Abstract Background Our primary goal is to describe the prevalence, severity, and risk of cognitive impairment (CI) by estimated glomerular filtration rate (eGFR, in mL/min/1.73 m2) in a cohort enriched for advanced chronic kidney disease (CKD; eGFR < 45), adjusting for albuminuria, as measured by urine albumin-to-creatinine ratio (UACR, in mg/g). As both eGFR and albuminuria are associated with CI risk in CKD, we also seek to determine the extent that eGFR remains a useful biomarker for risk of CI in those with CKD and concomitant albuminuria. Methods Chi-square tests measured the prevalence of severe CI and mild cognitive impairment (MCI) by eGFR level. Logistic regression models and generalized linear models measured risk of CI by eGFR, adjusted for UACR. Results Participants were 574 adults with a mean age of 69; 433 with CKD (eGFR < 60, nondialysis) and 141 controls (eGFR ≥ 60). Forty-eight percent of participants with CKD had severe CI or MCI. The prevalence of severe CI was highest (25%) in those with eGFR < 30. eGFR < 30 was only associated with severe CI in those without albuminuria (UACR < 30; OR = 3.3; p = .02) and was not associated with MCI in similar models. Conclusions One quarter of those with eGFR < 30 had severe CI. eGFR < 30 was associated with over threefold increased odds of severe CI in those with UACR < 30, but not with UACR > 30, suggesting that eGFR < 30 is a valid biomarker for increased risk of severe CI in those without concomitant albuminuria. Renal, Cognitive aging, Epidemiology, Dementia Chronic kidney disease (CKD) has been proposed as a model of accelerated brain aging based on the high number of individual and aggregate risk factors for cognitive impairment (CI) that are related to both renal and vascular comorbidities (1,2). There is a cross-sectional relation between estimated glomerular filtration rate (eGFR, in mL/min/1.73 m2) and cognitive function (3–6) and an association between reduced or decreasing eGFR and incident CI or dementia (7–9). Albuminuria, a systemic marker of endothelial dysfunction and microvascular disease (10,11), is also strongly associated with CI (12,13), but is less commonly measured than eGFR in older adults in routine clinical care (12). In our recent cross-sectional analyses from the prospective, longitudinal, BRain IN Kidney disease (BRINK) study, those with an eGFR < 30 performed significantly worse on individual measures of cognitive function compared to those with an eGFR ≥ 30. However, findings were not adjusted for albuminuria (3). Our primary goal is to describe the prevalence, severity, and risk of CI by eGFR level, adjusting for albuminuria, as measured by urine albumin-to-creatinine ratio (UACR, in mg/g). Given evidence that albuminuria is also associated with CI risk in CKD (12,13), we additionally seek to determine whether eGFR remains a useful biomarker for risk of CI in CKD with concomitant albuminuria. Lastly, we will characterize cognitive domain performance by CKD severity level to determine the extent that memory, executive function and language performance are impacted in advanced CKD (3,14,15). Methods Participants Study design, sample size calculations, and recruitment strategy for the BRINK study are described in detail elsewhere (3). In order to enrich our cohort with advanced CKD, participants were recruited from four community nephrology and diabetes clinics in the Minneapolis/St. Paul area. Eligibility criteria were as follows: 45 years or older, able to complete a 90-minute cognitive and physical function battery, and English as a first language. Exclusion criteria included recent psychosis, current chemical dependency, long-term high-dose narcotic use, nursing home residence, dialysis dependency, or kidney transplant recipient at the time of screening. Participants were excluded at screening if they were unable to complete the Modified Mini-Mental State Examination (3MS) (16). Definitions of categories related to indicators of CKD were based on Kidney Disease Improving Global Outcomes (KDIGO) clinical practice guidelines for eGFR only (17). Control participants were defined as having eGFR ≥ 60, although this group did include some with early CKD (ie, those with eGFRs of 60–90, with or without albuminuria). Control participants were recruited to approximate the mean age and similar racial and gender distribution of the combined CKD groups. We recruited a high proportion of participants with diabetes in our CKD cohort to reflect the approximate 50% prevalence of diabetes in the general CKD population (18). The CKD groups were defined as mild CKD (eGFR 45 to <60), moderate CKD (eGFR 30 to <45), and severe nondialysis CKD (eGFR < 30). UACR was defined as normal (UACR < 30), moderately increased (UACR 30–300) and severely increased (UACR > 300), with acknowledgment that KDIGO guidelines group normal and mildly increased albuminuria under one category (ie, UACR < 30). For this set of analyses, we use the term cohort to describe the entire BRINK sample and the term prevalence to describe the frequency of CI in the cohort. The institutional