Type 2 Diabetes Mellitus and Impaired Renal Function Are Associated With Brain Alterations and Poststroke Cognitive Decline
Background and Purpose—Type 2 diabetes mellitus (T2DM) is associated with diseases of the brain, kidney, and vasculature. However, the relationship between T2DM, chronic kidney disease, brain alterations, and cognitive function after stroke is unknown. We aimed to evaluate the inter-relationship between T2DM, impaired renal function, brain pathology on imaging, and cognitive decline in a longitudinal poststroke cohort.
Methods—The TABASCO (Tel Aviv brain acute stroke cohort) is a prospective cohort of stroke/transient ischemic attack survivors. The volume and white matter integrity, ischemic lesions, and brain and hippocampal volumes were measured at baseline using 3-T MRI. Cognitive tests were performed on 507 patients, who were diagnosed as having mild cognitive impairment, dementia, or being cognitively intact after 24 months.
Results—At baseline, T2DM and impaired renal function (estimated creatinine clearance [eCCl] <60 mL/min) were associated with smaller brain and hippocampal volumes, reduced cortical thickness, and worse white matter microstructural integrity. Two years later, both T2DM and eCCl <60 mL/min were associated with poorer cognitive scores, and 19.7% of the participants developed cognitive decline (mild cognitive impairment or dementia). Multiple analysis, controlling for age, sex, education, and apolipoprotein E4, showed a significant association of both T2DM and eCCl <60 mL/min with cognitive decline. Having both conditions doubled the risk compared with patients with T2DM or eCCl <60 mL/min alone and almost quadrupled the risk compared with patients without either abnormality.
Conclusions—T2DM and impaired renal function are independently associated with abnormal brain structure, as well as poorer performance in cognitive tests, 2 years after stroke. The presence of both conditions quadruples the risk for cognitive decline. T2DM and lower eCCl have an independent and additive effect on brain atrophy and the risk of cognitive decline.
Type 2 diabetes mellitus (T2DM) affects >415 million adults worldwide.1 T2DM is associated with incident cognitive impairment, dementia, and Alzheimer disease as a result of cerebrovascular disease and neurodegenerative processes.2 T2DM results in endothelial dysfunction affecting the kidney and the brain. Chronic kidney disease (CKD), affecting 13% of adults in the United States,3 is also associated with structural brain changes and cognitive impairment in the elderly.4
We have recently shown that impaired renal function, expressed as estimated creatinine clearance (eCCl) <60 mL/min, is associated with larger white matter (WM) lesion, a marker for cerebromicrovascular disease, as well as brain and hippocampal atrophy, and is a predictor for cognitive decline 2 years after stroke/transient ischemic attack (TIA).5 As both hyperglycemia and chronic renal failure frequently accompany cardiovascular disease or stroke,6 we have investigated whether patients with impaired renal function and T2DM have increased risk for brain microstructural changes and poststroke cognitive changes.
Included in the present study were survivors of mild-to-moderate first acute ischemic stroke/TIA with a total National Institutes of Health Stroke Scale (NIHSS)7 <17, who have participated in the prospective TABASCO study (Tel Aviv brain acute stroke cohort).8
We excluded patients whose stroke resulted from trauma or invasive brain procedures, hemorrhagic stroke, cognitive impairment before the stroke (determined by Informant Questionnaire on Cognitive Decline in the Elderly9 score ≥3.3), and severe aphasia or disability, which making the possibility of continuous follow-up unlikely, as well as subjects with type 1 diabetes mellitus (2 subjects). The neurological assessment included verification of stroke pathogenesis, NIHSS (ranges from 0 [no stroke] to 42, where NIHSS >20 is a severe stroke), and neuroimaging. Vascular risk factors were assessed according to Framingham Stroke Risk Profile score.10
Standard Protocol Approvals, Registrations, and Patient Consents
All participants signed informed consent forms and approved by the local ethics committee.
Definition of T2DM
T2DM was defined based on patient report, hemoglobin A1C (A1C) level >6.5%, and the prior use of antihyperglycemic agents. Absence of T2DM was defined as none of the above. All patients with an A1C ≥6.5% had a prior self-reported diagnosis of T2DM.
An A1C level <5.7% is considered normal, between 5.7% and 6.4% signals pre-diabetes, and T2DM is diagnosed when the A1C is >6.5%.
Duration of T2DM was also determined by self-reports.
