8 (2.5) 6.8 0.1
Digit Symbol Coding, 0–133 479 44.5 (15.2) 45.0 294 45.9 (14.7) 47.0 185 42.3 (15.7) 42.0 0.005
Stroop Interference 480 38.0 (20.4) 34.0 295 36.6 (20.0) 32 185 40.1 (20.8) 36 0.04
Open in a new tab

Note: Conversion factors for units: cholesterol in mg/dL to mmol/L, ×0.02586; glucose in mg/dL to mmol/L, ×0.05551 MMSE, Mini-Mental State Examination; BMI, body mass index; eGFR, estimated glomerular filtration rate; GMV, gray matter volume HDL, high-density lipoprotein; LDL, low-density lipoprotein; UACR, urine albumin-creatinine ratio; RAVLT, Rey Auditory Verbal Learning Test; SD, standard deviation; WMV, white matter volume; MIND, Memory in Diabetes; ACCORD, Action to Control Cardiovascular Risk in Diabetes; WMLV, white matter lesion volume.

*

Kidney disease defined as eGFR <60 ml/min/1.73m2 and/or UACR ≥30 mg/g.

Frequencies of kidney disease (defined as UACR >30 mg/g and/or eGFR <60 ml/min/1.73m2) were nearly identical in each study: 193 of 512 (37.7%) in AA-DHS MIND and 184 of 484 (38.0%) in ACCORD MIND. Levels of eGFR were also generally preserved (mean, 88–89 ml/min/1.73m2), with median UACRs of 10.0 mg/g in AA-DHS MIND and 13.64 mg/g in ACCORD MIND participants. Among AA-DHS MIND participants, 130 (25.4%) had UACR >30 mg/g and eGFR ≥60 ml/min/1.73m2, 33 (6.5%) had UACR <30 mg/g and eGFR <60 ml/min/1.73m2, 30 (5.9%) had UACR >30 mg/g and eGFR <60 ml/min/1.73m2, and 319 (62.3%) had UACR ≤30 mg/g and eGFR ≥60 ml/min/1.73m2. The number of ACCORD MIND participants with UACR >30 mg/g, eGFR <60 ml/min/1.73m2, or both were 148 (30.6%), 20 (4.1%), and 19 (3.9%), respectively. Table S2 and S3, respectively, display characteristics of AA-DHS MIND and ACCORD MIND participants based on levels of UACR. Systolic blood pressure, HbA1c, and use of statins were slightly higher in ACCORD MIND participants; baseline LDL cholesterol and high-density lipoprotein cholesterol levels were similar across studies.

The results of association analyses between kidney phenotypes UACR and eGFR with cerebral volumes are reported in Table 3. Model 1 adjusted for MRI scanner, TICV (AA-DHS) or total brain volume (ACCORD), age and sex; Model 2 adjusted for the covariates in Model 1 plus level of education, BMI, HbA1c and presence of hypertension; Model 3 adjusted for the covariates in Model 2 plus diabetes duration, smoking and history of CVD; Model 4 adjusted for covariates in Model 3 plus eGFR; and Model 5 adjusted for covariates in Model 3 plus UACR. Findings are reported separately in each study and in a meta-analysis. Statistically significant associations were found in the fully-adjusted meta-analysis between higher UACR and lower GMV (standardized coefficient β/σ = −2.0×10−4; p<0.05) and higher UACR and higher WMLV (β/σ = 2.0×10−4; p=0.05), but not with WMV. Significant association was also present in the fully-adjusted meta-analysis between higher eGFR and lower WMLV (β/σ = −6.0×10−3; p<0.001) and a non-significant trend was seen with higher eGFR and higher GMV (β/σ = 4.0×10−3; p=0.08), but not with WMV.

Table 3.

