Kidney Function and the Risk of Stroke and Dementia: The Rotterdam Study

Abstract

Longitudinal population-based data on effects of kidney dysfunction in the development of stroke and dementia remains inconclusive. We investigated associations of kidney function with risk of stroke and dementia in 5,993 community-dwelling individuals (mean age: 69.0 years, 57.2% women). We calculated estimated glomerular filtration rates based on creatinine, cystatin-C, and a combination of these two. During a mean follow-up of 11.6 years (69,790 person-years), 1,360 individuals suffered a stroke (n = 601) or developed dementia (n = 759). We found that an impaired kidney function was related to a higher risk of stroke, but not to dementia.

Keywords

Dementia epidemiology glomerular filtration rate kidney function stroke

INTRODUCTION

Kidney dysfunction has a high global prevalence of around 13% which increases to up to 35% in persons aged 70 years or older [1]. Consequently, kidney dysfunction is regarded as important contributor to global morbidity and mortality [2]. Aside from its well-established role in coronary heart disease [3], accumulating evidence suggests that impaired kidney function may also exert harmful effects on the brain, by inducing cognitive impairment [4, 5], and even increasing the risk of stroke and dementia [6 –8]. However, population-based data with long-term follow-up remains limited. Moreover, an important aspect pertaining to the occurrence of stroke and dementia that was not taken into account in previous studies, is that these conditions frequently co-occur and are closely interrelated [9 –11].

Against this background, we investigated the association of kidney function with the risk of stroke and dementia in a large sample of community-dwelling middle-aged and elderly persons.

MATERIALS AND METHODS

Setting

This study was performed in the Rotterdam Study [12], a prospective population-based cohort study in middle-aged and elderly persons. Every 4-5 years, all participants are re-invited to undergo extensive follow-up examinations at the research center.

For the current study, the research center visit between 1997 and 2001 served as baseline, given that then blood-based kidney function measures were determined. Out of a total of 7,796 participants taking part in this research round, kidney function measures were obtained in 6,399 (82.1%). We excluded participants with a history of stroke or dementia (n = 367), and those with insufficient stroke or dementia screening (n = 39), leaving 5,993 participants for analyses. The Rotterdam Study was approved by the Medical Ethics Committee of the Erasmus MC according to the “Population Screening Act: Rotterdam Study”, executed by the Ministry of Health, Welfare and Sports of the Netherlands. All participants provided written informed consent.

Kidney function assessment

We determined serum creatinine levels using an enzymatic assay method, standardized to isotope-dilution mass spectrometry-traceable measurements [13]. Additionally, cystatin-C was determined using a particle-enhanced immunonephelometric assay. We calculated the estimated glomerular filtration rate (eGFR) using Chronic Kidney Disease Epidemiology Collaboration formulas [14] in three different ways in order to give a complete representation of kidney function. First, we calculated the eGFR based on creatinine (eGFRcr), which is routinely used in clinical practice. Second, we calculated the eGFR based on cystatin-C (eGFRcys), a more novel and accurate assessment of kidney function [15]. Third, we computed the eGFR based on both creatinine and cystatin-C (eGFRcrcys).

Assessment of stroke

Stroke was defined according to the established WHO-criteria [16]. Prevalent strokes were assessed at baseline during interview and verified with medical records. Participants were continuously monitored for incident stroke through direct computerized linkage of the study database with general practitioners’ medical files. Additionally, we monitored nursing home files and files from general practitioners of participants that moved out of the study district. All potential strokes were reviewed by research physicians and verified by an experienced stroke neurologist. We classified strokes as ischemic or hemorrhagic based on neuroimaging-reports. If neuroimaging was unavailable, the stroke was classified as unspecified [17, 18]. Follow-up was virtually complete (98.6% of potential person-years) until January 1, 2016.