Estimation of Renal Function
Baseline renal function was estimated as creatinine clearance (eCCl) using the Cockcroft–Gault formula: CCl=(140−age)×(weight in kg)/(serum creatinine×72)×(0.85 for women).11 Participants whose eCCl <60 mL/min were considered as having CKD.12
Baseline and Follow-Up Cognitive Assessments
Patients completed a baseline neuropsychological assessment including the Montreal Cognitive Assessment13 and the NeuroTrax computerized cognitive testing (NeuroTrax, Bellaire, TX).14 These comprehensive computerized evaluations were repeated 6, 12, and 24 months after the event. The average of the 6 index scores (memory, executive functions, visuospatial perception, verbal function, attention, and motor skills) was computed as the total cognitive score. Data for each outcome parameter were stratified by age and education to give a distribution with a mean of 100 and a SD of 15. To minimize practice effects, alternate forms of the NeuroTrax tests were administered on the serial testing sessions.
Criteria for Cognitive Impairment
Patients with cognitive impairment were diagnosed as having either mild cognitive impairment (MCI) or dementia. To diagnose MCI, the modified Petersen criteria15 were applied: the subject had to complain and be impaired on at least 1 cognitive domain (≥1.5 SD) compared with age- and education-matched published norms on the Montreal Cognitive Assessment score16,17, to have no impairment of basic functional activities, and to not fulfill the DSM IV-TR criteria (Diagnostic and Statistical Manual of Mental Disorders) for dementia (American Psychiatric Association, Diagnostic and Statistical Manual for Mental Disorders, Fourth Edition, 2000). The norms for the NeuroTrax computerized cognitive testing were previously described.17
Patients who could not complete the Montreal Cognitive Assessment or whose NeuroTrax test was below ≥1.5 SD in follow-up examinations, as well as those with subjective cognitive complaints, were referred to an experienced cognitive neurologist (A.D.K.) for evaluation. Assessments were further reviewed by a consensus forum to determine whether the participant had dementia or MCI. The forum included the assessor, 3 senior neurologists specializing in memory disorders, and a neuropsychologist.
MRI and statistical analyses are provided in the online-only Data Supplement.
A total of 575 consecutive eligible cognitive intact patients who were admitted to the Department of Emergency Medicine at Tel Aviv Medical Center between April 1, 2008, and December 1, 2014, within 72 hours from onset of symptoms of TIA or stroke were initially evaluated. Of these, cognitive assessments at baseline and 2 years later were available for 507 subjects, and these were included in the final analysis (Table I in the online-only Data Supplement). The mean age was 67.4±9.7 years; 59.4% were men; 154 (30.4%) were diagnosed with TIA and 152 diagnosed with T2DM (30%). Brain MRI scans at baseline were available for 392 subjects. Participants with MRI were younger (66.8+9.7 versus 69.2+9.6) and had lower prevalence of cardiac disease (13.5% versus 30.2%) than those not included in the MRI study. No differences in cognitive results observed between participants included and not included in the MRI study.
Stroke pathogeneses (based on TOAST criteria (Trial of ORG 10172 in Acute Stroke Treatment),18 a classification of 5 subtypes of ischemic stroke) were as follows: 266 lacunar stroke (52.5%), 71 cardioembolic stroke (14%), 53 large-artery atherosclerotic stroke (10.5%), and 117 stroke of other or undetermined pathogenesis (23%).
Participants with T2DM had a higher body mass index and Framingham risk score for stroke, had elevated fasting blood glucose, A1C, and total cholesterol (as well as lower levels of high-density lipoprotein cholesterol), and were more likely to experience hypertension and dyslipidemia. Participants with T2DM were more likely to show new lesions on MRI although these were of smaller size (Table I in the online-only Data Supplement).
T2DM, Renal Function, and Cognitive Performance at Baseline
T2DM was significantly associated with poor performance in motor skills scores (Table 1; Table III in the online-only Data Supplement) while poor renal function (eCCl <60 mL/min, observed in 192 [37.9% of the participants]) was associated with poor performance in executive functions, visuospatial, and total cognitive scores (Table II in the online-only Data Supplement). The association disappeared, however, when adjusted for age, sex, education, and apolipoprotein E ε4 allele status (Table 1).
T2DM, Renal Function, and Long-Term Cognitive Performance Poststroke
Table II in the online-only Data Supplement presents the cognitive results of the longitudinal cognitive tests at admission, 6-, 12-, and 24-months after stroke. T2DM was independently associated with worse results in memory, executive functions, attention, and total cognitive scores after adjustment for age, sex, education, and apolipoprotein E ε4 status, which developed in a delayed manner after their stroke. Poor renal function was associated with worse results in memory, visuospatial, and total cognitive scores 2 years poststroke after adjustment (Table 1) and developed short term after their stroke.