Association between kidney function measures with MRI volumetric measures

Outcome Predictor Model AA-DHS MIND ACCORD MIND Meta-analysis
Parameter SE P Parameter SE P Parameter SE P I2
GMV (cm3) UACR 1 −4.0×10−4 1.0×10−4 <0.001 3.0×10−4 4.0×10−4 0.5 −4.0×10−4 1.0×10−4 <0.001 0.7
GMV (cm3) UACR 2 −4.0×10−4 1.0×10−4 <0.001 5.0×10−4 4.0×10−4 0.2 −3.0×10−4 1.0×10−4 <0.001 0.79
GMV (cm3) UACR 3 −4.0×10−4 1.0×10−4 <0.001 5.0×10−4 4.0×10−4 0.2 −3.0×10−4 1.0×10−4 0.01 0.78
GMV (cm3) UACR 4 −3.0×10−4 1.0×10−4 0.01 5.0×10−4 4.0×10−4 0.2 −2.0×10−4 1.0×10−4 0.05 0.75
GMV (cm3) eGFR 1 5.0×10−3 2.0×10−3 0.01 3.0×10−6 6.0×10−3 0.9 5.0×10−3 2.0×10−3 0.02 0
GMV (cm3) eGFR 2 6.0×10−3 2.0×10−3 <0.001 3.0×10−4 6.0×10−3 0.9 5.0×10−3 2.0×10−3 0.01 0
GMV (cm3) eGFR 3; 5 6.0×10−3 2.0×10−3 0.01 1.0×10−4 6.0×10−3 0.9 5.0×10−3 2.0×10−3 0.02 0
GMV (cm3) eGFR 5 4.0×10−3 2.0×10−3 0.07 6.0×10−4 6.0×10−3 0.9 4.0×10−3 2.0×10−3 0.08 0
WMV (cm3) UACR 1 −3.0×10−5 1.0×10−4 0.8 −3.0×10−4 4.0×10−4 0.5 −5.0×10−5 1.0×10−4 0.6 0
WMV(cm3) UACR 2 −5.0×10−6 1.0×10−4 0.9 −4.0×10−4 4.0×10−4 0.3 −4.0×10−5 1.0×10−4 0.7 0.1
WMV (cm3) UACR 3 4.0×10−5 1.0×10−4 0.8 −5.0×10−4 4.0×10−4 0.3 −8.0×10−6 1.0×10−4 0.9 0.31
WMV (cm3) UACR 4 2.0×10−5 1.0×10−4 0.9 −4.0×10−4 4.0×10−4 0.3 −2.0×10−5 1.0×10−4 0.8 0.20
WMV (cm3) eGFR 1 −4.0×10−4 2.0×10−3 0.9 3.0×10−3 6.0×10−3 0.6 1.0×10−4 2.0×10−3 0.9 0
WMV (cm3) eGFR 2 −7.0×10−5 2.0×10−3 0.9 3.0×10−3 6.0×10−3 0.6 3.0×10−4 2.0×10−3 0.9 0
WMV (cm3) eGFR 3; 5 −1.0×10−3 2.0×10−3 0.6 3.0×10−3 6.0×10−3 0.6 −7.0×10−4 2.0×10−3 0.8 0
WMV (cm3) eGFR 5 −1.0×10−3 2.0×10−3 0.6 3.0×10−3 6.0×10−3 0.6 −7.0×10−4 2.0×10−3 0.8 0
WMLV (cm3) UACR 1 4.0×10−4 1.0×10−4 <0.001 1.0×10−4 4.0×10−4 0.8 4.0×10−4 1.0×10−4 <0.001 0
WMLV (cm3) UACR 2 4.0×10−4 1.0×10−4 <0.001 1.0×10−4 4.0×10−4 0.8 4.0×10−4 1.0×10−4 <0.001 0
WMLV (cm3) UACR 3 4.0×10−4 1.0×10−4 <0.001 −1.0×10−5 4.0×10−4 0.9 3.0×10−4 1.0×10−4 <0.001 0
WMLV (cm3) UACR 4 3.0×10−4 1.0×10−4 0.04 −3.0×10−5 4.0×10−4 0.9 2.0×10−4 1.0×10−4 0.05 0
WMLV (cm3) eGFR 1 −8.0×10−3 2.0×10−3 <0.001 −5.0×10−3 6.0×10−3 0.4 −8.0×10−3 2.0×10−3 <0.001 0
WMLV (cm3) eGFR 2 −8.0×10−3 2.0×10−3 <0.001 −4.0×10−3 6.0×10−3 0.6 −8.0×10−3 2.0×10−3 <0.001 0
WMLV (cm3) eGFR 3; 5 −8.0×10−3 2.0×10−3 <0.001 −4.0×10−3 6.0×10−3 0.5 −8.0×10−3 2.0×10−3 <0.001 0
WMLV (cm3) eGFR 5 −7.0×10−3 2.0×10−3 <0.001 −4.0×10−3 6.0×10−3 0.5 −6.0×10−3 2.0×10−3 <0.001 0
Open in a new tab

Note: Model 1 covariates: MRI scanner, total intracranial volume (AA-DHS) or total brain volume (ACCORD), age and sex; Model 2 covariates: Model 1 + level of education, body mass index, hemoglobin A1c, and hypertension; Model 3 covariates: Model 2 + diabetes duration, smoking status, history of cardiovascular disease; Model 4 covariates: Model 3 + eGFR; Model 5 covariates: Model 3 + UACR

SE – standard error; eGFR, estimated glomerular filtration rate (in ml/min/1.73m2); GMV, gray matter volume UACR, urine albumin-creatinine ratio (in mg/g); WMV, white matter volume; WMLV, white matter lesion volume; AA-DHS, African American–Diabetes Heart Study; MIND, Memory in Diabetes; MRI, magnetic resonance imaging; ACCORD, Action to Control Cardiovascular Risk in Diabetes.

Table 4 displays results of association analyses between kidney disease phenotypes and cognitive performance. Model 1 adjusted for age, sex and level of education. Model 2 adjusted for these covariates plus BMI, HbA1c, hypertension and LDL cholesterol. Model 3 adjusted for the covariates in Model 2 plus diabetes duration, smoking and history of CVD; Model 4 adjusted for covariates in Model 3 plus eGFR; and Model 5 adjusted for covariates in Model 3 plus UACR. Statistically significant relationships were not observed in meta-analyses between kidney disease parameters and global cognition (MMSE/3MS) or memory (RAVLT). In contrast, higher UACR was associated with poorer function on Digit Symbol Coding (β/σ = −3.0×10−4; p<0.001) and revealed a non-significant trend toward higher Stroop interference (β/σ = 2.0×10−4; p=0.08).

