Association of Kidney Function Measures With Signs of Neurodegeneration and Small Vessel Disease on Brain Magnetic Resonance Imaging: The Atherosclerosis Risk in Communities (ARIC) Study.

RATIONALE & OBJECTIVE: Chronic kidney disease (CKD) is a risk factor for cognitive decline, but evidence is limited on etiology and morphological manifestation in the brain. We evaluated the association of estimated glomerular filtration rate (eGFR) and urine albumin-creatinine-ratio (UACR) with structural brain abnormalities visible on magnetic resonance imaging (MRI). We also assessed whether this association was altered when different biomarkers were used to estimate GFR.

STUDY DESIGN: Cross-sectional study nested in a cohort study.

SETTING & PARTICIPANTS: 1527 participants in the Atherosclerosis Risk in Communities (ARIC) Study.

PREDICTORS: Log-UACR and eGFR based on cystatin-C, creatinine, cystatin-C and creatinine, and beta-2-microglobulin (B2M).

OUTCOMES: Brain volume reduction, infarcts, micro-hemorrhages, white matter lesions.

ANALYTICAL APPROACH: Multivariable linear and logistic regression models fit separately for each predictor based on a predictor change equivalent to 1 interquartile range (1-IQR).

RESULTS: Lower eGFR was associated with reduced cortex volume (regression coefficient [95%-CI] per 1-IQR decrease: -0.07 [-0.12,-0.02]), increased log-white matter hyperintensity volume (0.07 [0.01,0.15]) and reduced white matter fractional anisotropy (-0.08 [-0.17,-0.01]). The results were similar when eGFR was estimated with different equations based on cystatin-C, creatinine, a combination of cystatin-C and creatinine, as well as B2M. Higher log-UACR was similarly associated with these outcomes as well as brain infarcts (odds ratio [95%-CI] per 1-IQR increase: 1.31 [1.13,1.52]) and micro-hemorrhages (1.30 [1.12,1.51]). Brain volume reduction was similar to regions usually susceptible to Alzheimer's disease (AD) and limbic-predominant age-related TDP-43 encephalopathy (LATE) compared to the rest of the cortex.

LIMITATIONS: No inference about longitudinal effects due to cross-sectional design.

CONCLUSIONS: UACR and eGFR are associated with structural brain damage across different domains of etiology. eGFR- and UACR-related brain atrophy is not selective for regions typically affected by AD and LATE. Hence, AD or LATE may not be leading contributors to neurodegeneration associated with CKD.