Higher A1C was associated with poor performance in memory, executive functions, attention, and total cognitive scores 2 years later after adjustment for age, sex, education, and vascular risk factors, including body mass index, hypertension, dyslipidemia, smoking, and apolipoprotein E4 (Table III in the online-only Data Supplement). No association was observed between A1C and cognitive scores at baseline.
Figure 1A shows the baseline A1C levels according to the cognitive status 2 years after the stroke/TIA.
No differences in A1C or eCCl values observed among those developing MCI and dementia.
Univariate and Multivariate Predictors of Cognitive Impairment
During 2-year follow-up period, 100 participants (19.7%) developed clinically significant cognitive impairment. Of these, 13 patients (2.6%) developed dementia and 87 patients (17.2%) developed MCI. No significant differences in A1C or eCCl values were observed between patients with MCI and dementia, and they were, therefore, grouped together as a cognitive impaired group.
Univariate and multivariate predictors for cognitive impairment are shown in Table 2. Univariate predictors included age (≥65), lower education (<12 years), new lesions in MRI, stroke severity (measured by NIHSS score on admission), hypertension, T2DM, dyslipidemia, elevated body mass index (>25 kg/m2), eCCl <60 mL/min, A1C ≥6.5%, and A1C ≥5.7% (considered as a pre-diabetic state). The lower A1C cutoff, including subjects with diabetes mellitus and pre-diabetes, was a better predictor for cognitive impairment than the higher A1C cutoff including only subjects with diabetes mellitus (odds ratio=2.67; 95% confidence interval, 1.38–5.14).
In a multivariate analysis for prediction of cognitive impairment, predictors retained were age ≥65, years of education <12, eCCl <60 mL/min, and T2DM. Patients with both eCCl <60 mL/min and T2DM have doubled their risk to develop cognitive impairment within 2 years from their index event compared with patients with T2DM or eCCl <60 mL/min alone and had ≈4-fold risk compared with patients without T2DM or eCCl <60 mL/min (odds ratio=3.8; 95% confidence interval, 1.52–9.52; ie, presence of both CKD and T2DM is equal to the risk contributed by increasing age; Table 2). The presence of new lesions, as well as stroke severity, did not retain as predictors of cognitive impairment in the multivariate model. Figure 2A and 2B demonstrates the appearance of dementia and a decline in the computerized cognitive score based on baseline T2DM and eCCl <60 mL/min status.
Of note, T2DM and CKD (eCCl <60 mL/min) were not associated (χ2=1.075; P=0.300) in our cohort.
Association Between T2DM, Renal Function, and MRI Measurements
T2DM was associated with thinner frontal, parietal, temporal, and occipital cortex, lower total gray matter (GM) and hippocampal volumes, worse global normal appearing WM microstructural integrity, the presence of ischemic lesion and lacunes but not with intracranial volume (ICV), WM lesion volume, or other measures small vessel disease (Table I in the online-only Data Supplement).
A longer duration of T2DM was associated with lower hippocampal volume normalized to ICV (r=−0.286; P=0.023; Figure IA in the online-only Data Supplement) but not with cognitive results. This association was diminished in subjects with both T2DM and CKD (Figure IB in the online-only Data Supplement). CKD (eCCl <60 mL/min) was associated with thinner frontal, parietal, temporal, and occipital cortex (P=0.001, P<0.001, P<0.001, P<0.001, respectively), lower ICV, total GM and hippocampal volumes normalized to ICV (P<0.001 for all; Figure IC in the online-only Data Supplement), larger WM lesion volume (P<0.001), worse global normal appearing WM microstructural integrity (P<0.001, P<0.001, respectively) but not with the presence of ischemic lesion and lacunes, enlarged perivascular spaces, or the presence of microbleeds.
A higher A1C was independently associated with larger ischemic lesion volume (r=0.165; P=0.024) and more brain atrophy (Table III in the online-only Data Supplement) after adjustment for age, sex, and education.
In addition, higher A1C was associated with stroke severity, as measured by NIHSS scoring at admission (r=0.155; P=0.003).