Table 4.

Association between kidney function measures with cognitive function

Outcome Predictor Model AA-DHS MIND ACCORD MIND Meta-analysis
Parameter SE P Parameter SE P Parameter SE P I2
Digit Symbol Coding UACR 1 −4.0×10−4 1.0×10−4 <0.001 −3.0×10−4 2.0×10−4 0.04 −4.0×10−4 1.0×10−4 <0.001 0
Digit Symbol Coding UACR 2 −5.0×10−4 2.0×10−4 0.003 −3.0×10−4 2.0×10−4 0.1 −4.0×10−4 1.0×10−4 0.001 0
Digit Symbol Coding UACR 3 −4.0×10−4 1.0×10−4 0.002 −2.0×10−4 2.0×10−4 0.3 −3.0×10−4 1.0×10−4 0.002 0.24
Digit Symbol Coding UACR 4 −4.0×10−4 1.0×10−4 0.003 −2.0×10−4 2.0×10−4 0.3 −3.0×10−4 1.0×10−4 0.003 0
Digit Symbol Coding eGFR 1 2.0×10-3 2.0×10-3 0.3 −2.0×10−3 2.0×10−3 0.5 6.0×10−4 2.0×10−3 0.7 0.27
Digit Symbol Coding eGFR 2 5.0×10−3 3.0×10-3 0.05 −4.0×10−4 2.0×10−3 0.9 2.0×10−3 2.0×10−3 0.2 0.6
Digit Symbol Coding eGFR 3 2.0×10−3 2.0×10-3 0.3 −2.0×10−3 3.0×10−3 0.5 6.0×10−4 2.0×10−3 0.7 0.25
Digit Symbol Coding eGFR 5 7.0×10−4 2.0×10-3 0.8 −2.0×10−3 3.0×10−3 0.3 −7.0×10−4 2.0×10−3 0.7 0
3MS/MMSE UACR 1 −2.0×10−4 1.0×10−4 0.2 2.0×10−5 2.0×10−4 0.9 −9.0×10−5 9.0×10−5 0.3 0
3MS/MMSE UACR 2 −2.0×10−4 2.0×10−4 0.2 4.0×10−5 2.0×10−4 0.8 −8.0×10−5 1.0×10−4 0.5 0.12
3MS/MMSE UACR 3 −2.0×10−4 1.0×10−4 0.1 3.0×10−5 2.0×10−4 0.9 −1.0×10−4 1.0×10−4 0.2 0.22
3MS/MMSE UACR 4 −2.0×10−4 1.0×10−4 0.1 3.0×10−5 2.0×10−4 0.8 −1.0×10−4 1.0×10−4 0.3 0.19
3MS/MMSE eGFR 1 4.0×10−4 2.0×10-3 0.8 −5.0×10−6 2.0×10−3 0.9 2.0×10−4 2.0×10−3 0.9 0
3MS/MMSE eGFR 2 2.0×10−3 3.0×10-3 0.5 8.0×10−4 2.0×10−3 0.7 1.0×10−3 2.0×10−3 0.5 0
3MS/MMSE eGFR 3 1.0×10−3 2.0×10-3 0.6 8.0×10−4 3.0×10−3 0.8 1.0×10−3 2.0×10−3 0.5 0
3MS/MMSE eGFR 5 4.0×10−4 2.0×10-3 0.9 6.0×10−4 3.0×10−3 0.8 5.0×10−4 2.0×10−3 0.8 0
RAVLT delayed recall UACR 1 1.0×10−4 1.0×10−4 0.4 −1.0×10−4 2.0×10−4 0.4 1.0×10−5 9.0×10−5 0.9 0.37
RAVLT delayed recall UACR 2 −1.0×10−4 2.0×10−4 0.6 −2.0×10−4 2.0×10−4 0.2 −2.0×10−4 1.0×10−4 0.2 0
RAVLT delayed recall UACR 3 6.0×10−5 1.0×10−4 0.6 −1.0×10−4 2.0×10−4 0.4 −1.0×10−5 1.0×10−4 0.9 0
RAVLT delayed recall UACR 4 5.0×10−5 1.0×10−4 0.7 −2.0×10−4 2.0×10−4 0.3 −3.0×10−5 1.0×10−4 0.8 0.07
RAVLT delayed recal eGFR 1 −7.0×10−4 2.0×10-3 0.7 −2.0×10−3 2.0×10−3 0.5 −1.0×10−3 2.0×10−3 0.5 0
RAVLT delayed recall eGFR 2 2.0×10−3 3.0×10-3 0.5 −2.0×10−3 2.0×10−3 0.5 −2.0×10−5 2.0×10−3 0.9 0.06
RAVLT delayed recall eGFR 3 −8.0×10−4 2.0×10-3 0.7 −3.0×10−3 3.0×10−3 0.3 −2.0×10−3 2.0×10−3 0.4 0
RAVLT delayed recall eGFR 5 −5.0×10−4 2.0×10-3 0.8 −4.0×10−3 3.0×10−3 0.2 −2.0×10−3 2.0×10−3 0.3 0
Stroop Interference UACR 1 2.0×10−4 1.0×10−4 0.05 2.0×10−4 2.0×10−4 0.2 2.0×10−4 9.0×10−5 0.02 0
Stroop Interference UACR 2 3.0×10−4 2.0×10−4 0.06 2.0×10−4 2.0×10−4 0.2 3.0×10−4 1.0×10−4 0.03 0
Stroop Interference UACR 3 2.0×10−4 1.0×10−4 0.05 1.0×10−4 2.0×10−4 0.5 2.0×10−4 1.0×10−4 0.04 0
Stroop Interference UACR 4 2.0×10−4 1.0×10−4 0.1 1.0×10−4 2.0×10−4 0.4 2.0×10−4 1.0×10−4 0.08 0
Stroop Interference eGFR 1 −3.0×10−3 2.0×10−3 0.1 −5.0×10−4 2.0×10−3 0.8 −2.0×10−3 2.0×10−3 0.2 0
Stroop Interference eGFR 2 −5.0×10−3 3.0×10−3 0.05 2.0×10−4 2.0×10−3 0.9 −2.0×10−3 2.0×10−3 0.2 0.54
Stroop Interference eGFR 3 −4.0×10−3 2.0×10−3 0.08 1.0×10−3 3.0×10−3 0.7 −2.0×10−3 2.0×10−3 0.3 0.53
Stroop Interference eGFR 5 −3.0×10−3 2.0×10−3 0.2 2.0×10−3 3.0×10−3 0.5 −1.0×10−3 2.0×10−3 0.6 0.46
Open in a new tab