review boards of collaborating institutions approved the study (Hennepin County Medical Center approval no: 11–3393; University of Minnesota: 1203M11122; Veterans Affairs Medical Center: 4364-B; and HealthPartners Institute: A12-282). Laboratory and Other Measures Nonfasting blood samples were drawn from an antecubital vein and urine samples were obtained at the baseline visit. Baseline eGFR was calculated from the serum creatinine using the CKD Epidemiology Collaboration creatinine equation (19). Albuminuria was measured with a spot urine sample UACR. Measurement of Baseline Cognitive Function Participants were administered the following baseline cognitive measures: Hopkins Verbal Learning Test-Revised (HVLT-R) Delayed Recall (20); Brief Visuospatial Memory Test-Revised (BVMT-R) Delayed Recall (21); Symbol Digits Modalities Test (SDMT) (22); Color Trails Test 1 and 2 (CTT1 and CTT2) (23); Letter (Multilingual Aphasia Exam) (24–26) and Category Fluency (Animals) (27); and the Digit Span Total (DS) subtest of the Wechsler Adult Intelligence Scale-Third Edition (WAIS-III) (28). Performance was measured using raw and T-scores for the following individual tests: HVLT-R, BVMT-R, CTT-1, CTT-2, SDMT, and Letter and Category Fluency. Administration and norming procedures were followed per published guidelines (20–28), with the exception of the administration of Category Fluency (Animals), which was based on extending the animal naming portion of the 3MS (16) to 60 seconds. A T-score of 50 denotes normal cognition; 1 SD = 10. T-scores were based on published norms that adjust for age and sometimes education (CTT-1, CTT-2, SDMT) and race (Letter and Category Fluency). For the DS and Letter Fluency, age-adjusted scaled scores (with mean of 10 and standard deviation [SD] of 3) were converted to T-scores. Identification of Cognitive Domains Factor analysis methods were used to identify cognitive domains that best classified results from the individual cognitive tests. A three-factor model was adequately supported with the use of seven cognitive measures (Digit Span Total was eliminated): memory (HVLT-R and BVMT-R), executive function/processing speed (SDMT, CTT-1, CTT-2), and language (Letter and Category Fluency). For the purpose of our analyses, we will use the term “executive function” to describe the executive function/processing speed domain. Classifying Severity of CI and Performance by Cognitive Domain To determine CI severity, the mean of each cognitive domain was calculated and then Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) (29) cognitive performance guidelines were applied to determine whether performance in one or more domains fell within the range of Major Neurocognitive Disorder (“dementia”) or Mild Neurocognitive disorder (“MCI”). However, the DSM-5 full criteria for dementia require verification of functional decline, and study partner report of functional status was not consistently available in our cohort. Consequently, we use the term “severe cognitive impairment (severe CI)” to describe the group with performance within the dementia range as set forth by the DSM-5 (Table 1). Table 1. Classification of CI in the BRINK Cohort BRINK CI Classificationa Cognitive Criteria Normal Within 1 SD of the appropriate norm for all domains MCI 1–2 SD range below appropriate norm in one or more domains Severe CI 2 or more standard deviations below appropriate norm in one or more domains BRINK CI Classificationa Cognitive Criteria Normal Within 1 SD of the appropriate norm for all domains MCI 1–2 SD range below appropriate norm in one or more domains Severe CI 2 or more standard deviations below appropriate norm in one or more domains Note: CI = Cognitive impairment; MCI = Mild cognitive impairment. aDSM-5 cognitive criteria for Mild and Major Neurocognitive Disorder (ie, “dementia”) (29) were used to define MCI and severe CI in this cohort. View Large Statistical Analyses Baseline mean differences in continuous demographic characteristics between eGFR groups were evaluated with analysis of variance or the Kruskal–Wallis test (for skewed variables). Chi-square tests were used to evaluate differences in baseline categorical demographic characteristics by eGFR group and unadjusted associations between CI (normal, MCI or severe CI) and eGFR group (control, mild, moderate, and severe CKD). Logistic regression models were used to estimate adjusted associations between eGFR group and severity of CI (MCI and severe CI). Generalized linear models were used to test for differences in adjusted mean domain scores (LS means) between eGFR groups. All CI severity models and models for domain performance were adjusted for age, gender, race (African American or non-African American), years of education, albuminuria (UACR), medical history of TIA/stroke, diabetes, and hypertension. As eGFR and UACR are strongly correlated, we also tested for an a priori interaction between eGFR and UACR in each model. The association between eGFR or UACR and CI may be stronger among older adults (30). Subsequently, we tested for a possible interaction