In this study, we found that eGFR was moderately correlated with cognitive
function, cortical thickness, and GM volume. Study participants with reduced kidney
function, had impaired cognitive performance, less cortical thickness, and smaller
hippocampal volume and GM volume than did subjects with normal renal function. In
further regression analysis, we found that renal function and age were major
contributing factors for cortical thickness, GM and hippocampal volume after
controlling for confounding factors.
A few studies have reported brain atrophy and WMH burden in patients with
CKD or renal failure (9, 10, 28-31). A meta-analysis of numerous structural and
functional neuroimaging studies that examined children and adults with CKD
identified several clear trends, including cerebral atrophy, WMH, cerebral infarction,
microbleeding, and cerebral blood flow pattern with affective disorders (31). In the
retrospective population-based Rotterdam Scan Study, subjects with low eGFR had
smaller brain and deep WM volumes and more WMH (30). No close relationship was
found between eGFR and GM or WM volume, which is not fully consistent with our
results (30). The demographic characteristics of participants in the Rotterdam Scan
Study differed from those in the present study subjects, such as lower percentage of
diabetes and hypertension, and the higher ratio of male patients to female patients
than those in the present population. Nevertheless, the Rotterdam Scan Study lacked
neuropsychological assessment, which precluded correlation of brain measurements
with cognitive performance. Most previous imaging analyses were focus on grading
of WMH and detection of ischemic infarction, which were based on qualitative visual
rating, rather than objective quantitative measurement. However, little about
quantitative measurement of brain structure in patients with CKD was reported.
Furthermore, the present study quantitatively showed the GM, WM and hippocampus
volumes decreased by 6.2%, 8.9% and 10.7%, respectively, in patients with CKD
compared with control subjects. The hippocampal volume seems to be more affected
by eGFR and age than global GM and WM volume.
In present study, significant cortical thinning in bilateral superior temporal sulci
and medial lingual gyri was found in these patients with CKD and they showed
21
lower global cognition scores and poorer performance in working memory (forward
digit span) than those in the control group. In patients with mild cognitive impairment
or dementia, cortical thinning has been found in association with cardiovascular risk
factors, such as hypertension and diabetes (32). However, eGFR and age were
independently related to GM volume and cortical thickness in analyses controlling for
the vascular component in the present study. The CKD group also had a significant
reduction in hippocampal volume. Hippocampal atrophy has been shown to be an
effective marker differentiating people with normal and impaired cognition,
regardless of WMH and lacunar infarcts (33). Thus, renal function seems to be an
independent factor affecting cortical atrophy and thinning via an uncertain mechanism
rather than cerebrovascular risk factors. Nevertheless, the difference in WMH
between patients with CKD and control subjects didn’t reach significant level in this
study. This result may be explained by the more frequent occurrence of WMH in
patients with advanced CKD and those receiving hemodialysis, whereas about half of
patients in our study had moderate CKD.
In line with previous studies (2, 5, 34), the present study also showed that
patients with CKD had poorer global cognitive function, executive function, verbal
and working memory than the control subjects. A recent meta-analysis found that
both cross-sectional and longitudinal studies have demonstrated a significantly
increased risk of cognitive impairment in patients with CKD (35). Furthermore,
another longitudinal cohort study conducted in Japan found that CKD was associated
with the risk of dementia in patients with v independent of vascular components (36).
These results are consistent with the hypothesis of a connection between renal
impairment and dementia risk.
The close relationship between cardiovascular risk factors and dementia with
small vessel disease is well established (37). Recently, CKD was found to be related
to neurological disorders and this brain–renal connection is thought to involve small
vessel disease in kidney and brain, based on hemodynamic similarities (38). The main
pathologic vascular feature of CKD and albuminuria, implied impaired kidney
vascular integrity, which might also be found in other organs with similar vascular
bed. Recent study showed lower eGFR was independently associated with lower
cerebral blood flow. The impaired cerebral autoregulation, probably leading to
23
hypoperfusion, was thought to increased risk of dementia (39). Furthermore it has
been suggested that various toxins have been involved in the pathogenesis of
cognitive impairment in uremic patients(40). Oxidative radicals, elevated serum
homocysteine level and inflammation have also been related to cognitive impairment
in dialysis patients. These toxic substance could lead to vascular endothelial
dysfunction and thus aggravate atherosclerosis and increase risk of dementia(41).
These studies implied that impaired renal function, as measured by decreased eGFR,
with alterations in water and electrolytes balance, accumulation of vasoactive species,
and chronic inflammation is related to cerebral small vessel disease, independent of
cardiovascular risk factors (30). Such findings are consistent with our study results:
impaired eGFR is correlated with reduced GM volume and global cognitive
performance, independent of other cardiovascular disease. In contrast, some study
showed that mild CKD was associated with an increased risk of Alzheimer disease
(AD) (36). The longitudinal BRain IN Kidney disease (BRINK) study found smaller
GM volume in region of interests (ROIs) associated with both AD (tempo-parietal
areas) and vascular cognitive impairment (VCI )(frontal lobes), worse cognition
function than patients without CKD(42, 43). Neuropathology Group of the Medical
Research Council Cognitive Function and Ageing Study (MRC CFAS) reported both
Alzheimer type and vascular pathology were found in most elder sample with
cognitive declineby necropsy(44). Thus the possibilities of mixed AD and VCI in
CKD group should be also taken into account.
The strengths of this study include the examination of patients with moderate
CKD and the use of automated MRI analysis, which make it possible to accurately
quantify cortical thinning and GM and WM atrophy and to investigate subcortical
WM lesions. The associations between CKD and cognitive performance, cortical GM
volume, cortical thickness, as well as eGFR were clearly demonstrated. To my
knowledge, this study is the first study on cognitive impairment in patients with CKD
based on comprehensive quantitative measurement of brain structure and renal
function.
This study has some limitations. First, diabetes was more prevalent among
patients with CKD than among the control subjects, reflecting a similar difference in
the Taiwanese population (the prevalence of diabetes is about 15% in individuals with
25
and 4% in those without CKD) (45). To eliminate the effect of this difference,
analysis with multiple linear regression model adjusted for age, education, diabetes,
and other comorbidities was applied. Second, patients in the present study were
recruited from a tertiary medical center, rather than from a community-based hospital,
which may have led to selection bias. Third, smoking is indeed a major cardiovascular
factor and it was not controlled in the present study. This might cause some bias.
Forth, the present study did not perform CSF markers check-up and PET isotope study,
the possibilities of prodrome of AD in CKD group cannot be excluded. Finally,
though cortical thickness, GM volume, and MMSE score were lower in the CKD
group than in the control subjects, the cause–effect relationships between renal
function, brain morphometric features and cognitive performance remain uncertain
because of the cross-sectional study design and lack of longitudinal follow-up.