Note: Model 1 covariates: age, sex and level of education;

Model 2 covariates: Model 1 + body mass index, hemoglobin A1c, hypertension and low-density lipoprotein -cholesterol; Model 3 covariates: Model 2 + diabetes duration, smoking status, history of cardiovascular disease; Model 4 covariates: Model 3 + eGFR; Model 5 covariates: Model 3 + UACR.

SE – standard error; 3MS, Modified Mini-Mental State Examination; eGFR, estimated glomerular filtration rate (in ml/min/1.73m2); MMSE, Mini-Mental State Examiniation; UACR, urine albumin-creatinine ratio (in mg/g); RAVLT, Rey Auditory Verbal Learning Test; AA-DHS, African American–Diabetes Heart Study; MIND, Memory in Diabetes; ACCORD, Action to Control Cardiovascular Risk in Diabetes.

Using standardized parameters obtained in the meta-analysis, effect sizes for increases of 100 mg/g in UACR and reductions of 10 ml/min/1.73m2 in eGFR were computed on cognitive performance and MRI volumes. Results were compared with the standardized effect of a year increase in age. In the fully-adjusted model, a 100 mg/g increase in UACR was equivalent to an aging effect of ~0.5 years on the Digit Symbol Coding, ~0.6 years on Stroop interference, and ~0.5 years on volume of abnormal white matter. A 10 ml/min/1.73m2 reduction in eGFR corresponded to an aging effect of ~1.3 years (Table S4).

Discussion

Independent from age, severity of diabetes (HbA1c), diabetes duration, hypertension, lipid levels, smoking or prevalent cardiovascular disease, albuminuria and reduced kidney function were each significantly associated with abnormal white matter volume (increased WMLV), a marker of cerebral microvascular disease, and lower gray matter volume in African Americans with diabetes. Findings may explain the poorer performance on Digit Symbol Coding that was observed in African Americans with kidney disease. Although diabetes and advanced kidney disease (end-stage kidney disease) are known to be associated with cognitive dysfunction and cerebral microvascular disease in European Americans and African Americans, results in African Americans with T2DM having milder degrees of kidney disease have been less clear, particularly in those with access to healthcare. These cognitive performance and cerebral MRI data in African Americans with T2DM and measures of kidney function and albuminuria encompass the largest such sample analyzed to date, to our knowledge. Diabetes, blood pressure, and lipid control were similar to those in participants from contemporary cohorts in other racial/ethnic groups, reducing the risk of bias of healthcare access on relationships.12, 33

The present findings extend those from an interim AA-DHS MIND analysis in 263 participants.13, 14 Although kidney disease associations with the brain were generally consistent in the interim study and the present analyses, the current better-powered study detected significant associations between eGFR with abnormal white matter volume (and a trend with GMV) that was not initially evident. Moreover, the previously reported association between UACR and 3MS was no longer present (an association between UACR and Digit Symbol Coding remained). In contrast to the relationships between mild kidney disease and the brain in African Americans, European Americans in DHS MIND (374 T2DM-affected European Americans with mild-to-moderate kidney disease) lacked significant association between UACR and eGFR with brain volumes and cognitive function.33 In addition, ACCORD MIND did not detect significant associations between baseline (or persistent) albuminuria with brain volumes after adjusting for age and systolic blood pressure in a predominantly non-African American T2DM-affected study population (67.7% European American, 17.3% African American, 8.8% Hispanic American, 6.2% other).15 Therefore, results from the ACCORD MIND and other published studies in European Americans with T2DM suggest that racial/ethnic-specific differences may exist in the relationships between mild kidney disease and the brain.34 These findings warrant further study in multi-ethnic cohorts or meta-analyses.