Ascertainment of dementia

Participants were screened for dementia at baseline and subsequent center visits with the Mini-Mental State Examination and the Geriatric Mental Schedule organic level [19]. Those with a Mini-Mental State Examination score below 26 or Geriatric Mental Schedule score above 0 underwent further investigation and informant interview. All participants also underwent routine cognitive assessment [19]. In addition, the entire cohort was continuously monitored for dementia through linkage of the study database with medical records from general practitioners and the regional institute for outpatient mental health care. A consensus panel led by a consultant neurologist evaluated all information and established the final diagnosis according to standard criteria for dementia (DSM-III-R), Alzheimer’s disease (NINCDS– ADRDA), and vascular dementia (NINDS-AIREN). Follow-up was virtually complete (96.4% of potential person-years) until January 1, 2016.

Assessment of covariables

Information on relevant covariables was obtained by physical examination, blood sampling, interview, and medical records [12]. Covariables assessed by physical examination included body mass index (BMI), and systolic and diastolic blood pressure [20]. Using blood samples, we obtained information on diabetes mellitus (defined as a fasting glucose level of 7.0 mmol/L or higher (or a non-fasting glucose level of 11.0 mmol/L or higher if fasting samples were not available), or use of anti-diabetic medication [12]), serum total cholesterol, and serum high-density lipoprotein (HDL) cholesterol. By interview, we assessed smoking status (never, past, or current smoking), the use of blood pressure lowering or lipid-lowering medication, educational level (low (primary education or lower vocational education), intermediate (secondary education or intermediate vocational education), or high educational level (higher vocational education or university) [21], and physical activity (expressed as the metabolic equivalent of task (MET) [22]. History of coronary heart disease (CHD) was assessed at baseline [23].

Statistical analyses

We used the additive inverse of the three measures of kidney function and then standardized this value in order to be able to demonstrate the effect of kidney function on stroke and dementia per standard deviation decrease in measures of kidney function. Next, we used Cox-regression models to examine the association between the different measures of kidney function (per 1-SD decrease) and the risk of a combined endpoint of stroke and dementia (whichever came first). In the first model, we adjusted for age and sex. In the second model, we additionally adjusted for BMI, systolic and diastolic blood pressure, use of blood-pressure lowering medication, diabetes mellitus, history of CHD, educational level, physical activity, total cholesterol, HDL cholesterol, use of lipid-lowering medication, and smoking status. Second, we assessed the association of kidney function with risk of stroke and risk of dementia separately, using similar Cox regression models. To assess the association of kidney function with stroke and dementia independently, we censored individuals that developed incident dementia in the stroke-analyses (n = 65), and those with incident stroke in the dementia-analyses (n = 93). Third, we repeated these analyses for stroke-subtypes (ischemic, hemorrhagic) and dementia-subtypes (Alzheimer’s Disease, non-Alzheimer’s Disease). Fourth, we performed several sensitivity analyses in which we assessed the risk of stroke and dementia according to the clinically used cut-off of 60 mL/min/1.73 m² for eGFR, and we repeated all analyses stratified for age below and above 70 years. Missing values (<5.0%) were handled using a 5-fold multiple imputation algorithm. All analyses were performed using IBM SPSS Statistics version 21.0 (IBM Corporation, Armonk, New York).

RESULTS

Baseline characteristics of the study population are shown in Table 1. When comparing characteristics of persons in the study with those not in the study due to a lack of kidney function measures, we found that these were slightly older (mean age 73.2 versus 69.0), and more often women (65.7% versus 57.2%). None of the cardiovascular risk factors differed statistically significantly between these groups.

Table 1

Baseline characteristics of the study population

Sample size	5,993
Women	3428 (57.2)
Age, y	69.0±8.2
BMI, kg/m²	27.0±4.0
Level of education
Lower education	3,350 (55.9)
Middle education	1,781 (29.7)
Higher education	795 (13.3)
MET-hours, 1 kcal/kg/h^*	78 (56–84)
Systolic blood pressure, mmHg	143±21
Diastolic blood pressure, mmHg	77±11
Diabetes mellitus	636 (10.6)
Serum total cholesterol, mmol/L	5.8±1.0
Serum HDL cholesterol, mmol/L	1.4±0.4
Smoking
Past	2,896 (48.3)
Current	1,182 (19.7)
Use of blood pressure-lowering medication	1,967 (32.8)
Use of lipid-lowering medication	735 (12.3)
History of coronary heart disease	483 (8.1)
eGFRcr, mL/min/1.73 m²	75.0±14.8
eGFRcys, mL/min/1.73 m²	72.9±18.2
eGFRcrcys, mL/min/1.73 m²	74.9±15.7