Effect of T2DM and Renal Function on MRI Measurements and Cognitive Performance
To study whether both impaired renal function and T2DM are associated with increased risk for brain microstructural changes and cognitive changes poststroke, we divided the study population into 4 subgroups: (1) nonimpaired renal function and no T2DM (n=174); (2) nonimpaired renal function and T2DM (n=69); (3) impaired renal function and no T2DM (n=100); and (4) impaired renal function and T2DM (n=49). Examples of hippocampal volume and cognitive results according to the group stratification are shown in Figure 1B and represent results from subjects who had both data on T2DM and eCCL at admission (n=448, of them, 392 had also diffusion tensor imaging-MRI data). As expected, groups 3 and 4 (eCCl <60 mL/min) were older than groups 1 and 2 and showed a clear trend for brain atrophy: frontal cortex width and hippocampal volume, normalized to ICV (P<0.001), WM integrity and volume (P≤0.001), as well as cognitive changes within 2 years after stroke/TIA (P<0.001, P=0.001, P<0.001, P=0.018, P<0.001, P<0.001 for memory, executive function, visuospatial, attention, total cognitive score, and Montreal Cognitive Assessment score, respectively) compared with groups 1 and 2. Group 4 (impaired renal function and T2DM) had significantly greater brain atrophy and worse cognitive scores 2 years after stroke. Because eCCl and the computerized cognitive scores are both corrected for age, no further adjustment for age was done.
Table 1 shows the magnitude of change in the β coefficients for T2DM or eCCl against the cognitive scores caused after addition of relevant MRI variables to the model after adjustment for age, sex, education, and apolipoprotein E4. The addition of total GM, WM, or hippocampal volume considerably attenuated the β coefficients of T2DM with cognitive scores (by 30.5%, 31%, 30.3%, respectively). However, the inclusion of new infarcts did not change these coefficients. We used mediation analysis (by PROCESS tool for SPSS19: a regression-based approach that estimates the indirect effect of an independent variable on a dependent variable via a mediator) to investigate the potential mediation effect. Thus, we suggest that the total GM, WM, or hippocampal volume mediated the relationship between T2DM and cognitive performance 2 years after stroke.
As for renal function, the addition of total GM, WM, or hippocampal volume slightly attenuated the β coefficients of eCCl with cognitive scores (by 11.2%, 14.8%, 12.2%, respectively). However, the inclusion of new infarcts did not change these coefficients. Thus, we suggest that the total GM, WM, or hippocampal volume partly mediates the relationship between eCCl and cognitive performance.
In this study, we found evidence that both, T2DM and impaired renal function, are independently associated with baseline imaging markers of damaged WM and GM integrity, as well as poorer performance on cognitive tests 2 years after stroke/TIA. Our results demonstrate an additive effect of T2DM and CKD on cognitive decline because presence of both conditions doubles the risk for poststroke cognitive decline relative to the presence of one. When compared with having neither T2DM nor CKD, the risk of cognitive decline is quadrupled.
From the longitudinal cognitive results, it seems that the full impact of T2DM seems late only 2 years after the index stroke whereas the impact of CKD seems immediately, probably not related to stroke, but reached its full impact after 2 years (thus the stroke may hasten its impact).
Of note, no differences in A1C or eCCl values were observed among those developing MCI and dementia, implying a threshold effect of T2DM or CKD, accelerated by the acute infarct.
Importantly, in patients with CKD, we observed impairment in memory and visuospatial scores; whereas in patients with T2DM, we observed impaired in memory, executive function, and attention scores. This could be explained by the sensitivity of executive functioning and processing speed to delicate diffuse vascular-related deterioration in WM integrity. Vascular disease could affect cognition not only through the influence on subcortical connections and WM changes but also aggravating cortical atrophy, suggesting a synergistic effect.
We are not aware of any previous study looking at the effect of both renal functions and glycemic status on poststroke cognitive functioning and regional brain microstructural changes. Previous studies have implicated both diabetes mellitus and impaired renal function as associated with poor stroke outcomes, but none have focused on the interaction of these comorbidities or their long-term effects on cognitive impairment in patients with stroke.
The mechanism linking T2DM, CKD, and cognitive decline in patients recovering from stroke is still unclear. It is possible that disruption in neuronal metabolic functions occurring mainly in brain regions most susceptible to cognitive impairment and less resilient to metabolic changes, found in both conditions, and may lead to worsened outcomes. This is reflected by cortical and hippocampal atrophy and loss of WM integrity, known to be promoted by hypoxic damage mediated by T2DM. High glucose concentrations promote accumulation of aberrant metabolites and oxidative stress, giving rise to neuronal dysfunction and apoptosis.20 Another possible pathway may be mediated by abnormalities in glucose metabolism in particular with regard to amyloid and tau metabolism, key elements also in the pathogenesis of Alzheimer disease.21
Although T2DM was associated in our study, as well as in previous studies,22 with cognitive decline and cortical atrophy, we have not examined an interaction with disease treatment and glycemic control. In the ACCORD MIND study (Action to Control Cardiovascular Risk in Diabetes Memory in Diabetes Study),23 intensive glucose-lowering treatment was associated with less brain atrophy. Cognitive outcomes, however, were not different versus standard treatment, which may imply structural changes preceding the functional changes. In a recently published study, pioglitazone, a commonly used antihyperglycemic agent, was shown to prevent recurrence of stroke in nondiabetic insulin resistant subjects.24 These results are aligned with the increased risk for cognitive impairment in the subgroup of subjects with elevated A1C. Furthermore, our results identified diabetic patients with chronic renal impairment as having a significantly increased risk for cognitive decline. As it stands to reason that both T2DM and CKD are being addressed medically after the ischemic event, prestroke aggressive treatment may impact poststroke cognitive performances. The latter is supported by structural alternations we found in patients with T2DM and CKD at baseline. As prestroke brain reserve seems to be impaired in these patients, the adding effect of acute infarct, although mild or even transient, may be the last straw in the cascade leading to vascular cognitive decline.
