CKD is a recognized risk factor for cognitive impairment (CI) (1–4), but reports of associations of both eGFR and albuminuria with CI and incident dementia in healthier, older populations without a history of cardiovascular disease (CVD) are limited. Decreased eGFR represents a decline in kidney function due to vascular and nonvascular etiologies and aging. Albuminuria is considered a measure of kidney damage, or glomerular angiopathy secondary to microvascular endothelial inflammation, and frequently precedes a decline in eGFR (5). Several studies have described the association between albuminuria and white matter hyperintensities in the brain—a measure of cerebral microvascular disease that is also associated with CI (6).
Vascular risk factors and associated pathology play large roles in the development of CKD and associated cerebrovascular disease in older populations in the United States, because over half of CKD cases are secondary to diabetes, hypertension, or both (7). However, in a population free of known CVD at baseline, the usual role of shared CVD risk factors for CKD and cerebrovascular disease in the development of CI in CKD may be weaker than in the general population. Thus, it is possible that previous estimates of the effects of CKD on cognitive function conducted in populations with high rates of CVD may not be generalizable to healthier populations.
It is important to determine to what extent eGFR and albuminuria can be used to identify those most at risk of CI in the large, and growing, portion of the older population without substantial underlying CVD, and, in turn, reduce related complications, such as medication and kidney disease management nonadherence and downstream hospitalizations. To do so, we conducted a longitudinal cohort analysis in the study population of the ASPirin in Reducing Events in the Elderly (ASPREE) trial, which was designed to enroll healthy, community-dwelling participants, free of cardiovascular events. Our goals were to estimate (1) the cross-sectional association between kidney function, measured by both eGFR and albuminuria (using the urine albumin-creatinine ratio [UACR]), and cognitive function at the baseline ASPREE assessment; and (2) the longitudinal association between baseline kidney function and change in cognitive function, incident CI no dementia (CIND), and incident dementia. We hypothesized that, in older individuals without a history of CVD or severe CKD, albuminuria would more likely be more associated than eGFR with CI, cognitive decline, and incident dementia.
Materials and Methods
Study Design and Population
We conducted a longitudinal cohort study using the ASPREE trial population in Australia and the United States. The ASPREE trial was a randomized, placebo-controlled trial of low-dose aspirin in healthy older individuals with mean 4.7 years of follow-up. We previously reported that aspirin did not reduce risk of incident dementia, mild CI, or cognitive decline (8). Full details regarding the rationale and study design for the ASPREE trial, including detailed inclusion and exclusion criteria, have been reported previously (Supplemental Appendix 1) (9).
Eligibility criteria included being free from CVD, physical disability (defined as independent in the Katz Activities of Daily Living scale), and dementia, and being expected to survive at least 5 years. Individuals with hypertension or diabetes were eligible, provided these conditions were well controlled. Individuals with a self-report or physician diagnosis of dementia at recruitment, or with a Modified Mini-Mental State Examination (3MS) (10) score of <78, were ineligible; however, it was possible that some individuals had mild levels of CI at baseline.
ASPREE participants were followed annually with in-person visits and with intermittent 6-month telephonic visits to assess change in medical conditions or study enrollment status.
ASPREE was approved by multiple institutional review boards in the United States and Australia, registered with the International Standard Randomized Controlled Trial Number Register (ISRCTN83772183) and ClinicalTrials.gov (NCT01038583), and undertaken in accordance with the Declaration of Helsinki. All participants provided written informed consent.
Primary Exposures: eGFR and UACR
The primary study exposures were the kidney biomarkers eGFR and UACR. CKD was defined using Kidney Disease Improving Global Outcomes criteria of eGFR <60 ml/min per 1.73 m2 (2,11). We calculated eGFR on the basis of serum creatinine using the Chronic Kidney Disease Epidemiology Collaboration equation (12). Albuminuria was defined as a UACR ≥3 mg/mmol (≥26.6 mg/g), on the basis of a single measurement at the baseline study visit.
Cognitive Assessments and Outcome Measures
Cognitive assessments were administered by trained and accredited staff at baseline and year 1, and then biennially over the follow-up period (year 3 and year 5, or at their close-out [last] visit during the trial phase, which could be visit 3 through 7). The cognitive battery included the 3MS, a 100-point measure of global cognition (9); the Delayed Recall component for episodic memory of the Hopkins Verbal Learning Test–Revised (HVLT-R) (13), a word list of 12 items; the Controlled Oral Word Association Test (COWAT) (14) for language (scored as the total number of words beginning with F generated in 1 minute); and the Symbol Digit Modalities Test (SDMT) (15), a measure of psychomotor speed (maximum score of 110). For all cognitive tests, a higher score indicates better performance.
Incident Dementia Outcome
Individuals with a suspected dementia diagnosis (a “trigger”) were referred for a dementia assessment that included further standardized cognitive tests and functional assessments. Dementia triggers were defined as a 3MS score <78 (16), a drop of more than ten points from the predicted score (17), a medical record report of memory concerns or other cognitive problems, a clinician diagnosis of dementia, or prescription of cholinesterase inhibitors in the participant’s medical records.
The dementia assessment included the Confusion Assessment Method (18), Alzheimer Disease Assessment Scale–Cognitive Subscale (19), Color Trails Test (20), Lurian overlapping figures test (21), and the Alzheimer Disease Cooperative Study Activities of Daily Living scale (22), along with the results of blood tests, brain computed tomography scanning or magnetic resonance imaging, and clinical notes from medical records (see Supplemental Appendix 2 for additional dementia assessment cognitive test results details).
The available information was reviewed by the dementia adjudication committee (consisting of a panel of neurologists, neuropsychologists, and geriatricians from Australia and the United States) who were blinded to treatment allocation. Dementia was adjudicated according to Diagnostic and Statistical Manual, Fourth Edition criteria (23). The date of diagnosis of dementia was taken as the date the dementia trigger occurred that resulted in a confirmed dementia diagnosis by the adjudication committee.
For participants without a dementia trigger, we defined those with significant cognitive decline as CIND if they had a >1.5-SD decline in cognitive score from their own baseline value on the HVLT-R Delayed Recall, SDMT, or COWAT. This definition did not include participants with evidence of only a transient decline (e.g., those with a >1.5-SD drop at one follow-up, but scoring above this threshold at a subsequent follow-up). Cognitive change was examined using the continuous scores on each of the cognitive tests over the follow-up period.
Descriptive statistics are presented as medians (interquartile range; IQR) for continuous variables, and frequency (percentage) for categoric variables. We used linear regression models to assess the association between kidney function measures and cognitive performance at baseline. In each model, one of baseline (binary) albuminuria, UACR doubling, or a 10-U eGFR lowering were included as independent variables (specifically, eGFR measurements scaled by a factor of −10, and UACR transformed by logarithm to the base two were included as a continuous variable); these measures were chosen to reflect clinically relevant declines in kidney function. Cognitive test scores were the dependent variables. We used multivariable mixed effects models to examine the association between baseline kidney function and the rate of cognitive change over time. Each model allowed for individual variation in cognitive performances at baseline and subsequent rate of change (i.e., random intercept and random slope model). We included an interaction term between visit time (years since randomization) and baseline kidney function to examine whether the rate of cognitive change depended on kidney function values at baseline.
Cox proportional hazards regression models were implemented to evaluate the effect of baseline kidney function on time to incident CIND and dementia. Cause-specific hazards were calculated, with all-cause deaths treated as competing events. Additionally, conditional covariate-adjusted survival curves (24) were presented to visualize the effect of baseline UACR levels on incidence of CIND and dementia (but not for eGFR, because initial Cox regression models demonstrated no significant association between baseline eGFR and incident CIND or dementia).
Lastly, to determine to what extent eGFR or UACR, individually or in combination, are associated with cognitive outcomes, additional analyses were performed to measure the effect of the interactions between an eGFR <60/≥60 ml/min per 1.73 m2 and a UACR of >3/≤3 mg/mmol, and, alternatively, an eGFR of <45/≥45 ml/min per 1.73 m2 and a UACR of >3/≤3 mg/mmol, on risks of baseline CI, cognitive decline, incident CIND, and dementia in both the cross-sectional and longitudinal models and the Cox regression models described above.
All models were adjusted for baseline characteristics: age, sex, education (<12 years or ≥12 years), treatment allocation to aspirin (versus placebo), diabetes, hypertension, and race (White/Australia, White/United States, Black, Hispanic, and other).
Participants with missing information in cognitive outcome measures, kidney function (UACR or eGFR), or any of the adjusted variables were excluded from the analyses. All analyses were performed using R version 4.0. Adjusted survival curves were computed using the survminer R package. Of note, study enrollment bias in this report was minimized due to the randomized trial recruitment design of the ASPREE study.
Baseline characteristics, overall and by different eGFR levels, are shown in Table 1 for the cohort (Supplemental Figure 1). Among 19,114 ASPREE participants, 18,560 (98%) and 18,131 (95%) had baseline measurement of eGFR and UACR, respectively. Median age was 74 years, 56% were female, 74% had hypertension, and 11% had diabetes. Median (IQR) eGFR was 74 (63–84) ml/min per 1.73 m2; 19% had an eGFR of <60 ml/min per 1.73 m2, and 16% an eGFR of 45–60 ml/min per 1.73 m2. Median (IQR) UACR was 0.8 (0.5–1.5) mg/mmol (7.1 [IQR, 4.4–13.3] mg/g), and UACR distribution was highly skewed: most (89%) had a UACR of <3 mg/mmol (≤26.6 mg/g), 9% had a UACR of 3 to <10 mg/mmol (88.5 mg/g), and only 3% had a UACR of ≥10 mg/mmol (88.5 mg/g) (not shown). Participants with an eGFR of <60 ml/min per 1.73 m2 (compared with those with an eGFR of ≥60 ml/min per 1.73 m2) tended to be older and more likely to have diabetes and hypertension.
Table 1. -
||Total Cohort (N=19,114)
||CKD (eGFR <45 ml/min per 1.73 m2) (N=606)
||Mild CKD (eGFR 45 to <60 ml/min per 1.73 m2) (N=2821)
||All CKD (eGFR <60 ml/min per 1.73 m2) (N=3427)
||eGFR ≥60 ml/min per 1.73 m2 (N=15,223)
|Age, yr, median (IQR)
Age group, yr, n (%)
Sex, n (%)
|eGFR, ml/min per 1.73 m2, median (IQR)
|UACR, mg/mmol, median (IQR)
|UACR <3 mg/mmol, n (%)
|UACR ≥3 mg/mmol, n (%)
Education, yr, n (%)
Race, n (%)
| White/United States
Country, n (%)
| United States
|Diabetes, n (%)
|Systolic BP, mm Hg, median (IQR)
|Diastolic BP, mm Hg, median (IQR)
|Hypertension, n (%)
|Total cholesterol, mmol/L, median (IQR)
|Current smoker, n (%)
IQR, interquartile range; UACR, urine albumin-creatinine ratio.
aDefined by self-report, blood glucose ≥126 mg/dl, or treatment for diabetes.
bDefined as the mean of three systolic BP readings >140 mm Hg, or three diastolic BP readings of >90 mm Hg, or by pharmacologic treatment for high BP.
Median follow-up time in the entire ASPREE cohort was 4.7 years; 77% (n=14,729) of the cohort completed their third annual visit for cognitive tests (i.e., at least 3 years of follow-up). By the end of the trial, 237 participants in this report withdrew, 1052 were deceased, and 296 were lost to follow-up. Further details regarding number of participants with annual visits with cognitive testing each year are described in Supplemental Table 1.
Cross-Sectional Association between Kidney Function and Cognitive Function
In adjusted analyses, baseline eGFR was not associated with scores on any of the four cognitive tests (Table 2). However, UACR ≥3 mg/mmol (≥26.6 mg/g) was significantly associated with baseline scores for three of the four tests (3MS, HVLT-R, and SDMT) in adjusted models. The association with the SDMT of β=−1.24 (95% CI, −1.66 to −0.81; P<0.001) appears clinically significant, because it was more than double the size of the mean change in SDMT in all participants: −0.50 points over 4.7 years of follow-up (with baseline mean [SD] SDMT of 36.7 [10.1]; Supplemental Table 2). The association between a UACR of ≥3 mg/mmol (≥26.6 mg/g) and scores on the 3MS, at β=−0.33 (95% CI, −53 to −0.13; P=0.001), and HVLT-R, at β=−0.24 (95% CI, −0.36 to −0.11; P<0.001), were statistically significant, but the absolute score changes were minor. Similar results were seen per doubling of the UACR value.
Table 2. -
Association between kidney function and cognitive function at baseline
||Estimate, 95% Confidence Interval
||Estimate, 95% Confidence Interval
eGFR (per 10-U decrease)
||−0.16 (−0.21 to −0.11)
||−0.01 (−0.06 to 0.03)
||−0.07 (−0.12 to −0.03)
||−0.03 (−0.08 to 0.01)
||−0.50 (−0.60 to −0.39)
||0.09 (−0.01 to 0.19)
||−0.12 (−0.15 to −0.10)
||−0.01 (−0.04 to 0.02)
UACR (≥3 versus <3 mg/mmol)
||−0.85 (−1.06 to −0.64)
||−0.33 (−0.53 to −0.13)
||−0.34 (−0.55 to −0.13)
||−0.06 (−0.26 to 0.14)
||−2.62 (−3.09 to −2.16)
||−1.24 (−1.66 to −0.81)
||−0.53 (−0.65 to −0.40)
||−0.24 (−0.36 to −0.11)
UACR (per two-fold increase)
||−0.32 (−0.39 to −0.24)
||−0.15 (−0.23 to −0.08)
||−0.14 (−0.22 to −0.07)
||−0.07 (−0.15 to 0.00)
||−1.09 (−1.25 to −0.92)
||−0.61 (−0.76 to −0.45)
||−0.18 (−0.23 to −0.13)
||−0.10 (−0.14 to −0.05)
Models were adjusted for age, sex, education (<12 and ≥12 years), treatment, diabetes, hypertension, and race (White/Australia, White/United States, Black, Hispanic, and other). 3MS, Modified Mini-Mental State Examination; COWAT, Controlled Oral Word Association Test; SDMT, Symbol Digit Modalities Test; HVLT4, Hopkins Verbal Learning Test 4; UACR, urine albumin-creatinine ratio.
aeGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation.
bThe estimate coefficients represent an increase in the performance of each cognitive test per 10-U decrease in eGFR value.
cThe estimate coefficients represent an increase in the performance of each cognitive test when the UACR level changed from <3 to ≥3 mg/mmol.
dThe estimate coefficients represent an increase in the performance of each cognitive test per two-fold increase in UACR value.
Association between Baseline Kidney Measures and Longitudinal Change in Performance on Cognitive Testing
A mean of 3.2 annual cognitive assessments per participant was performed over a median follow-up of 4.7 years. In fully adjusted linear mixed effects, eGFR at baseline was statistically significantly associated with decline of both the 3MS and SDMT scores, but the magnitudes of the associations were clinically negligible (Table 3).
Table 3. -
Association between baseline kidney function and change in cognitive function
||Modified Mini-Mental State Examination
||Controlled Oral Word Association Test
||Symbol Digit Modalities Test
||Hopkins Verbal Learning Test 4
|Estimate, 95% Confidence Interval
||Estimate, 95% Confidence Interval
||Estimate, 95% Confidence Interval
||Estimate, 95% Confidence Interval
eGFR (per 10-U decrease)
||0.01 (−0.04 to 0.05)
||−0.02 (−0.06 to 0.03)
||0.13 (0.03 to 0.22)
||−0.00 (−0.03 to 0.03)
| Time (+5yr)
||−1.26 (−1.79 to −0.74)
||0.96 (0.56 to 1.37)
||−3.41 (−4.11 to −2.71)
||−0.18 (−0.43 to 0.07)
||−0.11 (−0.18 to −0.04)
||−0.03 (−0.08 to 0.03)
||−0.11 (−0.20 to −0.01)
||−0.03 (−0.07 to 0.00)
UACR (≥3 versus <3)
||−0.31 (−0.50 to −0.13)
||−0.09 (−0.29 to 0.10)
||−1.09 (−1.49 to −0.68)
||−0.27 (−0.39 to −0.15)
| Time (+ 5yr)
||−0.39 (−0.50 to −0.29)
||1.19 (1.11 to 1.27)
||−2.55 (−2.69 to −2.41)
||0.06 (0.01 to 0.11)
||−0.67 (−0.99 to −0.35)
||−0.13 (−0.38 to 0.11)
||−0.76 (−1.18 to −0.33)
||−0.25 (−0.40 to −0.09)
||−0.15 (−0.22 to −0.08)
||−0.07 (−0.14 to 0.00)
||−0.57 (−0.72 to −0.42)
||−0.11 (−0.15 to −0.06)
| Time (+ 5yr)
||−0.16 (−0.32 to 0.00)
||1.22 (1.10 to 1.34)
||−2.40 (−2.61 to −2.19)
||0.15 (0.07 to 0.23)
||−0.28 (−0.39 to −0.16)
||−0.04 (−0.13 to 0.05)
||−0.22 (−0.37 to −0.06)
||−0.10 (−0.16 to −0.05)
All models were adjusted for age, sex, education (<12 and ≥12 years), diabetes, treatment, hypertension, and race (White/Australia, White/United States, Black, Hispanic, and other). 3MS, Modified Mini-Mental State Examination; COWAT, Controlled Oral Word Association Test; SDMT, Symbol Digit Modalities Test; HVLT4, Hopkins Verbal Learning Test 4; Est, estimate; UACR, urine albumin-creatinine ratio.
aNumber of included participants in the model for 3MS, COWAT, SDMT, and HTVLT are 18,638, 18,631, 18,606, and 18,606, respectively.
bNumber of included participants in the model for 3MS, COWAT, SDMT, and HTVLT are 18,120, 18,114, 18,088, and 18,091, respectively.
Baseline albuminuria (UACR ≥3 mg/mmol; ≥26.6 mg/g) was significantly associated with cognitive decline on all cognitive tests except the COWAT, and most strongly associated with SDMT performance, with β=−0.76 (95% CI, −1.18 to −0.33; P<0.001). Similar results were seen for UACR doubling.
Longitudinal Association between Baseline Kidney Measures and Incident CIND and Dementia
Over the median 4.7 years of follow-up, 2777 cases of CIND and 563 cases of adjudicated incident dementia occurred with either baseline eGFR or UACR measures. The overall rate of incident dementia was approximately 6.8/1000 patient-years (Figure 1). We found no statistically significant association between baseline eGFR and time to CIND or dementia (Table 4). In contrast, both the binary (UACR ≥3/<3 mg/mmol [≥26.6/<26.6 mg/g]) and continuous (UACR doubling) measures of UACR were associated with an increased risk for both incident CIND and ncident dementia. Specifically, baseline albuminuria corresponded to a 19% increased risk of CIND (hazard ratio, 1.19; 95% CI, 1.07 to 1.33), and 32% increased risk of incident dementia (hazard ratio, 1.32; 95% CI, 1.06 to 1.66). These associations are further illustrated in Figure 1. Of note, the three waves of incident CIND correspond to the annual visits that, per study design, most often included the cognitive assessments (baseline, and annual visits 1, 3, and 5).
Table 4. -
Association between baseline kidney function and incident cognitive impairment no dementia and dementia
||Hazard Ratio, 95% Confidence Interval
eGFR (per 10-U decrease)
||0.98 (0.95 to 1.01)
||0.95 (0.89 to 1.01)
UACR (≥3 versus <3)
||1.19 (1.07 to 1.33)
||1.32 (1.06 to 1.66)
||1.07 (1.03 to 1.11)
||1.13 (1.05 to 1.23)
All models were adjusted for age, sex, education (<12 and ≥12 years), diabetes, treatment group, hypertension, and race (White/Australia, White/United States, Black, Hispanic, and other). CIND, cognitive impairment no dementia; HR, hazard ratio; UACR, urine albumin-creatinine ratio.
No Interactions between eGFR and UACR on Baseline Cognitive Performance, Cognitive Decline, Incident CIND, and Incident Dementia
In adjusted linear regression models that included an interaction term for an eGFR of <60/≥60 or <45/≥45 ml/min per 1.73 m2 and UACR ≥3/<3 mg/mmol (≥26.6/<26.6 mg/g), there were no statistically significant interactions in the models for baseline cognitive function (Supplemental Table 3), cognitive decline (with a three-way interaction term for time; Supplemental Table 4), or in Cox proportional hazards regression models for incident CIND or dementia (Supplemental Table 5).
In a population of initially healthy, community-dwelling older adults without known CVD and only 3% prevalence of eGFR <45 ml/min per 1.73 m2, low levels of albuminuria (UACR of ≥3 mg/mmol; ≥26.6 mg/g) at study entry were associated with lower baseline cognitive function and more rapid cognitive decline over a mean of nearly 5 years. This association was seen in three of the four cognitive tests assessed: the 3MS (a global assessment), HVLT-R (verbal memory), and, especially, the SDMT (processing speed, attention, and visual scanning). More importantly, baseline albuminuria was also associated with increased rates of both incident CIND and dementia. Thus, in a relatively healthy elderly population with a median eGFR of 74 ml/min per 1.73 m2, mild albuminuria portended a substantially increased risk of cognitive decline in cognitive domains important for day-to-day function and dementia over a period of 5 years.
Although the association of baseline albuminuria on decline in SDMT scores was relatively modest, at 1.24 lower points over 5 years, it was still approximately 2.5 times greater than the mean SDMT decline in the entire cohort. UACR ≥3 mg/mmol (≥26.6 mg/g) was also associated with a 19% increased risk of CIND, and a 32% increased risk of dementia. The lack of association between UACR and COWAT, or verbal fluency, may be reflective of the low prevalence of poor performance on the COWAT at baseline, and on average mild decline in COWAT.
In contrast, eGFR was not associated with baseline cognitive performance on any tests. In addition, its association with longitudinal cognitive decline, seen only with the 3MS and SDMT scores, appeared modest in comparison to albuminuria, and it was not associated with incident CIND or dementia.
Finally, we found no significant interactions between lower eGFR and albuminuria on baseline cognitive function, incident CIND, or dementia. Thus, the association of albuminuria with the negative cognitive outcomes did not vary by eGFR level in this cohort with a low prevalence of eGFR <45 ml/min per 1.73 m2.
In a systematic review of the relation of UACR with CI, in populations with less advanced CKD (eGFR >45 ml/min per 1.73 m2), UACR was more frequently associated with CI than eGFR; however, in those with more severe CKD (eGFR <45 ml/min per 1.73 m2), eGFR was more often a significant risk factor (25–27). For example, in the Rancho Bernardo Study of older relatively healthy men, albuminuria but not eGFR was a predictor of cognitive function measured approximately 7 years later (26).
In the BRINK (BRain IN Kidney Disease) study of 574 participants with a mean eGFR of 46 ml/min per 1.73 m2, an eGFR of <30 ml/min per 1.73 m2 was associated with over three-fold increased odds of severe CI in those with a UACR of <2.7 ≥3 mg/mmol (≥26.6 mg/g),mg/mmol (<30 mg/g), but not for a UACR of >2.7 mg/mmol (>30 mg/g) (4). Because 97% of the ASPREE participants had an eGFR of >45 ml/min per 1.73 m2, our finding that UACR was a stronger risk factor than eGFR for CI and incident dementia adds to, and strengthens support for, these prior observations.
Although elevated UACR in the absence of low eGFR has been identified previously as being independently associated with CI and cognitive decline, our study is novel in that our study population was free of known CVD at baseline, whereas the majority of studies have been in populations with a higher prevalence of CVD or diabetes and moderate-to-severe CKD (28,29). Thus, our results are generalizable to the growing population of older people without known CVD, and in whom there may be lower suspicion for CI due to the known increased risk for CI in those with CVD.
The pathophysiology of the associations between the kidney biomarkers eGFR and UACR and cognitive function is complex and challenging to disentangle from the underlying vascular risk factors and high inflammatory load in CKD (27,30), which, together, drive parallel trajectories of microvascular endothelial dysfunction in both the kidney and the brain, leading to high rates of vasculopathic outcomes (31,32). Albuminuria is often considered a measure of kidney damage, or glomerular angiopathy due to microvascular endothelial inflammation, which frequently precedes a decline in eGFR. Albuminuria in our study was strongly associated with lower performance on the SDMT, a measure of processing speed, both cross-sectionally and longitudinally. Lower SDMT scores are observed in patients with frontal and subcortical microvascular disease, presenting as microinfarcts or white matter hyperintensities on brain magnetic resonance imaging (30), and in vascular CI (33). Decreased eGFR is often considered a more general measure of micro- and macrovascular disease, but may also more specifically reflect the effect of kidney toxins, or uremia, on neuronal function. Both kidney biomarker paradigms result in increased risk of structural and microvascular brain pathology and secondary CI.
The strengths of our study include the large, community-based, well-characterized study cohort across two countries, including approximately 9% non-White individuals, and the absence of CVD at baseline. Our cohort study benefited from the initial rigorous clinical trial design, which included annual creatinine and UACR, regularly scheduled cognitive assessments, follow-up–triggered dementia assessments, and dementia adjudication.
Study limitations include the overall minor decline in cognitive function over the follow-up period, which may have decreased ability to detect associations between kidney function and cognitive decline, and the use of a one-time measure of UACR at study baseline. Fewer participants had cognitive assessments as the study progressed, due, in part, to some being recruited later in the study, thus decreasing their number of annual assessments before trial end, and to participants being less willing to participate in cognitive testing as they become self-aware of cognitive decline, which is frequently observed in longitudinal studies of CI. However, 77% of the cohort had reached at least 3 years of follow-up. Our results are likely not generalizable to older populations with higher CVD risk profiles and higher prevalence of advanced CKD. We are also aware that obtaining a urine specimen in older patients in the clinic setting can be difficult for some, but, at the same time, have observed that urine samples have more frequently become a part of the annual physical examination for older patients, perhaps due to the increasing awareness of the growing prevalence of CKD in older patients.
The critical translatable implication of our results is that a low level of albuminuria, now easily measured in the clinic setting, could be used to identify those at elevated risk for prevalent CI, cognitive decline, CIND, and dementia in persons ≥70 years without concomitant advanced CKD (eGFR <30 ml/min per 1.73 m2). Importantly, baseline albuminuria identified those at increased risk of incident CIND early in the natural history of dementia, enabling interventions to decrease negative outcomes associated with cognitive decline promptly, when albuminuria is first noted.
We recommend consideration of two influential, translatable interventions in the clinic setting for those aged ≥70 years with a UACR of ≥3 mg/mmol (≥26.6 mg/g, or approximately 30 mg/g) and without an eGFR of <45 ml/min per 1.73 m2: (1) screen annually for CI, and, if present, (2) identify a caregiver or companion to attend clinic visits with the affected patient, and to provide medication and disease management supervision in the home, or obtain home health services to ensure this. Two options for cognitive screening tests are the 8- to 10-minute Montreal Cognitive Assessment (MoCA) (34), with high sensitivity to mild and greater levels of CI (although it requires an annual brief training and certification), or the Mini-Cog assessment, a 3-minute test that detects moderate to severe CI (35). Of note, at the time of the study’s design >10 years ago, ASPREE used the 3MS for global cognitive screening because the MoCA was not commonly used in large trials and the 3MS enabled easy comparisons of results between prior and concurrent studies. However, the MoCA is now frequently used as the primary global screening instrument or overall measure of cognitive function in many studies (36,37). We also chose not to use the Mini-Cog assessment because it has lower sensitivity to mild CI.
We are aware that our CI screening recommendation may be perceived to be in conflict with the recent US Preventive Services Task Force (USPSTF) recommendations against screening in asymptomatic older individuals (https://uspreventiveservicestaskforce.org/uspstf/recommendation/cognitive-impairment-in-older-adults-screening) (38), which were targeted at asymptomatic older adults regardless of their risk factors for CI. However, as discussed, we and others have demonstrated that individuals with albuminuria are at substantially increased risk of prevalent and incident CI and dementia. Second, asymptomatic does not equate to absence of CI, only absence of reported symptoms of CI. Many older individuals with memory or cognitive concerns hesitate to bring them to medical attention until brought in by family members—often after they have already reached a stage of mild-to-moderate dementia, or after a “sentinel event,” such as a hospitalization with delirium (a harbinger and litmus test for dementia). In addition, the USPSTF findings were made on the basis of literature reviews that were largely inconclusive because there were inadequate numbers of robust well-designed studies to measure the outcomes of cognitive screening, but that does not mean screening is nonbeneficial, rather it means that the studies needed to prove their benefit have not been done. In fact, one of their recommendations was to call for such studies. Early identification of CI would alert clinicians, families, and care partners to the need for increased vigilance to ensure adherence with medications and kidney disease management plans and decrease downstream risk for related hospitalizations and mortality (39,40).
In this large, healthy population of community-dwelling older people with a median eGFR of 74 ml/min per 1.73 m2, the presence of mild albuminuria was associated with modestly lower baseline performance, greater decline on tests of global cognitive function and processing speed, and substantially increased risk of incident CIND and dementia. Our findings suggest that screening global cognitive tests could be considered for those ≥70 years with a UACR of ≥3 mg/mmol and without an eGFR of <45 ml/min per 1.73 m2 to identify those at highest risk of cognitive decline and dementia.
A.M. Murray reports being a consultant to Alkahest, Inc. J.B. Wetmore reports receiving research funding from Amgen, AstraZeneca, BMS/Pfizer, Genentech, Merck, and OPKO Health; serving on advisory boards for Aurinia Pharmaceuticals, OPKO Health, and Reata Pharmaceuticals; having consultancy agreements via ad hoc consulting for Bristol Myers Squibb (BMS)–Pfizer Alliance; receiving honoraria from, and serving as a scientific advisor or member of, BMS-Pfizer Alliance (with occasional participation on ad hoc advisory boards); and receiving honoraria for academic continuing medical education from Healio and the Nephrology Self-Assessment Program. All remaining authors have nothing to disclose.
This work was supported by the National Institute on Aging and the National Cancer Institute grants U01AG029824 and U19AG062682, the Department of Health, Australian Government, National Health and Medical Research Council (NHMRC) grants 334047 and 1127060, Monash University, and the Victorian Cancer Agency. J. Ryan is funded by an NHMRC Dementia Research Leader Fellowship (APP1135727).
The authors thank the dedicated staff in Australia and the United States for conducting the trial, the ASPREE participants for their generous contributions of time, and the general practitioners and staff of the medical clinics who cared for the participants.
A.M. Murray was responsible for investigation; A.M. Murray, K.R. Polkinghorne, L.T.P. Thao, J.B. Wetmore, R. Wolfe, and R.L. Woods wrote the original draft; A.M. Murray, L.T.P. Thao, R. Wolfe, and R.L. Woods were responsible for formal analysis and methodology; A.M. Murray, L.T.P. Thao, and R.L. Woods were responsible for validation; A.M. Murray and R.L. Woods conceptualized the study, were responsible for project administration, and provided supervision; L. T.P. Thao was responsible for data curation, software, and visualization; and all authors reviewed and edited the manuscript.
Data Sharing Statement
All individual participant data (reidentifiable) that underlie the results reported in this article are available upon request to qualified researchers without limit of time, subject to approval and a standard data sharing agreement. Details regarding requests to access the data are available through the study website (ASPREE.org). The data will then be made available through a web-based data portal safe haven at Monash University (Melbourne, Australia). The ASPREE trial protocol and statistical analysis plan have been published (9,41).
This article contains supplemental material online at http://kidney360.asnjournals.org/lookup/suppl/doi:10.34067/KID.0005672021/-/DCSupplemental.
Supplemental Appendix 1. ASPREE study design and population.
Supplemental Appendix 2. Incident dementia outcome and dementia assessment.
Supplemental Figure 1. Consolidated Standards of Reporting Trials (CONSORT) flow diagram of participants in the Aspirin in Reducing Events in the Elderly (ASPREE) trial.
Supplemental Table 1. Sample sizes for those who had cognitive assessments at each annual visit (AV).
Supplemental Table 2. Summary of cognitive tests at baseline and annual visits (AV).
Supplemental Table 3. Association between renal function and cognitive function at baseline: parameter estimates for the interaction terms between eGFR and UACR.
Supplemental Table 4. Association between baseline renal function and change in cognitive function: parameter estimates for the interaction terms between eGFR and UACR and time.
Supplemental Table 5. Association between baseline renal functions and CIND and dementia: Parameter estimates for the interaction terms between eGFR and UACR.
1. Kurella Tamura M, Wadley V, Yaffe K, McClure LA, Howard G, Go R, Allman RM, Warnock DG, McClellan W: Kidney function and CI in US adults: The Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study. Am J Kidney Dis 52: 227–234, 2008 https://doi.org/10.1053/j.ajkd.2008.05.004
2. Yaffe K, Ackerson L, Kurella Tamura M, Le Blanc P, Kusek JW, Sehgal AR, Cohen D, Anderson C, Appel L, Desalvo K, Ojo A, Seliger S, Robinson N, Makos G, Go AS; 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 58: 338–345, 2010 https://doi.org/10.1111/j.1532-5415.2009.02670.x
3. Zammit AR, Katz MJ, Lai JY, Zimmerman ME, Bitzer M, Lipton RB: Association between renal function and cognitive ability domains in the Einstein aging study: A cross-sectional analysis. J Gerontol A Biol Sci Med Sci 70: 764–770, 2015 https://doi.org/10.1093/gerona/glu185
4. Burns CM, Knopman DS, Tupper DE, Davey CS, Slinin YM, Lakshminarayan K, Rossom RC, Pederson SL, Gilbertson DT, Murray AM: Prevalence and risk of severe cognitive impairment in advanced chronic kidney disease. J Gerontol A Biol Sci Med Sci 73: 393–399, 2018 https://doi.org/10.1093/gerona/glx241
5. Gupta J, Mitra N, Kanetsky PA, Devaney J, Wing MR, Reilly M, Shah VO, Balakrishnan VS, Guzman NJ, Girndt M, Periera BG, Feldman HI, Kusek JW, Joffe MM, Raj DS; CRIC Study Investigators: Association between albuminuria, kidney function, and inflammatory biomarker profile in CKD in CRIC. Clin J Am Soc Nephrol 7: 1938–1946, 2012 https://doi.org/10.2215/CJN.03500412
6. Weiner DE, Bartolomei K, Scott T, Price LL, Griffith JL, Rosenberg I, Levey AS, Folstein MF, Sarnak MJ: Albuminuria, cognitive functioning, and white matter hyperintensities in homebound elders. Am J Kidney Dis 53: 438–447, 2009 https://doi.org/10.1053/j.ajkd.2008.08.022
7. 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. Available at: https://www.usrds.org/previous-adrs/2010-2019/
Accessed June 1, 2021
8. Ryan J, Storey E, Murray AM, Woods RL, Wolfe R, Reid CM, Nelson MR, Chong TTJ, Williamson JD, Ward SA, Lockery JE, Orchard SG, Trevaks R, Kirpach B, Newman AB, Ernst ME, McNeil JJ, Shah RC; ASPREE Investigator Group: Randomized placebo-controlled trial of the effects of aspirin on dementia and cognitive decline. Neurology 95: e320–e331, 2020 https://doi.org/10.1212/WNL.0000000000009277
9. ASPREE Investigator Group: Study design of ASPirin in Reducing Events in the Elderly (ASPREE): A randomized, controlled trial. Contemp Clin Trials 36: 555–564, 2013 https://doi.org/10.1016/j.cct.2013.09.014
10. Teng EL, Chui HC: The Modified Mini-Mental State (3MS) examination. J Clin Psychiatry 48: 314–318, 1987
11. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group: KDIGO clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl 3: 1–150, 2013
12. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration): A new equation to estimate glomerular filtration rate [published correction appears in Ann Intern Med
155: 408, 2011]. Ann Intern Med 150: 604–612, 2009 https://doi.org/10.7326/0003-4819-150-9-200905050-00006
13. Benedict RHB, Schretlen D, Groninger L, Brandt J: Hopkins Verbal Learning Test – Revised: Normative data and analysis of inter-form and test–retest reliability. Clin Neuropsychol 12: 43–55, 2001 https://doi.org/10.1076/clin.126.96.36.1996
14. Ross TP, Calhoun E, Cox T, Wenner C, Kono W, Pleasant M: The reliability and validity of qualitative scores for the Controlled Oral Word Association Test. Arch Clin Neuropsychol 22: 475–488, 2007 https://doi.org/10.1016/j.acn.2007.01.026
15. Smith A: The Symbol-Digit Modalities Test: A Neuropsychologic Test of Learning and Other Cerebral Disorders. In: Learning Disorders, edited by Helmuth J, Seattle, WA, Special Child Publications, 1968, pp 83–91
16. Bland RC, Newman SC: Mild dementia or cognitive impairment: The Modified Mini-Mental State examination (3MS) as a screen for dementia. Can J Psychiatry 46: 506–510, 2001 https://doi.org/10.1177/070674370104600604
17. Tombaugh TN: Test-retest reliable coefficients and 5-year change scores for the MMSE and 3MS. Arch Clin Neuropsychol 20: 485–503, 2005 https://doi.org/10.1016/j.acn.2004.11.004
18. Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP, Horwitz RI: Clarifying confusion: the confusion assessment method. A new method for detection of delirium. Ann Intern Med 113: 941–948, 1990 https://doi.org/10.7326/0003-4819-113-12-941
19. Galasko D, Bennett D, Sano M, Ernesto C, Thomas R, Grundman M, Ferris S: An inventory to assess activities of daily living for clinical trials in Alzheimer’s disease. The Alzheimer’s Disease Cooperative Study. Alzheimer Dis Assoc Disord 11: S33–S39, 1997 https://doi.org/10.1097/00002093-199700112-00005
20. D’Elia L: Color Trails Test: Professional Manual, Odessa, FL, Psychological Assessment Resources, 1996
21. Alegret M, Boada-Rovira M, Vinyes-Junqué G, Valero S, Espinosa A, Hernández I, Modinos G, Rosende-Roca M, Mauleón A, Becker JT, Tárraga L: Detection of visuoperceptual deficits in preclinical and mild Alzheimer’s disease. J Clin Exp Neuropsychol 31: 860–867, 2009 https://doi.org/10.1080/13803390802595568
22. Graham DP, Cully JA, Snow AL, Massman P, Doody R: The Alzheimer’s Disease Assessment Scale-Cognitive subscale: Normative data for older adult controls. Alzheimer Dis Assoc Disord 18: 236–240, 2004
23. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), Washington, DC, American Psychiatric Association, 1994
24. Nieto FJ, Coresh J: Adjusting survival curves for confounders: A review and a new method. Am J Epidemiol 143: 1059–1068, 1996 https://doi.org/10.1093/oxfordjournals.aje.a008670
25. Georgakis MK, Dimitriou NG, Karalexi MA, Mihas C, Nasothimiou EG, Tousoulis D, Tsivgoulis G, Petridou ET: Albuminuria in association with cognitive function and dementia: A systematic review and meta-analysis. J Am Geriatr Soc 65: 1190–1198, 2017 https://doi.org/10.1111/jgs.14750
26. Jassal SK, Kritz-Silverstein D, Barrett-Connor E: A prospective study of albuminuria and cognitive function in older adults: The Rancho Bernardo study. Am J Epidemiol 171: 277–286, 2010 https://doi.org/10.1093/aje/kwp426
27. Sacre JW, Magliano DJ, Zimmet PZ, Polkinghorne KR, Chadban SJ, Anstey KJ, Shaw JE: Associations of chronic kidney disease markers with cognitive function: A 12-year follow-up study. J Alzheimers Dis 70: S19–S30, 2019 https://doi.org/10.3233/JAD-180498
28. Murray AM, Barzilay JI, Lovato JF, Williamson JD, Miller ME, Marcovina S, Launer LJ; Action to Control Cardiovascular Risk in Diabetes Memory in Diabetes (ACCORD-MIND) Substudy Investigators: Biomarkers of renal function and cognitive impairment in patients with diabetes. Diabetes Care 34: 1827–1832, 2011 https://doi.org/10.2337/dc11-0186
29. Barzilay JI, Lovato JF, Murray AM, Williamson J, Ismail-Beigi F, Karl D, Papademetriou V, Launer LJ: Albuminuria and cognitive decline in people with diabetes and normal renal function. Clin J Am Soc Nephrol 8: 1907–1914, 2013 https://doi.org/10.2215/CJN.11321112
30. Lu R, Kiernan MC, Murray A, Rosner MH, Ronco C: Kidney-brain crosstalk in the acute and chronic setting. Nat Rev Nephrol 11: 707–719, 2015 https://doi.org/10.1038/nrneph.2015.131
31. Bugnicourt JM, Godefroy O, Chillon JM, Choukroun G, Massy ZA: Cognitive disorders and dementia in CKD: The neglected kidney-brain axis. J Am Soc Nephrol 24: 353–363, 2013 https://doi.org/10.1681/ASN.2012050536
32. Segura B, Jurado MA, Freixenet N, Bargalló N, Junqué C, Arboix A: White matter fractional anisotropy is related to processing speed in metabolic syndrome patients: A case-control study. BMC Neurol 10: 64, 2010 https://doi.org/10.1186/1471-2377-10-64
33. Dong Y, Slavin MJ, Chan BP, Venketasubramanian N, Sharma VK, Collinson SL, Sachdev PS, Chen CL: Improving screening for vascular cognitive impairment at three to six months after mild ischemic stroke and transient ischemic attack. Int Psychogeriatr 26: 787–793, 2014 https://doi.org/10.1017/S1041610213002457
34. Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, Cummings JL, Chertkow H: The Montreal Cognitive Assessment, MoCA: A brief screening tool for mild Cognitive impairment [published correction appears in J Am Geriatr Soc
67: 1991, 2019]. J Am Geriatr Soc 53: 695–699, 2005 https://doi.org/10.1111/j.1532-5415.2005.53221.x
35. Borson S, Scanlan J, Brush M, Vitaliano P, Dokmak A: The mini-cog: A cognitive ‘vital signs’ measure for dementia screening in multi-lingual elderly. Int J Geriatr Psychiatry 15: 1021–1027, 2000 https://doi.org/10.1002/1099-1166(200011)15:11<1021::AID-GPS234>3.0.CO;2-6
36. Williamson JD, Pajewski NM, Auchus AP, Bryan RN, Chelune G, Cheung AK, Cleveland ML, Coker LH, Crowe MG, Cushman WC, Cutler JA, Davatzikos C, Desiderio L, Erus G, Fine LJ, Gaussoin SA, Harris D, Hsieh MK, Johnson KC, Kimmel PL, Tamura MK, Launer LJ, Lerner AJ, Lewis CE, Martindale-Adams J, Moy CS, Nasrallah IM, Nichols LO, Oparil S, Ogrocki PK, Rahman M, Rapp SR, Reboussin DM, Rocco MV, Sachs BC, Sink KM, Still CH, Supiano MA, Snyder JK, Wadley VG, Walker J, Weiner DE, Whelton PK, Wilson VM, Woolard N, Wright JT Jr, Wright CB; SPRINT MIND Investigators for the SPRINT Research Group: Effect of intensive vs standard blood pressure control on probable dementia: A randomized clinical trial. JAMA 321: 553–561, 2019 https://doi.org/10.1001/jama.2018.21442
37. Röhr S, Zülke A, Luppa M, Brettschneider C, Weißenborn M, Kühne F, Zöllinger I, Samos FAZ, Bauer A, Döhring J, Krebs-Hein K, Oey A, Czock D, Frese T, Gensichen J, Haefeli WE, Hoffmann W, Kaduszkiewicz H, König HH, Thyrian JR, Wiese B, Riedel-Heller SG: Recruitment and baseline characteristics of participants in the AgeWell.de Study—A Pragmatic cluster-randomized controlled lifestyle trial against cognitive decline. Int J Environ Res Public Health 18: 408, 2021
38. Owens DK, Davidson KW, Krist AH, Barry MJ, Cabana M, Caughey AB, Doubeni CA, Epling JW Jr, Kubik M, Landefeld CS, Mangione CM, Pbert L, Silverstein M, Simon MA, Tseng CW, Wong JB; US Preventive Services Task Force: Screening for cognitive impairment in older adults: US Preventive Services Task Force Recommendation Statement. JAMA 323: 757–763, 2020 https://doi.org/10.1001/jama.2020.0435
39. Chang TI, Desai M, Solomon DH, Winkelmayer WC: Kidney function and long-term medication adherence after myocardial infarction in the elderly. Clin J Am Soc Nephrol 6: 864–869, 2011 https://doi.org/10.2215/CJN.07290810
40. Walsh CA, Cahir C, Tecklenborg S, Byrne C, Culbertson MA, Bennett KE: The association between medication non-adherence and adverse health outcomes in ageing populations: A systematic review and meta-analysis. Br J Clin Pharmacol 85: 2464–2478, 2019 https://doi.org/10.1111/bcp.14075
41. Wolfe R, Murray AM, Woods RL, Kirpach B, Gilbertson D, Shah RC, Nelson MR, Reid CM, Ernst ME, Lockery J, Donnan GA, Williamson J, McNeil JJ: The aspirin in reducing events in the elderly trial: Statistical analysis plan. Int J Stroke 13: 335–338, 2018 https://doi.org/10.1177/1747493017741383