Values are mean±standard deviation for continuous variables or frequency (percentage) for dichotomous variables. Data represent original data without imputed values. Missing values were present for body mass index (0.9%), blood pressure (0.4%), coronary heart disease (2.1%), level of education (1.1%), MET-hours (3.9%), total cholesterol (0.5%), HDL cholesterol (1.3%), diabetes mellitus (0.2%), lipid-lowering medication (4.1%), blood pressure– lowering medication (4.9%), and smoking (0.8%). BMI, body mass index, HDL, high density lipoprotein; eGFR, estimated glomerular filtration rate calculated from creatinine (cr), cystatin-C (cys), and a combination of both (crcys). ^*Median (interquartile range) presented.

During a mean follow-up time of 11.6 years (69,790 person-years), 1,360 participants suffered a stroke (n = 666) or developed dementia (n = 852). Of these, 158 were diagnosed with both stroke and dementia.

Lower values of eGFRcys and eGFRcrcys associated with a higher risk of the combined endpoint of stroke and dementia [Model 2-adjusted hazard ratio (HR) per 1-SD decrease in eGFRcys:1.12(95% confidence interval (CI):1.04;1.21), and in eGFRcrcys:1.09(95% CI:1.01;1.17)].

When assessing the risk of stroke separately, lower values of all three eGFR-calculations associated with higher risks of stroke (Table 2). The most prominent associations were found for eGFRcys and eGFRcrcys [Model 2-adjusted HR per 1-SD decrease:1.26(95% CI:1.13;1.41) and 1.21(95% CI:1.09;1.35)]. We found that these associations with any stroke were mainly driven by hemorrhagic stroke (Table 2). Additionally, eGFRcys and eGFRcrcys were also independently associated with a higher risk of ischemic stroke [HR per 1-SD decrease: 1.22 (95% CI: 1.07;1.39), and 1.16 (95% CI: 1.03;1.32)].

Table 2

Kidney function and the risk of stroke censored for incident dementia

		Any stroke		Ischemic stroke		Hemorrhagic stroke
		n_cases/n_atrisk = 601/5993		n_cases/n_atrisk = 445/5993		n_cases/n_atrisk = 67/5993
		Hazard Ratio (95% CI)		Hazard Ratio (95% CI)		Hazard Ratio (95% CI)
		Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
eGFRcr	Per 1-SD decrease	1.13 (1.03–1.25)	1.11 (1.01–1.23)	1.09 (0.97–1.22)	1.07 (0.95–1.20)	1.49 (1.12–1.98)	1.49 (1.12–1.98)
eGFRcys	Per 1-SD decrease	1.34 (1.20–1.49)	1.26 (1.13–1.41)	1.29 (1.14–1.47)	1.22 (1.07–1.39)	1.85 (1.35–2.55)	1.77 (1.28–2.44)
eGFRcrcys	Per 1-SD decrease	1.27 (1.14–1.41)	1.21 (1.09–1.35)	1.22 (1.08–1.38)	1.16 (1.03–1.32)	1.76 (1.31–2.37)	1.71 (1.27–2.31)

n, number of persons; CI, confidence interval; eGFR, estimated glomerular filtration rate calculated from creatinine (cr), cystatin-C (cys), and a combination of both (crcys); SD, standard deviation. Model 1: Adjusted for age and sex. Model 2: Adjusted for age, sex, body-mass index, systolic and diastolic blood pressure, use of blood-pressure lowering medication, diabetes mellitus, history of coronary heart disease, educational level, physical activity, total cholesterol, HDL cholesterol, use of lipid-lowering medication, and smoking status.

We found no association of kidney function with the risk of all-cause dementia, Alzheimer’s Disease, or non-Alzheimer Disease (Table 3).