between age and UACR and between age and eGFR in adjusted logistic regression models for severe CI and MCI and in generalized linear models for cognitive domains. Alpha level was 0.05 for significant associations and 0.10 for significant interaction terms. SAS version 9.4 was used for all statistical analyses. Results Baseline Characteristics The cohort of 574 BRINK participants included 92 (16.0%) with mild CKD (eGFR 45 to <60), 193 (33.6%) with moderate CKD (eGFR 30 to <45), 148 (25.8%) with severe CKD (eGFR < 30), and 141(24.6%) controls (eGFR ≥ 60). Mean age was 69.2 ± 9.8 years, half were male, mean education level was 14.3 ± 2.8 years, and 20.7% classified their race as African American or other. Mean eGFR for the control group was 81.8 ± 14.8, compared with 34.3 ± 11.8 in participants with reduced eGFR. Median UACR for the control group was 3.1, compared to 52.6 for participants with reduced eGFR. The CKD groups had higher rates of diabetes and hypertension and higher mean systolic blood pressure compared with controls. There were no significant differences between eGFR groups in race distribution, history of prior TIA/stroke, diastolic blood pressure, or apolipoprotein E (APOE) ε4 allele frequency (Table 2). Table 2. Demographic and Other Characteristics of BRINK Baseline Cohort eGFR in mL/min/1.73 m2 Characteristics Sample (N = 574) <30 (n = 148) 30 to <45 (n = 193) 45 to <60 (n = 92) ≥60 (n = 141) p Value Age in years, mean ± SD 69.2 ± 9.8 68.9 ± 10.4 70.3 ± 9.6 70.3 ± 9.7 67.2 ± 9.3 .02a Age category, n (%) .008b 45–54 y 36 (6.3) 15 (10.1) 6 (3.1) 2 (2.2) 13 (9.2) 55–64 y 152 (26.5) 38 (25.7) 56 (29.0) 23 (25.0) 35 (24.8) 65–74 y 213 (37.1) 47 (31.8) 63 (32.6) 41 (44.6) 62 (44.0) ≥75 y 173 (30.1) 48 (32.4) 68 (35.2) 26 (28.3) 31 (22.0) Male, n (%) 286 (49.8) 88 (59.5) 91 (47.2) 47 (51.1) 60 (42.6) .03b Race, n (%) .09b,c African American 91 (15.8) 33 (22.3) 20 (10.4) 15 (16.3) 23 (16.3) White 455 (79.3) 111 (75.0) 163 (84.5) 72 (78.3) 109 (77.3) Other 28 (4.9) 4 (2.7) 10 (5.2) 5 (5.4) 9 (6.4) Years of education, mean ± SD 14.3 ± 2.8 13.5 ± 2.7 14.1 ± 2.6 14.4 ± 3.2 15.2 ± 2.5 <.001a eGFR, mL/min/1.73 m2, mean ± SDd 46.0 ± 24.0 21.3 ± 6.0 36.6 ± 3.9 50.8 ± 4.0 81.8 ± 14.8 <.001a Urine ACR, mg/g, mediane (Q1, Q3) 28.1 (0, 192.7) 215.2 (34.8, 1337.3) 36.3 (2.5, 197.6) 17.5 (0, 79.8) 3.1 (0, 12.3) <.001f Diabetes, n (%) 287 (50.0) 89 (60.1) 88 (45.6) 51 (55.4) 59 (41.8) .006b Hypertension, n (%) 520 (90.6) 145 (98.0) 185 (95.9) 86 (93.5) 104 (73.8) <.001b Prior TIA/stroke, n (%) 95 (16.6) 28 (18.9) 36 (18.7) 18 (19.6) 13 (9.2) .06b Systolic BP, mmHg, mean ± SD 132.2 ± 18.8 136.4 ± 21.1 133.0 ± 18.4 130.3 ± 16.7 127.8 ± 17.0 .001a Diastolic BP, mmHg, mean ± SD 68.9 ± 12.0 68.3 ± 14.0 68.6 ± 11.5 69.6 ± 12.1 69.2 ± 10.6 .83a At least one APOE ε4 allele, n (%)g 160 (28.4) 40 (27.6) 60 (32.1) 21 (23.1) 39 (27.9) .46b eGFR in mL/min/1.73 m2 Characteristics Sample (N = 574) <30 (n = 148) 30 to <45 (n = 193) 45 to <60 (n = 92) ≥60 (n = 141) p Value Age in years, mean ± SD 69.2 ± 9.8 68.9 ± 10.4 70.3 ± 9.6 70.3 ± 9.7 67.2 ± 9.3 .02a Age category, n (%) .008b 45–54 y 36 (6.3) 15 (10.1) 6 (3.1) 2 (2.2) 13 (9.2) 55–64 y 152 (26.5) 38 (25.7) 56 (29.0) 23 (25.0) 35 (24.8) 65–74 y 213 (37.1) 47 (31.8) 63 (32.6) 41 (44.6) 62 (44.0) ≥75 y 173 (30.1) 48 (32.4) 68 (35.2) 26 (28.3) 31 (22.0) Male, n (%) 286 (49.8) 88 (59.5) 91 (47.2) 47 (51.1) 60 (42.6) .03b Race, n (%) .09b,c African American 91 (15.8) 33 (22.3) 20 (10.4) 15 (16.3) 23 (16.3) White 455 (79.3) 111 (75.0) 163 (84.5) 72 (78.3) 109 (77.3) Other 28 (4.9) 4 (2.7) 10 (5.2) 5 (5.4) 9 (6.4) Years of education, mean ± SD 14.3 ± 2.8 13.5 ± 2.7 14.1 ± 2.6 14.4 ± 3.2 15.2 ± 2.5 <.001a eGFR, mL/min/1.73 m2, mean ± SDd 46.0 ± 24.0 21.3 ± 6.0 36.6 ± 3.9 50.8 ± 4.0 81.8 ± 14.8 <.001a Urine ACR, mg/g, mediane (Q1, Q3) 28.1 (0, 192.7) 215.2 (34.8, 1337.3) 36.3 (2.5, 197.6) 17.5 (0, 79.8) 3.1 (0, 12.3) <.001f Diabetes, n (%) 287 (50.0) 89 (60.1) 88 (45.6) 51 (55.4) 59 (41.8) .006b Hypertension, n (%) 520 (90.6) 145 (98.0) 185 (95.9) 86 (93.5) 104 (73.8) <.001b Prior TIA/stroke, n (%) 95 (16.6) 28 (18.9) 36 (18.7) 18 (19.6) 13 (9.2) .06b Systolic BP, mmHg, mean ± SD 132.2 ± 18.8 136.4 ± 21.1 133.0 ± 18.4 130.3 ± 16.7 127.8 ± 17.0 .001a Diastolic BP, mmHg, mean ± SD 68.9 ± 12.0 68.3 ± 14.0 68.6 ± 11.5 69.6 ± 12.1 69.2 ± 10.6 .83a At least one APOE ε4 allele, n (%)g 160 (28.4) 40 (27.6) 60 (32.1) 21 (23.1) 39 (27.9) .46b Note: ACR = Albumin-creatinine ratio; APOE = Apolipoprotein E; BP = Blood pressure; eGFR = Estimated glomerular filtration rate, CKD-Epi formula; SD = Standard deviation; Q1 = 25th percentile; Q3 = 75th percentile. ap-value from ANOVA F-test. bp-value from chi-square test. cGroups categorized as white versus non-white for chi-square test. dMean eGFR for the control group was 81.8 ± 14.8, compared with 34.3 ± 11.8 in participants with reduced eGFR. eMedian UACR for the control group was 3.1, compared to 52.6 for participants with reduced eGFR. fp-value from Kruskal–Wallis test. gN = 563 with genotyping. View Large Baseline Prevalence of CI At baseline 48% of those with CKD (eGFR < 60) and 41% of controls (eGFR ≥ 60) were characterized as severe CI or MCI. The highest prevalence of severe CI was in those with severe kidney disease (eGFR < 30) at 25.0%, compared with 15.6% in the control group. Prevalence of MCI was highest (31.8%) in those with mild kidney disease (eGFR 45 to <60). Unadjusted chi square analysis revealed that eGFR was significantly associated with CI level (p = .03; Table 3). Additional unadjusted chi-square analyses investigated the association between MCI or severe CI and eGFR using three different eGFR cutpoints: <30 vs ≥30, <45 vs ≥45, and <60 vs ≥60. The only significant association was between eGFR with cutpoint at 30 and severe CI (25.0% for eGFR < 30 vs 13.9% for eGFR ≥ 30, p = .002). There were no significant associations between eGFR cutpoints and MCI. Table 3. Prevalence of CI Severity by eGFR Groupa eGFR (mL/min/1.73 m2) N Normal n = 308 MCI n = 170 Severe CI n = 96 <30 148 64 (43.2%) 47 (31.8%) 37 (25.0%) 30 - <45 193 111 (57.5%) 56 (29.0%) 26 (13.5%) 45 - <60 92 50 (54.4%) 31 (33.7%) 11 (12.0%) ≥60 141 83 (58.9%) 36 (25.5%) 22 (15.6%) eGFR (mL/min/1.73 m2) N Normal n = 308 MCI n = 170 Severe CI n = 96 <30 148 64 (43.2%) 47 (31.8%) 37 (25.0%) 30 - <45 193 111 (57.5%) 56 (29.0%) 26 (13.5%) 45 - <60 92 50 (54.4%) 31 (33.7%) 11 (12.0%) ≥60 141 83 (58.9%) 36 (25.5%) 22 (15.6%) Note: CI = Cognitive impairment; eGFR = Estimated glomerular filtration rate; MCI = Mild cognitive impairment aChi-square test of association p = .03. View Large In a logistic regression model for severe CI adjusted for age, education, gender, race, albuminuria, TIA/stroke, diabetes, and hypertension, there was a significant interaction between eGFR (<30 vs ≥30) and UACR (p = .05). Stratified models to examine this interaction revealed that eGFR (<30 vs ≥30) was significantly associated with severe CI in those with UACR <30 (OR = 3.3, 95% CI = 1.27, 8.50, p = .02); this association was no longer significant among those with a UACR 30–300 or UACR > 300 (Table 4). There were no significant differences in the prevalence of severe CI between UACR categories (p = .13; Table 4). Table 4. Adjusted Odds of Severe CI by UACR Status in Severe CKDa UACR Category N (%) with Severe CIb Effect OR Estimate 95% Wald Confidence Limits p Value UACR <30 (n = 291) 39 (13.4%) eGFR: <30 vs ≥30 3.28 1.27, 8.50 .02 UACR 30–300 (n = 159) 33 (20.8%) eGFR: <30 vs ≥30 2.22 0.90, 5.48 .08 UACR >300 (n = 119) 20 (16.8%) eGFR: <30 vs ≥30 0.76 0.27, 2.14 .60 UACR Category N (%) with Severe CIb Effect OR Estimate 95% Wald Confidence Limits p Value UACR <30 (n = 291) 39 (13.4%) eGFR: <30 vs ≥30 3.28 1.27, 8.50 .02 UACR 30–300 (n = 159) 33 (20.8%) eGFR: <30 vs ≥30 2.22 0.90, 5.48 .08 UACR >300 (n = 119) 20 (16.8%) eGFR: <30 vs ≥30 0.76 0.27, 2.14 .60 Note: CI = Cognitive impairment; eGFR = Estimated glomerular filtration rate; OR = Odds ratio; UACR = Urine albumin-to-creatinine ratio. aAll results adjusted for age, education, gender, race, history of TIA/stroke, diabetes and hypertension. bThe chi-square test for prevalence of CI by UACR group was not significant (p = .13). View Large Secondary Analyses Related to African American Race and Age The relation between eGFR group and CI level was not consistently linear (Table 3). For example, while the highest prevalence of severe CI was in the eGFR < 30 group (25.0%), the prevalence of severe CI was higher in the eGFR control group (15.6%) than the eGFR 45 to <60 (12%) and 30 to <45 groups (13.5%). We also noted a nonlinear trend in the distribution of African American race across groups (Table 2), with a higher percentage (16.3%) in the control group relative to those with mild to moderate kidney disease (eGFR 30 to <45; 10.4%), but lower relative to those with severe CKD (eGFR <30; 22.3%). African American race is associated with higher risk for diabetes, hypertension CKD, and CI (31,32). Therefore, we explored whether African American race may be contributing to this nonlinear pattern of severe CI results. Upon collapsing the mild and moderate CKD groups into one “mild to moderate” category (eGFR 30 to <60), chi-square analyses revealed that there was a significant difference in the distribution of African Americans across the three eGFR groups: 22.3% in eGFR < 30, 12.3% in eGFR 30 to <60 and 16.3% in eGFR ≥ 60 (p = .03). Furthermore, in the BRINK cohort, African Americans, relative to non-African Americans, had significantly higher rates of diabetes (69.2% vs 46.4%; p < .001) and TIA/stroke (24.2% vs 15.1%; p = .03) but not hypertension (p = 0.32). African Americans had less years of education on average (12.5 [2.9] vs 14.6 [2.7]; p < .001). All models for severe CI and MCI were adjusted for AA race, diabetes, stroke/TIA, age, and education to account for these multiple and potentially confounded contributors to CI in the BRINK cohort. Lastly, in fully-adjusted logistic regression models, we found no significant interaction between age group and eGFR (p = .33) or UACR (p = .60) on severe CI. Baseline Cognitive Domain Scores Performance in each cognitive domain varied by eGFR