It is possible that glycemic control modulates relationships between albuminuria and cognitive performance. In 2,957 T2DM-affected ACCORD MIND participants with HbA1c values >7.5%, a UACR >30 mg/g was associated with the poorest cognitive function, specifically with reduced RAVLT and Digit Symbol Coding.11 In contrast, levels of cystatin C and eGFR were not associated with cognitive performance. In a longitudinal analysis, Barzilay et al. found that relatively young ACCORD MIND participants with persistent or progressive albuminuria showed significantly greater declines in the speed of information processing, relative to those without albuminuria; 16.2% of these participants were African American.12 This finding was evident even in those with normal eGFR. We note that blood pressures differed somewhat between reports; the average baseline systolic blood pressure in participants with albuminuria in the Barzilay et al. report was 141 mm Hg,12 contrasted with 144 mm Hg in those with kidney disease from the current study. Therefore, higher blood pressures in African Americans compared to non-African Americans could contribute to differences in the effect of albuminuria on brain structure and cognitive function.

Measures of brain structure and function have been shown to be poorer in persons with albuminuria and reduced eGFR, including in study populations with a low prevalence of diabetes.35–38 although a large European study failed to detect these associations.39 The Nurses’ Health Study demonstrated that among older women, even slightly increased levels of UACR (>5 mg/g) were associated with cognitive decline, whereas reductions in eGFR were not.40 In the Systolic Blood Pressure Intervention Trial (SPRINT), cerebral MRI and cognitive testing were performed for 218 hypertensive non-diabetic study participants with low eGFR; slightly more than 30% were African American.41 In these non-diabetic SPRINT participants, increasing albuminuria was associated with higher abnormal white matter volumes (WMLV); cognitive data were not provided. Excess albuminuria is a marker of endothelial dysfunction and related systemic microvascular disease. This can affect the coronary, neurologic, pulmonary and cerebral circulations.42–46

Although this current analysis included the largest sample of African Americans with T2DM for cognitive performance and cerebral volumes of which we are aware, limitations exist. This was a cross-sectional analysis with a single measure of UACR. As such, causation between MRI-based cerebral changes and cognitive performance could not be demonstrated. Although it is likely that most participants had access to adequate healthcare based on baseline blood pressures and clinical data, healthcare access was not directly assessed. Associations between mild kidney disease with GMV, WMLV, and Digit Symbol Coding were stronger in the AA-DHS MIND than in ACCORD MIND. ACCORD examined several interventions in a sample limited to individuals with T2DM at high risk for CVD and with a baseline HbA1C >7.5%. AA-DHS MIND was observational, recruited a broader mix of participants and was better powered to detect MRI associations than ACCORD MIND due to the larger number of participants with cerebral imaging. Hence, ACCORD entry criteria may have resulted in a more narrowly defined group. White matter hyperintensity volume is known to be a reliable and common measure of macroscopic white matter disease; however, it is not a comprehensive measure of cerebral small vessel disease nor is it a sensitive measure of more extensive microscopic white matter integrity such as those obtained from diffusion tensor imaging.47 Finally, the MMSE may provide a more limited assessment of memory functions than the 3MS.

We conclude that low level albuminuria and mildly reduced kidney function are associated with increased cerebral abnormal white matter volume and reduced gray matter volume in African Americans with T2DM. Poorer performance on Digit Symbol Coding (slower processing speed and working memory) was also present in those with higher levels of albuminuria. These results identify a subgroup of African Americans with T2DM who are at higher risk for developing reduced cognitive function and shed light on possible treatment pathways for reducing the burden of cognitive impairment–related disability in African Americans with diabetes.

Supplementary Material

1

Supplementary Table S1 (PDF). Demographic and laboratory characteristics of African American ACCORD MIND participants with MRI scan.

2

Supplementary Table S2 (PDF). Demographic and laboratory characteristics of AA-DHS MIND participants, by UACR.

3

Supplementary Table S3 (PDF). Demographic and laboratory characteristics of African American ACCORD MIND participants, by UACR.

4

Supplementary Table S4 (PDF). Comparison between association effect sizes of UACR and eGFR with effect of aging on cognitive function and MRI volumetric measures.

Acknowledgments

The authors thank all study participants and research staff at Wake Forest School of Medicine AA-DHS MIND14 and ACCORD MIND15 study sites.

Support: Support for this work was provided by National Institute of Health grants R01 DK071891, R01 NS075107, and R01 MD009055. ACCORD MIND was funded through an intra-agency agreement between the National Institute on Aging and the National Heart, Lung, and Blood Institute (AG-0002) and the National Institute on Aging Intramural Research Program. ACCORD was funded by the National Heart, Lung, and Blood Institute (N01-HC-95178, N01-HC-95179, N01-HC-95180, N01-HC-95181, N01-HC-95182, N01-HC-95183, and N01-HC-95184). Abbott Laboratories, Amylin Pharmaceuticals, AstraZeneca, Bayer HealthCare, Closer Healthcare, GlaxoSmithKline, King Pharmaceuticals, Merck, Novartis, Novo Nordisk, Omron Healthcare, Sanofi-Aventis, and Schering-Plough provided study drugs, equipment, or supplies. The funders of this project had no role in study design, collection, analysis or interpretation of the data; writing the report; or decision to submit for publication.

Financial Disclosure: The authors declare that they have no other relevant financial interests.

Contributions: Research idea and study design: BIF, JD, DWB, JAM; data acquisition: BIF, JAM, CTW, BCW, JX, SCS, KMS, RNB, LCL, LJL, JIB, RMC, MDS; data analysis/interpretation: JD, KMS, BIF, CEH, TMH, JDW, NDP, MEM, JX, LJL, JIB; statistical analysis: JD, MEM, LJL, JAM; supervision: KMS, CEH, TMH, LJL; Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved.

Peer Review: Evaluated by two external peer reviewers, a statistician, and An Acting Editor-in-Chief.

Footnotes

Supplementary Material

Table S1: Demographic and laboratory characteristics of African American ACCORD MIND participants with MRI scan.

Table S2: Demographic and laboratory characteristics of AA-DHS MIND participants, by UACR.

Table S3: Demographic and laboratory characteristics of African American ACCORD MIND participants, by UACR.

Table S4: Comparison between association effect sizes of UACR and eGFR with effect of aging on cognitive function and MRI volumetric measures.

Note: The supplementary material accompanying this article (doi:_______) is available at www.ajkd.org