Table 3

Kidney function and the risk of dementia censored for incident stroke

		All-cause dementia		Alzheimer’s Disease		Non-Alzheimer’s Disease
		n/N = 759/5993		n/N = 647/5993		n/N = 112/5993
		Hazard Ratio (95% CI)		Hazard Ratio (95% CI)		Hazard Ratio (95% CI)
		Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
eGFRcr	Per 1-SD decrease	0.98 (0.90–1.08)	0.99 (0.90–1.08)	0.98 (0.89–1.09)	0.98 (0.89–1.09)	1.00 (0.78–1.26)	1.02 (0.80–1.29)
eGFRcys	Per 1-SD decrease	0.98 (0.89–1.09)	1.01 (0.90–1.12)	0.96 (0.86–1.07)	0.98 (0.87–1.10)	1.15 (0.88–1.49)	1.14 (0.88–1.49)
eGFRcrcys	Per 1-SD decrease	0.97 (0.88–1.07)	0.98 (0.89–1.10)	0.95 (0.86–1.06)	0.97 (0.87–1.08)	1.08 (0.83–1.38)	1.08 (0.84–1.39)

n: number of persons, CI: confidence interval, eGFR: estimated Glomerular filtration rate calculated from creatinine (cr), cystatin-C (cys), and a combination of both (crcys), SD: standard deviation. Non-Alzheimer’s Disease includes vascular dementia, Parkinson’s disease dementia, and other/undetermined. Model 1: Adjusted for age and sex. Model 2: Adjusted for age, sex, body-mass index, systolic and diastolic blood pressure, use of blood-pressure lowering medication, diabetes mellitus, history of coronary heart disease, educational level, physical activity, total cholesterol, HDL cholesterol, use of lipid-lowering medication, and smoking status.

After assessing the risk of stroke and dementia using the commonly used eGFRcrcys cut-off of 60 mL/min/1.73m², we found that an eGFRcrcys below 60 mL/min/1.73m² was associated with stroke only [Model 2-adjusted HR:1.27(95% CI:1.03;1.57)]. After stratification by age, we found that all eGFR measures were associated with a higher risk of stroke in the age-group of 70 and above, and not with dementia (Supplementary Table 1).

DISCUSSION

We found that impaired kidney function was associated with a higher risk of stroke, but not with dementia during 12 years of follow-up.

Strengths of our study include the large sample size, the comprehensive kidney function assessment, and the long follow-up period for stroke and dementia. Regarding kidney function, it was reported that the equation including both creatinine and cystatin-C performs best in diagnosing chronic kidney disease [24]. Yet, in current clinical practice, eGFR based on creatinine is still mostly used followed by eGFR based on cystatin-C. In order to increase the generalizability of our results, we decided to calculate all three measures of eGFR. In general, we found that eGFR based on cystatin-C showed most prominent associations with our outcomes. Limitations of our study should also be mentioned. First, kidney function was assessed at one occasion only, precluding the possibility to examine the influence of fluctuations in kidney function on the development of stroke and dementia. Second, we did not have information on socioeconomic status which may be an important confounder of the association between kidney function and stroke and dementia. However, by using level of education as its proxy to adjust for in our analyses we largely overcome this given the close relation between the two [25].

In line with previous reports [6, 8], we found that impaired kidney function was consistently associated with stroke independent of cardiovascular risk factors, and in particular in persons aged 70 years or older. Yet, we found no statically significant association of kidney function with dementia. An important notion with regard to this finding is the censoring that we applied in the analyses on the risk of stroke and dementia separately. By censoring the stroke analyses for prevalent and incident dementia and censoring the dementia analyses for prevalent and incident stroke, the possibility that these associations were mediated by the other outcome is being limited, which is especially important in this case because of the close link between stroke and dementia terms of etiology and their frequent co-occurrence [11]. Our findings indicate that previously reported associations of kidney function with dementia may have been mediated by stroke, either clinically overt or covert. Yet, there is evidence for associations of chronic kidney disease with risk of dementia in specific Asian-populations burdened with vascular disease [26]. However, these may not be directly generalizable to our population. Of note, we observed slightly higher effect estimates when investigating non-Alzheimer’s Disease only, indicating a potential association of kidney dysfunction with non-Alzheimer’s Disease dementia types. Unfortunately, our analyses were insufficiently powered to study the subtypes of dementia in the non-Alzheimer Disease group separately.

Several hypotheses on the mechanism underlying the link between kidney dysfunction and cerebrovascular diseases have been postulated. First, the kidney and the brain are organs sharing a similar arterial structure with low vessel resistance and exposure to high-volume blood flow [27]. Their structural and functional integrity is critically dependent on adequate organ perfusion and intact vascular homeostasis. Therefore, it has been suggested that systemic vascular damage can cause simultaneous damage in both organs and in fact impaired kidney function may reflect vascular disease-related cerebral pathology [27, 28]. Second, the kidney is a key regulator of water and electrolyte balance, vascular tone, oxidative stress response, and cytokine production. Hence, impaired kidney function directly promotes disturbances in these processes which may lead to chronic inflammation and uremia-induced oxidative stress which are crucial factors in the development and progression of vascular damage leading to cardio- and cerebrovascular disease [7].

In summary, our results indicate that an impaired kidney function is an important determinant of the risk of stroke, but not of dementia.

Footnotes

ACKNOWLEDGMENTS

The dedication, commitment, and contribution of inhabitants, general practitioners, and pharmacists of the Ommoord district to the Rotterdam Study are gratefully acknowledged.

Sources of support: The Rotterdam Study is supported by the Erasmus MC; Erasmus University Rotterdam; Netherlands Organisation for Scientific Research (NWO); Netherlands Organisation for Health Research and Development (ZonMW); Research Institute for Diseases in the Elderly (RIDE); Netherlands Genomics Initiative; Ministry of Education, Culture and Science; Ministry of Health, Welfare and Sports; European Commission (DG XII); and Municipality of Rotterdam.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Hill

, Fatoba

, Oke

, Hirst

, O’Callaghan

, Lasserson

, Hobbs

(2016) Global prevalence of chronic kidney disease - a systematic review and meta-analysis. PLoS One 11, e0158765.

Mann

, Gerstein

, Pogue

, Bosch

, Yusuf

(2001) Renal insufficiency as a predictor of cardiovascular outcomes and the impact of ramipril: The HOPE randomized trial. Ann Intern Med 134, 629–636.

, Chertow

, Fan

, McCulloch

, Hsu

(2004) Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 351, 1296–1305.

Etgen

, Chonchol

, Förstl

, Sander

(2012) Chronic kidney disease and cognitive impairment: A systematic review and meta-analysis. Am J Nephrol 35, 474–482.

Shen

, Ruan

, Yu

, Sun

(2017) Chronic kidney disease-related physical frailty and cognitive impairment: A systemic review. Geriatr Gerontol Int 17, 529–544.

Deckers

, Camerino

, van Boxtel

MPJ

, Verhey

FRJ

, Irving

, Brayne

, Kivipelto

, Starr

, Yaffe

, de Leeuw

(2017) Dementia risk in renal dysfunction A systematic review and meta-analysis of prospective studies. Neurology 88, 198–208.

Bugnicourt

J-M

, Godefroy

, Chillon

J-M

, Choukroun

, Massy

(2013) Cognitive disorders and dementia in CKD: The neglected kidney-brain axis. J Am Soc Nephrol 24, 353–363.

Masson

, Webster

, Hong

, Turner

, Lindley

, Craig

(2015) Chronic kidney disease and the risk of stroke: A systematic review and meta-analysis. Nephrol Dial Transplant 30, 1162–1169.

Pendlebury

, Rothwell

(2009) Prevalence, incidence, and factors associated with pre-stroke and post-stroke dementia: A systematic review and meta-analysis. Lancet Neurol 8, 1006–1018.

10.

Snowdon

, Greiner

, Mortimer

, Riley

, Greiner

, Markesbery

(1997) Brain infarction and the clinical expression of Alzheimer disease: The Nun Study. JAMA 277, 813–817.

11.