group (<30 vs ≥30), as measured by least squares (LS) mean T scores adjusted for age, education, gender, race, albuminuria (UACR), TIA/stroke, diabetes, hypertension and, if present, the interaction between eGFR and UACR, age and eGFR, or age and UACR. In adjusted regression models, there was an interaction noted between eGFR and UACR on both memory (p = .03) and executive function performance (p = .005). Stratified models demonstrated that eGFR < 30 was significantly associated with lower performance in memory (p = .02) and executive function (p = .004) in those with normal UACR (<30) and those with UACR 30–300 (p = .02 for memory; p = .02 for executive function), but eGFR was not associated with lower memory (p = .54) or executive function performance (p = .21) in those with UACR > 300 (Table 5). In the adjusted regression model for language, eGFR < 30 was associated with reduced performance (p = .04); there was no significant interaction between eGFR and UACR (p = .17). There was a significant interaction between age and UACR on language performance (p = .02). In those aged 70 years or older, there was a lower LS mean language domain score for those with UACR 30–300 compared to UACR < 30 but these findings were only marginally significant (p = .05) There was no significant association between eGFR <45 vs ≥45 and memory (p = .28), executive function (p = .35), or language performance (p = .57) or between eGFR <60 vs ≥ 60 and memory (p =.74), executive function (p = .39), or language performance (p = 0.32). Table 5. Adjusted Mean Difference in Cognitive Domain Scores for eGFR < 30 vs ≥30 Stratified by Albuminuria Statusa UACR ≤ 30 n = 289/291c UACR 30–300 n = 156 UACR ≥ 300 n = 119 Cognitive Domainb Estimated (SE) p Value Estimated (SE) p Value Estimated (SE) p Value Memory −4.37 (1.87) .02 −4.35 (1.85) .02 1.15 (1.88) .54 Executive Function −5.04 (1.74) .004 −4.02 (1.74) .02 1.78 (1.42) .21 UACR ≤ 30 n = 289/291c UACR 30–300 n = 156 UACR ≥ 300 n = 119 Cognitive Domainb Estimated (SE) p Value Estimated (SE) p Value Estimated (SE) p Value Memory −4.37 (1.87) .02 −4.35 (1.85) .02 1.15 (1.88) .54 Executive Function −5.04 (1.74) .004 −4.02 (1.74) .02 1.78 (1.42) .21 Note: eGFR = Estimated glomerular filtration rate; SE = Statndard error; UACR = Urine albumin-to-creatinine ratio aThe Language Estimated (SE) was −1.81 (0.88); p = .04. The interaction between eGFR and UACR was not significant for this domain (p = .17), so models stratified by UACR group were not run. bAll results adjusted for age, education, gender, African American race, history of TIA / stroke, diabetes, and hypertension. cN = 289 for memory domain and 291 for the executive domain due to missing memory domain scores. dDifference in adjusted mean domain scores (LS mean) between eGFR < 30 and eGFR ≥ 30. View Large Discussion In this community-based clinic cohort enriched for advanced CKD, we found that 48% of those with CKD (eGFR < 60) had some level of CI at study baseline. Furthermore, one-quarter of those with severe CKD (eGFR < 30) demonstrated severe CI. For those without albuminuria (UACR < 30), the odds of severe CI were over three times higher for eGFR < 30 compared with eGFR ≥ 30. However, as UACR increased, eGFR < 30 was no longer associated with the odds of severe CI. Overall, these findings suggest that eGFR < 30 is a useful biomarker for increased risk of severe CI in those without concomitant albuminuria. We are not aware of previous studies that have measured the burden or risk of CI by eGFR level within an enriched CKD cohort with a comprehensive battery of neuropsychological measures, while adjusting for concomitant albuminuria. A recent cross-sectional analysis of a large community sample in the Netherlands (the Maastricht Study) revealed that both eGFR and albuminuria were individually associated with cognitive function but this relation was limited to older adults (30). However, the Maastricht study was not enriched for advanced CKD and included less than 4% with an eGFR < 60 (33), compared with 75% of participants in the BRINK cohort. We anticipate that quantifying the prevalence of severe CI in CKD by eGFR level will increase the awareness of this burden among clinicians. Our results extend our previous work with individual cognitive measures (3) by clarifying that eGFR < 30 is associated with risk for severe CI in those without moderate to severely increased albuminuria (UACR ≥ 30). Our findings complement previous findings from the Reasons for Geographic and Racial Disparities in Stroke (REGARDS) study which utilized a brief screening measure to identify CI (34). REGARDS results demonstrated that when stratified by UACR, reduced eGFR defined as eGFR < 60 was associated with higher odds for CI (OR =1.30), but solely at lower UACR levels (defined as UACR < 10) compared to other increased UACR groups. Our results also support their conclusion that albuminuria and low eGFR are complementary, but not additive risk factors for CI. The lack of association between eGFR and CI in patients with moderate to severe albuminuria may be due to a number