Pioglitazone, a peroxisome proliferator-activated receptor γ agonists, was found to reduce recurrent stroke and major vascular events in ischemic stroke patients with insulin resistance, pre-diabetes, and diabetes mellitus.25 It also exert neuroprotective effects in the brain and was found to have a protective effect against dementia in patients with diabetes mellitus.26 Whether pioglitazone, or other antiglycemic agents, have a role in preventing vascular dementia in poststroke patients is yet to be investigated in clinical trials.
Some limitations of this study should be taken into consideration; enrolling only post-mild stroke or TIA patients may affect the generalizability of our results. Participants were also receiving multiple medications (antihyperglycemic agents, statins, blood pressure–lowering medication etc). We partially accounted for this potentially masking effect of treatment by using dichotomous variables. However, we were not able to take into account the degree of success of these pharmacological interventions. Our study did not take into account other parameters of metabolic control, such as fasting blood glucose, parameters of insulin resistance, and hypoglycemic events. Another important limitation is determining duration of diabetes mellitus based on self-reporting because a pre-diabetic state may exist even before those self-reports and impact cognition.
As the general population ages, prevention of cognitive decline becomes a key element for future trials and novel medications. During the past decade, many new therapies for diabetes mellitus were introduced based on clinical trials measuring change in A1C, mortality, and cardiovascular events. The impact of these drugs on cognition, however, was not directly evaluated and does not play a role in current treatment decision algorithms. Risk prognostication and advanced brain imaging may guide the selection of appropriate therapy to stop or delay vascular cognitive decline in patients with T2DM.
T2DM and impaired renal function are independently associated with abnormal brain structure, as well as poorer performance in cognitive tests, 2 years after stroke. The presence of both conditions quadruples the risk for cognitive decline. T2DM and lower eCCl have an independent and additive effect on brain atrophy and the risk for cognitive decline.
Sources of Funding
This study is supported by grants RAG11482 from the American Federation for Aging Research (to Dr Ben Assayag) and grant 2011344 from the US–Israel Bi-national Science Foundation (to Dr Ben Assayag). These funding agencies had no role in the conduct and publication of this study.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.117.017709/-/DC1.
- Received April 13, 2017.
- Revision received June 26, 2017.
- Accepted June 28, 2017.
- © 2017 American Heart Association, Inc.
- Federation ID
- Moran C,
- Phan TG,
- Chen J,
- Blizzard L,
- Beare R,
- Venn A,
- et al
- Sink KM,
- Divers J,
- Whitlow CT,
- Palmer ND,
- Smith SC,
- Xu J,
- et al
- Auriel E,
- Kliper E,
- Shenhar-Tsarfaty S,
- Molad J,
- Berliner S,
- Shapira I,
- et al
- Brott T,
- Adams HP Jr,
- Olinger CP,
- Marler JR,
- Barsan WG,
- Biller J,
- et al
- Wolf PA,
- D’Agostino RB,
- Belanger AJ,
- Kannel WB
- Doniger GM,
- Dwolatzky T,
- Zucker DM,
- Chertkow H,
- Crystal H,
- Schweiger A,
- et al
- Arnáiz E,
- Almkvist O,
- Ivnik RJ,
- Tangalos EG,
- Wahlund LO,
- Winblad B,
- et al
- Pendlebury ST,
- Mariz J,
- Bull L,
- Mehta Z,
- Rothwell PM
- Adams HP Jr,
- Bendixen BH,
- Kappelle LJ,
- Biller J,
- Love BB,
- Gordon DL,
- et al
- Hayes AF
- Launer LJ,
- Miller ME,
- Williamson JD,
- Lazar RM,
- Gerstein HC,
- Murray AM,
- et al
- Lee M,
- Saver JL,
- Liao HW,
- Lin CH,
- Ovbiagele B
- Chou PS,
- Ho BL,
- Yang YH