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Kurella M, Chertow GM, Luan J, Yaffe K. Cognitive impairment in chronic kidney disease. J Am Geriatr Soc. 2004;52(11):1863–1869. doi: 10.1111/j.1532-5415.2004.52508.x. [DOI] [PubMed] [Google Scholar]
  • 2.Van Harten B, e Leeuw FE, einstein HC, cheltens P, iessels GJ. Brain imaging in patients with diabetes: a systematic review. Diabetes Care. 2006;29(11):2539–2548. doi: 10.2337/dc06-1637. [DOI] [PubMed] [Google Scholar]
  • 3.Weiner DE, Bartolomei K, Scott T, et al. Albuminuria, cognitive functioning, and white matter hyperintensities in homebound elders. Am J Kidney Dis. 2009;53(3):438–447. doi: 10.1053/j.ajkd.2008.08.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Schneider BC, Gross AL, Bangen KJ, et al. Association of Vascular Risk Factors With Cognition in a Multiethnic Sample. J Gerontol B Psychol Sci Soc Sci. 2014 doi: 10.1093/geronb/gbu040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Moodalbail DG, Reiser KA, Detre JA, et al. Systematic review of structural and functional neuroimaging findings in children and adults with CKD. Clin J Am Soc Nephrol. 2013;8(8):1429–1448. doi: 10.2215/CJN.11601112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Tsuruya K, Yoshida H, Kuroki Y, et al. Brain atrophy in peritoneal dialysis and CKD stages 3–5: a cross-sectional and longitudinal study. Am J Kidney Dis. 2015;65(2):312–321.19. doi: 10.1053/j.ajkd.2014.07.011. [DOI] [PubMed] [Google Scholar]
  • 7.Miwa K, Tanaka M, Okazaki S, et al. Chronic kidney disease is associated with dementia independent of cerebral small-vessel disease. Neurology. 2014;82(12):1051–1057. doi: 10.1212/WNL.0000000000000251. [DOI] [PubMed] [Google Scholar]
  • 8.Vemuri P, Knopman DS, Jack CR, et al. Association of Kidney Function Biomarkers with Brain MRI Findings: The BRINK Study. J Alzheimers Dis. 2017;55(3):1069–1082. doi: 10.3233/JAD-160834. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Seliger SL, Wendell CR, Waldstein SR, Ferrucci L, Zonderman AB. Renal function and long-term decline in cognitive function: the Baltimore Longitudinal Study of Aging. Am J Nephrol. 2015;41(4–5):305–312. doi: 10.1159/000430922. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Freedman BI, Divers J, Palmer ND. Population ancestry and genetic risk for diabetes and kidney, cardiovascular, and bone disease: modifiable environmental factors may produce the cures. Am J Kidney Dis. 2013;62(6):1165–1175. doi: 10.1053/j.ajkd.2013.05.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Murray AM, Barzilay JI, Lovato JF, et al. Biomarkers of renal function and cognitive impairment in patients with diabetes. Diabetes Care. 2011;34(8):1827–1832. doi: 10.2337/dc11-0186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Barzilay JI, Lovato JF, Murray AM, et al. Albuminuria and cognitive decline in people with diabetes and normal renal function. Clin J Am Soc Nephrol. 2013;8(11):1907–1914. doi: 10.2215/CJN.11321112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Palmer ND, Sink KM, Smith SC, et al. Kidney disease and cognitive function: african american-diabetes heart study MIND. Am J Nephrol. 2014;40(3):200–207. doi: 10.1159/000367669. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Sink KM, Divers J, Whitlow CT, et al. Cerebral structural changes in diabetic kidney disease: African American-Diabetes Heart Study MIND. Diabetes Care. 2015;38(2):206–212. doi: 10.2337/dc14-1231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Barzilay JI, Morgan TM, Murray AM, et al. Brain MRI Volume Findings in Diabetic Adults With Albuminuria: The ACCORD-MIND Study. J Gerontol A Biol Sci Med Sci. 2016;71(6):803–810. doi: 10.1093/gerona/glv187. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–612. doi: 10.7326/0003-4819-150-9-200905050-00006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Wechsler D. Manual for the Wechsler Adult Intelligence Scale-Revised. New York: Psychological Corporation; 2014. [Google Scholar]
  • 18.Houx PJ, Jolles J, Vreeling FW. Stroop interference: aging effects assessed with the Stroop Color-Word Test. Exp Aging Res. 1993;19(3):209–224. doi: 10.1080/03610739308253934. [DOI] [PubMed] [Google Scholar]
  • 19.Teng EL, Chui HC. The Modified Mini-Mental State (3MS) examination. J Clin Psychiatry. 1987;48(8):314–318. [PubMed] [Google Scholar]
  • 20.Nasreddine ZS, Phillips NA, Bedirian V, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53(4):695–699. doi: 10.1111/j.1532-5415.2005.53221.x. [DOI] [PubMed] [Google Scholar]
  • 21.Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12(3):189–198. doi: 10.1016/0022-3956(75)90026-6. [DOI] [PubMed] [Google Scholar]
  • 22.Maldjian JA, Whitlow CT, Saha BN, et al. Automated white matter total lesion volume segmentation in diabetes. AJNR Am J Neuroradiol. 2013;34(12):2265–2270. doi: 10.3174/ajnr.A3590. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Whitlow CT, Sink KM, Divers J, et al. Effects of Type 2 Diabetes on Brain Structure and Cognitive Function: African American-Diabetes Heart Study MIND. AJNR Am J Neuroradiol. 2015;36(9):1648–1653. doi: 10.3174/ajnr.A4321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Ashburner J, Friston KJ. Voxel-based morphometry--the methods. Neuroimage. 2000;11(6 Pt 1):805–821. doi: 10.1006/nimg.2000.0582. [DOI] [PubMed] [Google Scholar]
  • 25.Structural Brain Mapping Group. 2017 http://dbn,beyri,ybu-jena.de/vbm.html.
  • 26.Schmidt P, Gaser C, Arsic M, et al. An automated tool for detection of FLAIRhyperintense white-matter lesions in Multiple Sclerosis. Neuroimage. 2012;59(4):3774–3783. doi: 10.1016/j.neuroimage.2011.11.032. [DOI] [PubMed] [Google Scholar]
  • 27.Goldszal AF, Davatzikos C, Pham DL, Yan MX, Bryan RN, Resnick SM. An imageprocessing system for qualitative and quantitative volumetric analysis of brain images. J Comput Assist Tomogr. 1998;22(5):827–837. doi: 10.1097/00004728-199809000-00030. [DOI] [PubMed] [Google Scholar]
  • 28.Lao Z, Shen D, Liu D, et al. Computer-assisted segmentation of white matter lesions in 3D MR images using support vector machine. Acad Radiol. 2008;15(3):300–313. doi: 10.1016/j.acra.2007.10.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Williamson JD, Miller ME, Bryan RN, et al. The Action to Control Cardiovascular Risk in Diabetes Memory in Diabetes Study (ACCORD-MIND): rationale, design, and methods. Am J Cardiol. 2007;99(12A):112i–122i. doi: 10.1016/j.amjcard.2007.03.029. [DOI] [PubMed] [Google Scholar]
  • 30.Launer LJ, Miller ME, Williamson JD, et al. Effects of intensive glucose lowering on brain structure and function in people with type 2 diabetes (ACCORD MIND): a randomised open-label substudy. Lancet Neurol. 2011;10(11):969–977. doi: 10.1016/S1474-4422(11)70188-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Box GEP, Cox DR. An analysis of tranformations. Journal of the Royal Statistical Society, Series B. 1964;26:211–246. [Google Scholar]
  • 32.Yang Z, Wu E, Desmond AF. Inference for general parametric function in Box-Cox-type transformation models. Can J Stat. 2008;36(2):301–319. [Google Scholar]
  • 33.Murea M, Hsu FC, Cox AJ, et al. Structural and functional assessment of the brain in European Americans with mild-to-moderate kidney disease: Diabetes Heart Study-MIND. Nephrol Dial Transplant. 2015;30(8):1322–1329. doi: 10.1093/ndt/gfv030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Schneider AL, Barzilay JI. Diabetes, the brain, and cognition: More clues to the puzzle. Neurology. 2016;87(16):1640–1641. doi: 10.1212/WNL.0000000000003242. [DOI] [PubMed] [Google Scholar]
  • 35.Kuriyama N, Mizuno T, Ohshima Y, et al. Intracranial deep white matter lesions (DWLs) are associated with chronic kidney disease (CKD) and cognitive impairment: a 5-year follow-up magnetic resonance imaging (MRI) study. Arch Gerontol Geriatr. 2013;56(1):55–60. doi: 10.1016/j.archger.2011.11.009. [DOI] [PubMed] [Google Scholar]
  • 36.Sedaghat S, Cremers LG, de GM, et al. Kidney function and microstructural integrity of brain white matter. Neurology. 2015;85(2):154–161. doi: 10.1212/WNL.0000000000001741. [DOI] [PubMed] [Google Scholar]
  • 37.Ruebner RL, Laney N, Kim JY, et al. Neurocognitive Dysfunction in Children, Adolescents, and Young Adults With CKD. Am J Kidney Dis. 2016;67(4):567–575. doi: 10.1053/j.ajkd.2015.08.025. [DOI] [PubMed] [Google Scholar]
  • 38.Kurella TM, Muntner P, Wadley V, et al. Albuminuria, kidney function, and the incidence of cognitive impairment among adults in the United States. Am J Kidney Dis. 2011;58(5):756–763. doi: 10.1053/j.ajkd.2011.05.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Helmer C, Stengel B, Metzger M, et al. Chronic kidney disease, cognitive decline, and incident dementia: the 3C Study. Neurology. 2011;77(23):2043–2051. doi: 10.1212/WNL.0b013e31823b4765. [DOI] [PubMed] [Google Scholar]
  • 40.Sajjad I, Grodstein F, Kang JH, Curhan GC, Lin J. Kidney dysfunction and cognitive decline in women. Clin J Am Soc Nephrol. 2012;7(3):437–443. doi: 10.2215/CJN.05330611. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Tamura MK, Pajewski NM, Bryan RN, et al. Chronic kidney disease, cerebral blood flow, and white matter volume in hypertensive adults. Neurology. 2016 Mar 29;86(13):1208–16. doi: 10.1212/WNL.0000000000002527. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Moran A, Palmas W, Field L, et al. Cardiovascular autonomic neuropathy is associated with microalbuminuria in older patients with type 2 diabetes. Diabetes Care. 2004;27(4):972–977. doi: 10.2337/diacare.27.4.972. [DOI] [PubMed] [Google Scholar]
  • 43.von Scholten BJ, Hansen CS, Hasbak P, Kjaer A, Rossing P, Hansen TW. Cardiac Autonomic Function Is Associated With the Coronary Microcirculatory Function in Patients With Type 2 Diabetes. Diabetes. 2016;65(10):3129–3138. doi: 10.2337/db16-0437. [DOI] [PubMed] [Google Scholar]
  • 44.Harris B, Klein R, Jerosch-Herold M, et al. The association of systemic microvascular changes with lung function and lung density: a cross-sectional study. PLoS One. 2012;7(12):e50224. doi: 10.1371/journal.pone.0050224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Martens RJ, Henry RM, Houben AJ, et al. Capillary Rarefaction Associates with Albuminuria: The Maastricht Study. J Am Soc Nephrol. 2016;27(12):3748–3757. doi: 10.1681/ASN.2015111219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Kim TN, Lee EJ, Hong JW, et al. Relationship Between Sarcopenia and Albuminuria: The 2011 Korea National Health and Nutrition Examination Survey. Medicine (Baltimore) 2016;95(3):e2500. doi: 10.1097/MD.0000000000002500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Maillard P, Fletcher E, Lockhart SN, et al. White matter hyperintensities and their penumbra lie along a continuum of injury in the aging brain. Stroke. 2014;45(6):1721–1726. doi: 10.1161/STROKEAHA.113.004084. [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

1

Supplementary Table S1 (PDF). Demographic and laboratory characteristics of African American ACCORD MIND participants with MRI scan.

NIHMS882031-supplement-1.pdf (255.3KB, pdf)
2

Supplementary Table S2 (PDF). Demographic and laboratory characteristics of AA-DHS MIND participants, by UACR.

NIHMS882031-supplement-2.pdf (213.1KB, pdf)
3

Supplementary Table S3 (PDF). Demographic and laboratory characteristics of African American ACCORD MIND participants, by UACR.

NIHMS882031-supplement-3.pdf (213.7KB, pdf)
4

Supplementary Table S4 (PDF). Comparison between association effect sizes of UACR and eGFR with effect of aging on cognitive function and MRI volumetric measures.

NIHMS882031-supplement-4.pdf (216.3KB, pdf)

RESOURCES