of factors. Albuminuria is associated with vascular endothelial function, which in the cerebral vasculature may result in breakdown of the blood brain barrier; this may result in cognitive changes associated with cerebrovascular and white matter changes seen on brain MRI in CKD (10). It may be that UACR is more strongly associated with these neuropathological factors than eGFR, whereas an elevated eGFR may be more closely tied to neuronal toxins seen in uremia (35). However, it remains the case that eGFR is often measured in routine clinical care, whereas albuminuria is not, except in those with diabetes or severe hypertension (12). In terms of cognitive domain performance, eGFR < 30 was associated with lower performance in the cognitive domains of memory and executive function in those without severely increased UACR (>300); eGFR < 30 was associated with reduced performance in language, independent of UACR levels. These findings support previous work that advanced CKD is associated with reduction in memory, executive function, and language performance (3,14,15). However, we did not find an association between reduced eGFR and MCI. There are a limited number of studies in the CKD literature addressing MCI risk in CKD. It has been suggested that MCI may be more closely associated with an Alzheimer’s disease/neurodegenerative process compared to the vascular processes that are hypothesized to contribute to CI in CKD (36). The Einstein Aging Study, demonstrated that low eGFR (<45), compared to high eGFR (≥60) was associated with both amnestic and nonamnestic MCI subtypes in adjusted analyses (14). Our results may differ from these findings due to our use of more liberal DSM-5 clinical criteria to define MCI (ie, impairment defined as 1–2 SD range below appropriate norms) versus the Petersen research criteria of 1.5 to 2 SD below appropriate norms (37). Lastly, the sample size for our MCI models was reduced to only include those with MCI or normal cognitive function, thereby potentially limiting power to detect a significant association between eGFR and MCI. We sought to recruit controls that approximated the distributions of key demographic variables in our CKD group, resulting in a similar percentage of African Americans in our control and CKD group. African Americans also had significantly higher rates of diabetes and TIA/stroke, and fewer years of education. Taken together, these factors likely contributed to the higher than expected rates of severe CI among controls in the BRINK cohort. Findings from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) MIND study demonstrated that baseline or persisting albuminuria (corrected for eGFR) were not associated with MRI measures associated with cognitive decline (38). However, in related analyses of African American ACCORD MIND and African American-Diabetes Heart Study MIND (AA-DHS MIND) participants, even mild kidney dysfunction was associated with reduced performance in screening measures for CI, processing speed, executive function (39), as well as reduced grey matter volume and higher white matter lesion volume (40). These findings underscore the importance of pursuing further investigation of the complex relationships between the multiple factors associated with CI in CKD populations, especially among underrepresented groups (39,40). Strengths of our study include utilizing the DSM-5 cognitive criteria to define severe CI and MCI in a well-characterized cohort enriched with advanced CKD patients and measuring performance by cognitive domain based on clinically relevant neuropsychological measures. By characterizing the overall burden and severity of CI, we found that 48% of CKD participants had some level of CI at study baseline, further raising the awareness of the burden of CI in CKD. In addition, we measured the effect of eGFR adjusted for albuminuria in our analyses of risk for CI, thereby providing specific guidance for clinicians as regarding risk of CI in advanced CKD patients without albuminuria. Several study limitations deserve mention. As this was a cross-sectional analysis, causal inferences cannot be made regarding eGFR, albuminuria, and CI. The original study was not designed to subtype dementia (ie, vascular dementia vs Alzheimer’s disease); therefore, we are unable to hypothesize the etiology of severe CI this cohort. We used the DSM-5 cognitive criteria for MCI and severe CI, but did not have comprehensive informant based functional status information to meet the full criteria for dementia. However, we believe that classification of severe CI is sufficient to inform providers and families with guidance in the clinical management of CKD patients. Our findings may improve CKD patient care by identifying those at highest risk for severe CI and improving awareness of its relatively high prevalence in advanced CKD. Our results suggest that cognitive screening or referral for more detailed neuropsychological assessment is warranted for patients with eGFR < 30, especially if albuminuria is not suspected. Future BRINK analyses will characterize the trajectory and