Ivan

, Seshadri

, Beiser

, Au

, Kase

, Kelly-Hayes

, Wolf

(2004) Dementia after stroke: The Framingham Study. Stroke 35, 1264–1268.

12.

Hofman

, Brusselle

, Murad

, van Duijn

, Franco

, Goedegebure

, Ikram

, Klaver

, Nijsten

, Peeters

(2015) The Rotterdam Study: 2016 objectives and design update. Eur J Epidemiol 30, 661–708.

13.

Sedaghat

, Hoorn

, van Rooij

FJA

, Hofman

, Franco

, Witteman

JCM

, Dehghan

(2013) Serum uric acid and chronic kidney disease: The role of hypertension. PLoS One 8, e76827.

14.

Inker

, Schmid

, Tighiouart

, Eckfeldt

, Feldman

, Greene

, Kusek

, Manzi

, Van Lente

, Zhang

(2012) Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med 367, 20–29.

15.

Dharnidharka

, Kwon

, Stevens

(2002) Serum cystatin C is superior to serum creatinine as a marker of kidney function: A meta-analysis. Am J Kidney Dis 40, 221–226.

16.

Truelsen

, Begg

, Mathers

(2002) Global burden of cerebrovascular disease in the year 2000. World Health Organization, Geneva.

17.

Wieberdink

, Ikram

, Hofman

, Koudstaal

, Breteler

(2012) Trends in stroke incidence rates and stroke risk factors in Rotterdam, the Netherlands from 1990 to 2008. Eur J Epidemiol 27, 287–295.

18.

Bos

, Koudstaal

, Hofman

, Ikram

(2014) Modifiable etiological factors and the burden of stroke from the Rotterdam study: A population-based cohort study. PLoS Med 11, e1001634.

19.

de Bruijn

, Bos

, Portegies

, Hofman

, Franco

, Koudstaal

, Ikram

(2015) The potential for prevention of dementia across two decades: The prospective, population-based Rotterdam Study. BMC Med 13, 132.

20.

Bos

, Koudstaal

, Hofman

, Breteler

(2007) Decreased glomerular filtration rate is a risk factor for hemorrhagic but not for ischemic stroke: The Rotterdam Study. Stroke 38, 3127–3132.

21.

de Bruijn

RFAG

, Bos

, Portegies

MLP

, Hofman

, Franco

, Koudstaal

, Ikram

(2015) The potential for prevention of dementia across two decades: The prospective, population-based Rotterdam Study. BMC Med 13, 132.

22.

de Bruijn

RFAG

, Schrijvers

EMC

, de Groot

, Witteman

JCM

, Hofman

, Franco

, Koudstaal

, Ikram

(2013) The association between physical activity and dementia in an elderly population: The Rotterdam Study. Eur J Epidemiol 28, 277–283.

23.

Bos

, Koudstaal

, Hofman

, Witteman

JCM

, Breteler

MMB

(2006) Uric acid is a risk factor for myocardial infarction and stroke: The Rotterdam study. Stroke 37, 1503–1507.

24.

Inker

, Schmid

, Tighiouart

, Eckfeldt

, Feldman

, Greene

, Kusek

, Manzi

, Van Lente

, Zhang

, Coresh

, Levey

, Investigators

C-E

(2012) Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med 367, 20–29.

25.

Winkleby

, Jatulis

, Frank

, Fortmann

(1992) Socioeconomic status and health: How education, income, and occupation contribute to risk factors for cardiovascular disease. Am J Public Health 82, 816–820.

26.

Miwa

, Tanaka

, Okazaki

, Furukado

, Yagita

, Sakaguchi

, Mochizuki

, Kitagawa

(2014) Chronic kidney disease is associated with dementia independent of cerebral small-vessel disease. Neurology 82, 1051–1057.

27.

O’Rourke

, Safar

(2005) Relationship between aortic stiffening and microvascular disease in brain and kidney: Cause and logic of therapy. Hypertension 46, 200–204.

28.

Safar

, Boudier

(2005) Vascular development, pulse pressure, and the mechanisms of hypertension. Hypertension 46, 205–209.