characteristics of cognitive decline to enable potential identification of mechanisms and modifiable risk factors for CI in this unique and complex population. Funding This work was supported by National Institute on Aging Grant # R01 AG03755, Satellite Health Inc., the Minneapolis Medical Research Foundation and grant number UL1TR000114 from the National Center for Research Resources, National Institutes of Health. Conflict of Interest None reported. Acknowledgments The authors would like to thank the BRINK study participants. The authors acknowledge Liz Amiot, Rachel Vasseur, and Ashley Johnson for their dedication as research coordinators for this study. The authors also thank Molly Prozinski for editing and formatting the manuscript for nontechnical content. References 1. Murray AM. The brain and the kidney connection: a model of accelerated vascular cognitive impairment. Neurology . 2009; 73: 916– 917. doi: 10.1212/WNL.0b013e3181b99a2e Google Scholar CrossRef Search ADS PubMed 2. Shuchi A, Johansen KL, Kurella Tamura M. Aging and chronic kidney disease: the impact on physical function and cognition. J Gerontol A Biol Sci Med Sci . 2014; 69A: 315– 322. doi: 10.1093/gerona/glt109 Google Scholar CrossRef Search ADS 3. Murray AM, Bell EJ, Tupper DEet al. The Brain in Kidney Disease (BRINK) cohort study: design and baseline cognitive function. Am J Kidney Dis . 2016; 67: 593– 600. doi: 10.1053/j.ajkd.2015.11.008 Google Scholar CrossRef Search ADS PubMed 4. Kurella Tamura M, Muntner P, Wadley Vet al. Albuminuria, kidney function, and the incidence of cognitive impairment among adults in the United States. Am J Kidney Dis . 2011; 58: 756– 763. doi: 10.1053/j.ajkd.2011.05.027 Google Scholar CrossRef Search ADS PubMed 5. Yaffe K, Ackerson L, Kurella Tamura Met al. ; Chronic Renal Insufficiency Cohort Investigators. Chronic kidney disease and cognitive function in older adults: findings from the chronic renal insufficiency cohort cognitive study. J Am Geriatr Soc . 2010; 58: 338– 345. doi: 10.1111/j.1532-5415.2009.02670.x Google Scholar CrossRef Search ADS PubMed 6. Torres RV, Elias MF, Seliger S, Davey A, Robbins MA. Risk for cognitive impairment across 22 measures of cognitive ability in early-stage chronic kidney disease. Nephrol Dial Transplant . 2016;32:299–306. doi: 10.1093/ndt/gfw005 7. Buchman AS, Tanne D, Boyle PA, Shah RC, Leurgans SE, Bennett DA. Kidney function is associated with the rate of cognitive decline in the elderly. Neurology . 2009; 73: 920– 927. doi: 10.1212/WNL.0b013e3181b72629 Google Scholar CrossRef Search ADS PubMed 8. Davey A, Elias MF, Robbins MA, Seliger SL, Dore GA. Decline in renal functioning is associated with longitudinal decline in global cognitive functioning, abstract reasoning and verbal memory. Nephrol Dial Transplant . 2013; 28: 1810– 1819. doi: 10.1093/ndt/gfs470 Google Scholar CrossRef Search ADS PubMed 9. Etgen T, Sander D, Chonchol Met al. Chronic kidney disease is associated with incident cognitive impairment in the elderly: the INVADE study. Nephrol Dial Transplant . 2009; 24: 3144– 3150. doi: 10.1093/ndt/gfp230 Google Scholar CrossRef Search ADS PubMed 10. Knopman DS. Invited commentary: albuminuria and microvascular disease of the brain–a shared pathophysiology. Am J Epidemiol . 2010; 171: 287– 9; author reply 290. doi: 10.1093/aje/kwp429 Google Scholar CrossRef Search ADS PubMed 11. Seliger SL, Salimi S, Pierre V, Giffuni J, Katzel L, Parsa A. Microvascular endothelial dysfunction is associated with albuminuria and CKD in older adults. BMC Nephrol . 2016; 17: 82. doi: 10.1186/s12882-016-0303-x Google Scholar CrossRef Search ADS PubMed 12. Georgakis MK, Dimitriou NG, Karalexi MAet al. Albuminuria in association with cognitive function and dementia: a systematic review and meta-analysis. J Am Geriatr Soc . 2017; 65: 1190– 1198. doi: 10.1111/jgs.14750 Google Scholar CrossRef Search ADS PubMed 13. Deckers K, Camerino I, van Boxtel MPet al. Dementia risk in renal dysfunction: a systematic review and meta-analysis of prospective studies. Neurology . 2017; 88: 198– 208. doi: 10.1212/WNL.0000000000003482 Google Scholar CrossRef Search ADS PubMed 14. Zammit AR, Katz MJ, Zimmerman ME, Bitzer M, Lipton RB. Low eGFR is associated with dysexecutive and amnestic mild cognitive impairment. Alzheimers Dement (Amst) . 2015; 1: 152– 159. doi: 10.1016/j.dadm.2014.12.002 Google Scholar PubMed 15. Kurella M, Yaffe K, Shlipak MG, Wenger NK, Chertow GM. Chronic kidney disease and cognitive impairment in menopausal women. Am J Kidney Dis . 2005; 45: 66– 76. doi:10.1053/j.ajkd.2004.08.044 Google Scholar CrossRef Search ADS PubMed 16. Teng EL, Chui HC. The Modified Mini-Mental State (3MS) examination. J Clin Psychiatry . 1987; 48: 314– 318. doi:10.1007/springerreference_183334 Google Scholar PubMed 17. KDIGO CKD-MBD Work Group. KDIGO clinical practice guideline for the diagnosis, evaluation, prevention, and treatment of Chronic Kidney Disease-Mineral and Bone Disorder (CKD-MBD). Kidney Int Suppl . 2009;113: S1– 130. doi: 10.1038/ki.2009.188 18. Tuttle KR, Bakris GL, Bilous RWet al. Diabetic kidney disease: a report from an ADA Consensus Conference. Am J Kidney Dis . 2014; 64: 510– 533. doi: 10.1053/j.ajkd.2014.08.001 Google Scholar CrossRef Search ADS PubMed 19. Levey AS, Stevens LA, Schmid CHet al. ; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med . 2009; 150: 604– 612. doi:10.7326/0003-4819-150-9-200905050-00006 Google Scholar CrossRef Search ADS PubMed 20. Brandt J. Hopkins Verbal Learning Test Revised. Professional Manual . Lutz, FL: Psychological Assessment Resources, Inc.; 2001. 21. Benedict RHB. Brief Visuospatial Memory Test-Revised: Professional Manual . Lutz, FL: Psychological Assessment Resources, Inc.; 1997. 22. Smith A. Symbol Digits Modalities Test. Professional Manual . Los Angeles, CA: Western Psychological Services; 1982. 23. D’Elia LF, Satz P, Uchiyama CL, White T. Color Trails Test (CTT) . Lutz, FL: Psychological Assessment Resources, Inc.; 1994. 24. Benton AL, Hamsher K, Sivan AB. Multilingual Aphasia Examination, Third Edition. Manual of Instructions . 3rd ed. Lutz, FL: PAR; 1978. 25. Lucas JA, Ivnik RJ, Smith GEet al. Mayo’s Older African Americans Normative Studies: norms for Boston Naming Test, Controlled Oral Word Association, Category Fluency, Animal Naming, Token Test, WRAT-3 Reading, Trail Making Test, Stroop Test, and Judgment of Line Orientation. Clin Neuropsychol . 2005; 19: 243– 269. doi: 10.1080/13854040590945337 Google Scholar CrossRef Search ADS PubMed 26. Ivnik RJ, Malec JF, Smith GE, Tangalos EG, Petersen RC. Neuropsychological tests’ norms above age 55: COWAT, BNT, MAE Token, WRAT-R Reading, AMNART, Stroop, TMT, and JLO. Clin Neuropsychol . 1996; 10: 262– 278. doi:10.1080/13854049608406689 Google Scholar CrossRef Search ADS 27. Gladsjo JA, Schuman CC, Evans JD, Peavy GM, Miller SW, Heaton RK. Norms for letter and category fluency: demographic corrections for age, education, and ethnicity. Assessment . 1999; 6: 147– 178. doi: 10.1177/107319119900600204 Google Scholar CrossRef Search ADS PubMed 28. Wechsler D. Manual for the Wechsler Adult Intelligence Scale-Revised . San Antonio, TX: The Psychological Corporation; 1981. 29. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders: DSM-5 . 5th ed. Arlington, VA: American Psychiatric Association; 2013. 30. Martens RJ, Kooman JP, Stehouwer CDet al. Estimated GFR, Albuminuria, and Cognitive Performance: the Maastricht Study. Am J Kidney Dis . 2017; 69: 179– 191. doi: 10.1053/j.ajkd.2016.04.017 Google Scholar CrossRef Search ADS PubMed 31. Chin AL, Negash S, Hamilton R. Diversity and disparity in dementia: the impact of ethnoracial differences in Alzheimer disease. Alzheimer Dis Assoc Disord . 2011; 25: 187– 195. doi: 10.1097/WAD.0b013e318211c6c9 Google Scholar CrossRef Search ADS PubMed 32. United States Renal Data System. 2016 USRDS Annual Data Report: Epidemiology of Kidney Disease in the United States . Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 2016. PubMed PubMed 33. Elias MF, Torres RV, Davey A. Albuminuria and cognitive performance: new evidence for consideration of a risk factor precursor model from the Maastricht study. Am J Kidney Dis . 2017; 69: 163– 165. doi: 10.1053/j.ajkd.2016.11.002 Google Scholar CrossRef Search ADS PubMed 34. Kurella Tamura M, Wadley V, Yaffe Ket al. Kidney function and cognitive impairment in US adults: the Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study. Am J Kidney Dis . 2008; 52: 227– 234. doi: 10.1053/j.ajkd.2008.05.004 Google Scholar CrossRef Search ADS PubMed 35. Vemuri P, Knopman DS, Jack CRet al. Association of kidney function biomarkers with brain MRI findings: the BRINK study. J Alzheimers Dis . 2017; 55: 1069– 1082. doi: 10.3233/JAD-160834 Google Scholar CrossRef Search ADS PubMed 36. Barzilay JI, Fitzpatrick AL, Luchsinger Jet al. Albuminuria and dementia in the elderly: a community study. Am J Kidney Dis . 2008; 52: 216– 226. doi: 10.1053/j.ajkd.2007.12.044 Google Scholar CrossRef Search ADS PubMed 37. Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome. Arch Neurol . 1999; 56: 303– 308. doi:10.1001/archneur.56.3.303 Google Scholar CrossRef Search ADS PubMed 38. Barzilay JI, Morgan TM, Murray AMet al. Brain MRI volume findings in diabetic adults with albuminuria: the ACCORD-MIND study. J Gerontol A Biol Sci Med Sci . 2016; 71: 803– 810. doi: 10.1093/gerona/glv187 Google Scholar CrossRef Search ADS PubMed 39. Palmer ND, Sink KM, Smith SCet al. Kidney disease and cognitive function: African American-diabetes heart study MIND. Am J Nephrol . 2014; 40: 200– 207. doi: 10.1159/000367669 Google Scholar CrossRef Search ADS PubMed 40. Freedman BI, Sink KM, Hugenschmidt CEet al. ; AA-DHS MIND and ACCORD MIND Investigators. Associations of early kidney disease with brain magnetic resonance imaging and cognitive function in African Americans with type 2 diabetes mellitus. Am J Kidney Dis . 2017; 70: 627– 637. doi: 10.1053/j.ajkd.2017.05.006 Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: email@example.com.
The Journals of Gerontology Series A: Biomedical Sciences and Medical Sciences – Oxford University Press
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera