The Role of Brain Structure in Explaining Physical Functioning in Male Veterans With Impaired Kidney Function

ABSTRACT

Background

Older adults with chronic kidney disease (CKD) experience deficits in physical function and cardiorespiratory fitness at higher rates than older adults without impaired renal function. Emerging evidence suggests covert brain pathology may be contributing to these functional impairments. However, much of the previous work has not controlled for common cardiac comorbidities also associated with brain pathology and reduced physical function. The present analysis investigates differences in global brain structure, physical performance, and cardiorespiratory fitness between older adults with CKD and hypertensive controls without kidney disease. This analysis also explores the role of brain structure in mediating the relationship between renal function and physical function while controlling for common cardiac comorbidities.

Methods

Forty‐one older male veterans with CKD and 30 hypertensive controls were recruited from clinics at the VA Maryland Health Care System to complete physical function assessments (Six Minute Walk Test (6MWT), Short Physical Performance Battery, Timed Up and Go), a graded exercise treadmill test to evaluate fitness, and brain magnetic resonance imaging to quantify total gray matter volume, total white matter volume, and total white matter lesion volume. Group differences were assessed, then mediation of brain structure on the relationships between continuous measures of renal and physical function was evaluated with all participants combined.

Results

Older male veterans with CKD had lower gray matter volume (597.5 ± 46.9 vs. 624.4 ± 51.4 cc), poorer endurance (6MWT: 431.2 ± 112.9 vs. 484.1 ± 99.4 m), and lower cardiorespiratory fitness (VO₂ peak: 19.4 ± 5.2 vs. 25.1 ± 7.5 mL/kg/min) compared with hypertensive controls. Global gray matter volume, but not total white matter volume nor white matter lesion volume, mediated the relationships between renal and physical function measures. Mediation held after accounting for common cardiac comorbidities and risk factors.

Conclusions

CKD‐specific pathways, distinct from traditional hemodynamic or ischemic mechanisms, may be contributing to the impaired physical functioning observed among those with impaired renal function.

Summary

• Key points

  • ○
    Older male veterans with chronic kidney disease have lower gray matter volume, cardiorespiratory fitness, and poorer physical function compared to hypertensive controls.
  • ○
    Gray matter volume, but not total white matter volume nor white matter lesion volume, mediates the relationships between renal function and physical function. These relationships held even after accounting for cardiometabolic risk factors.
• Why does this matter?

  • ○
    A nephrological‐neurological approach considering both renal function and brain structure is indicated in addressing physical performance limitations among those with chronic kidney disease. Gray matter volume quantification may assist in identifying fall risk among adults with impaired renal function, given its mediating role on physical performance measures.

Brain structure as a mediator of the relationship of impaired kidney function and physical function.

1. Introduction

Chronic Kidney Disease (CKD) affects one in five older adults in the US, including nearly 600,000 older veterans, yet is often unrecognized by affected individuals and even their healthcare providers [1, 2]. Older adults with CKD suffer from a markedly higher risk of physical dysfunction, with more than one third of all older adults with CKD also being considered physically frail [3]. Deficits in physical function, capacity, and activity have been described, especially in end stage renal disease (ESRD). However, many of these prior studies have relied on self‐reported measures of function limitations or functional disability, and common co‐morbidities (e.g., cardiac disease) have not consistently been accounted for in comparisons with non‐ESRD controls [4, 5, 6]. Given the extreme physiologic disturbances present in dialysis‐dependent patients, deficits in physical function are perhaps expected and unsurprising. However, more recent studies have also provided evidence of deficits in physical function and lower physical activity in earlier stages of CKD and across a spectrum of renal function [7, 8, 9]. Still, objective measures of impaired physical function and exercise capacity have not been well described in the context of renal dysfunction. Further, the factors which account for physical dysfunction in those with impaired renal function are unclear.

Recent neuroimaging investigations suggest a greatly increased burden of covert brain pathology in association with CKD and reduced kidney function [10, 11, 12, 13]. Outside the context of CKD, these neuroimaging findings of subclinical ischemic brain disease—silent brain infarcts, white matter hyperintensities, atrophy, and hypoperfusion—are associated with reduced physical function [14, 15, 16, 17, 18]. However, there is limited work evaluating subclinical ischemic brain disease as a mediator of the relations of renal function to physical function, across the spectrums of renal and physical function.

Despite the high prevalence of CKD in aging adults and the high burden of physical function deficits, there have been limited investigations into the pattern of these deficits, their relations with renal biomarkers, and the mechanisms that link CKD with physical dysfunction. Therefore, the purpose of this analysis was to objectively measure physical functioning and physiological capacity in older veterans with CKD, and to understand the role of brain structure in mediating impaired physical functioning in those with impaired kidney function. We hypothesized that older adults with CKD suffer from an increased burden of subclinical ischemic brain disease, which is associated with impaired physical function independent of common co‐morbidities including cardiac disease, hypertension, and diabetes.

2. Methods

2.1. Study Population

Community‐dwelling veterans with CKD and hypertensive controls were recruited from the VA Maryland Health Care System. Participants were recruited through nephrology, diabetes, and primary care clinics. Participants were eligible if aged 60–90 years with known hypertension. Hypertension was verified by medical chart review, medical history obtained by the participant, and/or medication review. CKD was defined according to KDIGO criteria as persistent estimated glomerular filtration rate (eGFR) < 60 cc/min/1.73 m² or urine albumin/creatinine ratio (uACR) > 30 mg/g for at least 3 months [19]. eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) equation [20]. Participants were excluded if any of the following conditions were present: eGFR < 15 cc/min/1.73 m²; treatment with dialysis; NYHA class III‐IV heart failure; coronary revascularization or acute coronary syndrome within the prior 3 months; vascular claudication; prior stroke; dementia; severe anemia (Hb < 9 g/dL); known primary or secondary brain pathologies; consumption of > 14 alcoholic beverages per week; and poorly controlled hypertension (BP > 180/95) or diabetes (HbA1C > 11%). Participants provided written informed consent prior to participation in study procedures. All study procedures were approved by the University of Maryland Baltimore Institutional Review Board (HP‐43880) and the Veterans Affairs Research and Development Committee. Study procedures were carried out in accordance with the Declaration of Helsinki.

2.2. Study Procedures

Participants who provided written informed consent then completed screening with a study physician including detailed medical history, physical examination, and a blood draw. Health history captured from chart review and the history and physician‐directed physical exam captured information on concussion history, presence of co‐morbid conditions (coronary artery disease, congestive heart failure, peripheral arterial disease, cardiovascular disease, diabetes, chronic obstructive pulmonary disease), and number of active medications. Enrolled participants completed research testing which included measuring anthropometry along with tests of physical function, a graded exercise treadmill test, and brain magnetic resonance imaging (MRI). Resting blood pressure was measured from a seated position after 5 min of rest across three intervals, 1 min apart, using an automated sphygmomanometer. The average of these measurements was used.

2.2.1. Brain Imaging Acquisition and Analysis

Neuroimaging was performed at the University of Maryland Medical Center on a 3.0 Tesla scanner (Philips Achieva, Philips, Amsterdam, Netherlands) equipped with a 12‐channel head–neck coil. All participants received the following scan sequences: T1‐weighted images, T2‐weighted images, FLAIR, DTI, and MP‐RAGE. All MRIs were reviewed by a radiologist for incidental clinical findings prior to analysis. Participants with pathological findings, such as tumors, identified on MRI were excluded (N = 3). Imaging acquisition and analysis was performed in accordance with previously published methodology [21]. Briefly, a multi‐atlas fusion methodology was used to segment the brain into anatomical regions of interest (ROIs) and compute volumetric measures for normal and abnormal (with lesion) brain tissue within each ROI. Total white matter lesion volume, total brain gray matter, and total brain white matter volume were calculated and used in analyses. Greater white matter lesion volume, less total brain gray matter, and/or a lower total brain white matter volume were considered suggestive of subclinical ischemic brain disease in this cohort of adults free from known clinically overt cerebrovascular disease.

2.2.2. Physical Function and Exercise Test Data Collection

Participants completed a battery of physical function assessments including the Six Minute Walk Test (6MWT), Short Physical Performance Battery (SPPB), and Timed Up and Go (TUG). For the 6MWT, an assessment of submaximal exercise capacity, participants were instructed to “walk as far as possible” for 6 min around a 100‐ft oval track [22]. Total distance walked was recorded. The SPPB assessed lower extremity function over three components: (1) usual gait speed over four meters; (2) time to complete five unassisted chair rises; and (3) standing balance battery consisting of maintaining balance with three different foot positions: feet together, semi tandem, full tandem. Each component was scored 0–4 and the total scores were summed for a final score ranging from 0 to 12, with higher scores indicating better lower extremity function and mobility [23]. The TUG assessed functional mobility by asking participants to rise from a straight‐backed chair, walk three meters, turn around, walk back to the chair, and sit down as fast as possible [24]. The fastest time to complete two trials was used. Participants were allowed to use their assistive devices for physical function tests.

Cardiorespiratory fitness was assessed as peak oxygen consumption during a Graded Exercise Treadmill Test. Participants walked on a treadmill at a pre‐selected speed with progressive increases in incline until volitional fatigue, at which time the test was terminated as described previously [25]. Briefly, speed was determined during a quick warm‐up of treadmill walking where the belt speed was increased by 0.1 mph increments until target gait speed was reached. The graded test consisted of 2 min of walking at the target gait speed with no incline, followed by 2 min at 4% incline, after which incline increased by 2% every 1 min until volitional fatigue. Oxygen uptake, CO₂ production, and ventilation were measured continuously using a metabolic cart (COSMED Quark CPET, COSMED, Rome, Italy). The two highest VO₂ values observed towards the end of the test were averaged and considered peak oxygen consumption (VO₂ peak).

2.3. Statistical Modeling and Analyses

Female participants (n = 3) and participants with racial identities other than Black or White (n = 4) were excluded from analysis given their small number and failure of bootstrapping models when controlling for sex and race. Descriptive statistics are presented as mean ± standard deviation for continuous variables and frequencies (percentages) for categorical variables. Pulse pressure (computed as the difference between systolic and diastolic blood pressure), low‐density lipoprotein cholesterol, triglycerides, body mass index, and hemoglobin A1C (Hb1AC) were converted into z‐scores and summed to reflect a standardized cumulative cardiovascular risk score (zCVR).

2.3.1. Differences Between Participants With CKD and Hypertensive Controls

Differences in demographics, health characteristics, and physical functioning between participants with CKD and hypertensive controls were evaluated with independent samples t‐tests for continuous data and independent samples proportions tests for categorical data.

2.3.2. Relationship Between Renal and Physical Function

To understand the linear relationships between renal function (eGFR, uACR) and physical function (6MWT, SPPB, TUG, VO₂ peak), data from all participants were combined into a single group. Normality of measures was assessed with Q‐Q plots and log transformations were applied for non‐normality violations to adjust for non‐normality as indicated: eGFR, uACR, VO₂ peak, SPPB, TUG, and total white matter lesion volume were log transformed for analyses. Linear relationships were evaluated by Pearson correlations and regression residual plots were assessed for violations of homoscedasticity. No violations were observed.

2.3.3. Mediation of Brain Structure

To investigate whether the relationship between kidney function (eGFR, uACR) and physical function (6MWT, SPPB, TUG, VO₂ peak) was mediated by structural brain changes (total white matter lesion volume, total gray matter volume, and total white matter volume), a mediational analysis controlling for age and race was performed using Hayes PROCESS Macro Procedure Version 4.2 in SPSS Version 29 (IBM, Armonk, NY, USA). Figure 1 depicts a simplified path model used to assess the indirect effect of a kidney function predictor, P, on a physical function outcome, O, through a brain structure mediator, M. The total effect (c) is the unmediated effect of P on O, the direct effect (c′) is the effect of P on O considering mediation by M, and the indirect effect is represented by a*b, which is the effect of P on O through M. Significance of the mediation effect, that is, the indirect effect, was evaluated by the 95% confidence intervals of 5000 bootstrapped samples and unstandardized effects. To evaluate if cardiovascular health was driving the observed associations and mediation, sensitivity analyses were performed to evaluate whether the mediational models held their significance after accounting for cumulative cardiovascular risk score (zCVR).

FIGURE 1.

Example mediation path model where c′ is the direct effect of the independent variable, P, on the dependent variable, O, considering M. a*b is the indirect effect of the independent variable, P, on the dependent variable, O, through the mediator, M, and c is the total, unmediated effect of P on O, including both direct and indirect effects.

3. Results

One hundred forty‐four veterans were screened and 120 were enrolled and consented to participate in the study (Figure S1). Seventy‐one male participants completed all testing, had complete data, and were included in the final analysis. Participants analyzed were older adults aged 60–87 (average: 69.7 ± 6.7 years) and predominantly identified as Black (n = 43, 60.6%). Participants with CKD were similar in age, racial identity distribution, and most health characteristic markers when compared to hypertensive controls (Table 1). As expected, participants with CKD had a lower eGFR, higher uACR, and higher HbA1C compared to hypertensive controls. Further, the prevalence of cardiovascular disease and diabetes was higher in those with CKD compared with hypertensive controls.

TABLE 1.

Participant demographics and health characteristics by group presented as mean ± standard deviation for continuous data and N (%) for categorical data.

		Chronic kidney disease (N = 41)	Hypertensive control (N = 30)	p	All participants (N = 71)
Age, years		70.9 ± 6.1	68.2 ± 7.2	0.09	69.7 ± 6.7
Race	Black	25 (61)	18 (60)	0.93	43 (60.6)
	White	16 (39)	12 (40)		28 (39.4)
Concussion history		7 (17.1)	4 (13.3)	0.67	11 (15.5)
Coronary artery disease		12 (29.3)	4 (13.3)	0.10	16 (22.5)
Congestive heart failure		5 (12.2)	1 (3.3)	0.16	6 (8.5)
Peripheral arterial disease		3 (7.5)	2 (6.7)	0.89	5 (7.1)
Cardiovascular disease		14 (34.1)	5 (16.7)	0.08	19 (27.1)
Diabetes		29 (70.7)	7 (23.3)	< 0.001	36 (50.7)
COPD		2 (4.89)	1 (3.3)	0.75	3 (4.2)
Number of medications		9.7 ± 3.7	8.2 ± 3.9	0.11	9.1 ± 3.8
eGFR, cc/min/1.73 m2		44.9 ± 14.6	79.8 ± 14.0	< 0.001	59.6 ± 22.5
uACR, mg/g		397.6 ± 809.9	7.5 ± 10.4	0.01	232.8 ± 642.3
Cumulative cardiovascular risk score (zCVR) b		0.56 ± 2.52	−0.97 ± 2.60	0.02	−0.08 ± 2.65
Pulse Pressure, mmHg a , b		64.3 ± 19.1	57.2 ± 11.0	0.07	61.3 ± 16.4
LDL mg/dL a		95.5 ± 33.4	101.9 ± 35.5	0.44	98.2 ± 34.2
Triglycerides mg/dL a		129.8 ± 66.6	99.7 ± 61.0	0.06	117.1 ± 65.6
Body mass index, kg/m2a		30.1 ± 5.5	29.7 ± 5.3	0.77	29.9 ± 5.4
HbA1C, mmol/mol a , b		6.9 ± 1.1	6.0 ± 1.0	< 0.001	6.5 ± 1.1

Note: Bolded values indicate statistically significant differences between groups with p < 0.05.

Abbreviations: COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; HbA1C, hemoglobin A1C; LDL, low‐density lipoprotein; uACR, urine albumin/creatinine ratio.

^a Used to calculate cumulative cardiovascular risk score (zCVR).

^b N = 70.

Brain structure and physical function measures were compared between veterans with CKD and hypertensive controls to assess if older male veterans with CKD suffer from an increased burden of subclinical ischemic brain disease and to assess physical function differences between adults with CKD and those without (Table 2). Participants with CKD had significantly lower gray matter volume compared to hypertensive controls and poorer physical functioning as measured with the 6MWT and TUG. Participants with CKD scored modestly higher than hypertensive controls on the SPPB and had lower cardiorespiratory fitness assessed by VO₂ peak. There were no differences in total white matter volume or total white matter lesion volume between those with and without CKD. Differences between groups in gray matter volume and VO₂ peak remained statistically significant after adjusting for cumulative cardiometabolic risk, race, and age (Table S1).

TABLE 2.

Comparison of brain structure, physical function, and cardiorespiratory fitness measures between participants with CKD and hypertensive controls showing unadjusted p‐values.

Measure, units	Chronic kidney disease (N = 41)	Hypertensive control (N = 30)	Unadjusted p‐value	All participants combined (N = 71)
Gray matter volume, cc	597.5 ± 46.9 a	624.4 ± 51.4 b	0.03	608.6 ± 50.3 c
White matter volume, cc	458.9 ± 46.4 a	476.6 ± 48.1 b	0.13	466.2 ± 47.6 c
Total white matter lesion volume, cc	3.8 ± 4.7 a	4.1 ± 6.7 b	0.85	3.9 ± 5.6 c
Six Minute Walk Test, m	431.2 ± 112.9	484.1 ± 99.4	0.045	453.9 ± 109.8
Timed Up and Go, s	8.0 ± 2.5	7.0 ± 1.6	0.05	7.6 ± 2.2
Short physical performance battery, out of 12	10.7 ± 1.2	9.6 ± 1.9	0.01	10.1 ± 1.8
VO2 peak, mL/kg/min	19.4 ± 5.2	25.1 ± 7.5	< 0.001	21.8 ± 6.8

Note: Adjusted comparisons for cumulative cardiometabolic risk, race, and age are presented in the Supporting Information. Bolded values indicate statistically significant differences between groups with p < 0.05.

^a N = 40.

^b N = 28.

^c N = 68.

Mediation analyses were performed to assess if the relationship between renal function and subclinical ischemic brain disease is associated with impaired physical function, and sensitivity analyses accounting for zCVR were done to assess if these relationships are explained by differences in cardiovascular risk. Results of mediation and sensitivity analyses for gray matter volume as a mediator on the relationships between kidney function and physical function are included in Table 3. Gray matter volume mediated the relationship between eGFR and all physical function measures, 6MWT, b = 30.64, 95% CI [3.60, 66.99]; VO₂ peak, b = 0.08, 95% CI [0.010, 0.18]; SPPB, b = 0.05, 95% CI [0.002, 0.16]; TUG, b = −0.06, 95% CI [−0.15, −0.004]. After accounting for a summary score for cardiovascular risk factors (zCVR), gray matter volume continued to mediate the relationship between eGFR and 6MWT, 95% CI [4.74, 69.71]; VO₂ peak, 95% CI [0.01, 0.18]; SPPB, 95% CI [0.002, 0.17]; and TUG, 95% CI [−0.16, −0.003].

TABLE 3.

95% confidence intervals for the mediation effect of gray matter volume on the relationship between kidney function predictors and physical function outcomes, including sensitivity analysis adjusting for zCVR in addition to age and race.

Mediator, M	Independent variable, P	Dependent variable, O	Indirect effect of P on O through M (a*b)	Sensitivity analysis (a*b accounting for zCVR)
Gray matter volume (cc)	eGFR (cc/min/1.73 m2)	6MWT (m)	[3.6, 67.0]	[4.7, 69.7]
		SPPB (out of 12)	[0.002, 0.2]	[0.002, 0.2]
		TUG (s)	[−0.2, −0.004]	[−0.2, −0.003]
		VO2 peak (mL/kg/min)	[0.01, 0.2]	[0.01, 0.2]
	uACR (mg/g)	6MWT (m)	[−11.1, −0.1]	[−11.3, −1.1]
		SPPB (out of 12)	[−0.02, 0.00]	[−0.03, −0.0001]
		TUG (s)	[0.0006, 0.03]	[0.001, 0.03]
		VO2 peak (mL/kg/min)	[−0.03, −0.002]	[−0.03, −0.002]

Note: Bolded values indicate statistical significance with 95% confidence intervals not crossing 0.

Abbreviations: 6MWT, Six Minute Walk Test; eGFR, estimated glomerular filtration rate; SPPB, short physical performance battery; TUG, timed up and go; uACR, urine albumin/creatinine ratio; zCVR, standardized cumulative cardiovascular risk score.

Gray matter volume mediated the relationship between albuminuria (uACR) and all physical function measures as displayed in Figure 2, 6MWT, b = −4.87, 95% CI [−11.06, −0.76]; VO₂ peak, b = −0.01, 95% CI [−0.03, −0.002]; TUG, b = 0.010, 95% CI [0.0006, 0.03]; and SPPB, b = −0.009, 95% CI [−0.02, 0.00]. In sensitivity analyses after accounting for cumulative cardiovascular risk factors, gray matter volume continued to mediate the relationship between uACR and all physical function measures: 6MWT, 95% CI [−11.31, −1.10]; VO₂ peak, 95% CI [−0.03, −0.002]; TUG, 95% CI [0.001, 0.03]; SPPB, 95% CI [−0.03, −0.0001].

FIGURE 2.

Mediation pathways with unstandardized coefficients (SD), p‐values for gray matter volume on the relationships between eGFR (cc/min/1.73 m²) and (i) 6MWT (m); (ii) SPPB (out of 12); (iii) TUG (s); and (iv) VO₂ peak (mL/kg/min); and the relationships between uACR (mg/g) and (v) 6MWT (m); (vi) SPPB (out of 12); (vii) TUG (s); and (viii) VO₂ peak (mL/kg/min).

Total white matter volume mediated the relationship only between uACR and TUG (95% CI [0.00, 0.02]), and this mediation became non‐significant after accounting for cumulative cardiovascular risk factors (Table S2). White matter lesion volume did not mediate the relationship between kidney function and any of the physical function measures (Tables S3 and S4). Comprehensive information on mediation pathway outputs for all models analyzed are provided in Supporting Information (Tables S2–S8).

4. Discussion

The present analysis evaluated the relationship between renal dysfunction and physical functioning in older male veterans with CKD and assessed the role of brain structure in mediating these relationships, while considering cardiometabolic risk factors. We hypothesized that older adults with CKD suffer from an increased burden of subclinical ischemic brain disease, which is associated with impaired physical function independent of cardiometabolic comorbidities. As expected, participants with CKD had a greater prevalence of cardiovascular disease and diabetes when compared to hypertensive controls. Further, while it is likely that higher HbA1C underestimates lack of glycemic control for those with CKD [26], HbA1C was still significantly higher in participants with CKD compared to hypertensive controls. Older veterans with CKD had significantly lower gray matter volume when compared to hypertensive controls and poorer physical functioning on nearly all measures evaluated, including poorer cardiorespiratory fitness.

This is in line with previous work by Odden et al., who reported those with stage 3 or greater CKD had a 1.4‐MET lower aerobic capacity on graded treadmill testing compared to non‐CKD controls, after adjustment for co‐morbidity [27]. In the present analysis, those with CKD had a VO₂ peak 5.67 mL/kg/min lower than hypertensive controls, equating to a 1.62‐MET lower aerobic capacity. Aerobic capacity among these stroke‐free CKD participants is similar to that observed among post‐stroke patients, those with heart failure, and older adults with peripheral arterial disease reported previously [28, 29]. Differences in aerobic capacity between groups remained statistically significant even after adjusting for cumulative cardiometabolic risk, race, and age.

Compared to hypertensive controls, those with CKD also had markedly lower submaximal endurance, as measured by the 6MWT. The difference in average group performance on the 6MWT was 52.9 m, exceeding the minimal clinically important difference established in non‐CKD‐specific older adult cohorts [30, 31]. However, it is important to note that participants in both groups, on average, walked more than 400 m, exceeding the cut point for increased mortality risk previously defined in adults with advanced CKD [32]. Additionally, these group differences were no longer statistically significant after adjusting for cumulative cardiometabolic risk, race, and age.

Gray matter volume, but not white matter lesion volume, mediated the relationships between renal function and physical function. Total white matter volume mediated the relationship between only one renal and physical function measure; mediation was non‐significant after accounting for cardiovascular risk factors in the model. Mediation of gray matter volume on these relationships continued to be statistically significant even after accounting for cumulative cardiovascular risk factors, indicating the relationships between renal and physical function as explained by brain structure are not explained entirely by cardiometabolic factors associated with renal dysfunction, but may be unique to the renal dysfunction itself.

This mechanistic insight extends prior observational studies linking CKD to functional decline by identifying cortical atrophy as a critical neural substrate underlying these associations. The 4.3% lower gray matter volume in CKD participants compared to hypertensive controls (597.5 vs. 624.4 cm³, p = 0.03) and its persistent mediation of renal‐physical function relationships after adjusting for cardiovascular risk (zCVR) suggest CKD‐specific neurotoxicity pathways distinct from traditional hemodynamic or ischemic mechanisms.

4.1. Contrasting Neurodegenerative Pathways

The specificity of gray matter volume's mediation, contrasted with null effects from white matter and white matter lesion volume, challenges cerebrocentric models of CKD‐related disability rooted in small vessel ischemia [33]. While prior work emphasized white matter hyperintensities and silent infarcts in CKD‐associated cognitive decline [10, 11, 12, 13, 34], our mediation models reveal gray matter volume as a primary factor in physical dysfunction. This divergence aligns with evidence implicating uremic toxins (e.g., indoxyl sulfate) with blood–brain barrier disruption in CKD [35]. The cohort's gray matter volume loss mirrors patterns observed in proteinuric CKD models, potentially linking albuminuria's direct neurotoxicity with accelerated cortical atrophy [36].

The absence of white matter lesion mediation in the present analysis contrasts with general aging studies linking white matter hyperintensities to mobility limitations [37]. This divergence underscores CKD‐specific neurodegeneration patterns where metabolic disturbances may overshadow ischemic white matter lesions in early disease [35, 38, 39, 40]. White matter tract integrity may be temporarily preserved, while oxidative stress associated with kidney disease preferentially damages neuronal cell bodies. Our results extend Odden et al. seminal work by identifying gray matter volume as the neural substrate underlying their observed CKD‐related aerobic capacity reduction [27].

4.2. Metabolic Over Hemodynamic Mechanisms

The preserved mediation effects after adjusting for cumulative cardiovascular risk challenges the idea that functional decline associated with CKD is explained by a greater burden of cardiovascular disease. Instead, in line with our initial hypothesis that associations of renal function with subclinical ischemic brain disease and impaired physical function would be independent of common cardiovascular comorbidities, our findings may support a mechanism of shared microvascular etiologies where endothelial glycocalyx damage and inflammatory cytokine activation simultaneously compromise renal and neural tissues. The mediation of gray matter volume for both eGFR (b = 30.64 for 6MWT) and uACR (b = −4.87 for 6MWT) underscores dual pathways in which glomerular filtration (eGFR) likely influences toxin clearance critical for neuronal health and albuminuria (uACR) reflects endothelial dysfunction and blood–brain barrier breakdown, exacerbating neuroinflammation. The lower aerobic capacity observed in the present analysis (VO₂ peak 19.43 vs. 25.10 mL/kg/min, p < 0.001) with CKD parallels deficits seen in stroke populations, highlighting CKD's profound central nervous system involvement in exercise intolerance beyond cardiopulmonary limitations.

4.3. Clinical Implications and Future Directions

These findings highlight the brain changes which may occur early on in those with identified renal dysfunction. Future research directions should focus on creating longitudinal monitoring frameworks in which serial neuroimaging could track the progression of CKD effects, analogous to albuminuria monitoring in nephropathy. Notably, the present analysis did not adjust for multiple comparisons. Effect sizes and confidence intervals can guide interpretation and may be used to power larger studies evaluating the findings of this work longitudinally. Longitudinal studies should also clarify whether gray matter volume loss precedes or follows functional decline, informing prevention strategies, and explore more in‐depth analyses of comorbidity duration and severity, genetic risk factors, and protective social or behavioral contexts to understand how condition management may be adjusted based on what factors contributed to development.

Although modifying CKD management specifically to preserve brain health may not presently be part of routine practice, our findings highlight the importance of understanding the mechanisms by which CKD contributes to the changes in brain structure and function. Identifying key pathways, such as the role of specific uremic toxins or metabolic disturbances, could lead to adjunctive therapies or strategies aimed at neuroprotection, even if renal function cannot be meaningfully improved. Additionally, recognizing that neurocognitive sequelae often develop alongside renal dysfunction underscores the need to incorporate compensatory approaches and supportive measures into existing treatment frameworks, with the goal of promoting adherence and optimizing functional independence. Future research should examine whether interventions targeting these mechanistic pathways, such as approaches to neutralize or remove uremic toxins or medications with potential neuroprotective effects, confer benefit with respect to cognitive and physical outcomes for adults living with CKD.

It is important to note that the present analysis includes only male veteran participants; the findings of this work may not generalize to non‐veterans or female veterans. Future work should investigate if the relationships identified in the present analysis differ by sex and/or veteran status. Nonetheless, by establishing gray matter volume's mediating role between renal dysfunction and physical disability, this study bridges observational reports and provides a neuroanatomical basis for targeted interventions for older male veterans. The findings advocate for integrated nephrological‐neurological care models addressing both renal and neural health in CKD management.

Author Contributions

All authors have read and approved the submission of this manuscript. J.E.G., S.L.S., and S.R.W. contributed to the study concept and design. J.E.G., K.P.W., S.L.S., and S.R.W. contributed to the data acquisition. C.D., G.E., H.B.M., J.E.G., J.S.R., and S.R.W. contributed to data analysis. H.B.M., J.E.G., J.S.R., K.P.W., S.L.S., and S.R.W. contributed to the interpretation of data. H.B.M. and S.R.W. were responsible for statistical analyses. H.B.M., J.E.G., and J.S.R. drafted the manuscript. All authors contributed to critical revisions of the manuscript for important intellectual content. S.L.S. provided study supervision.

Funding

S.L.S. is supported by VA 1I01RX000159 and P30AG028747. J.S.R. was provided with protected time to prepare this manuscript as part of the Advanced Fellowship in Geriatrics, supported by the U.S. Department of Veterans Affairs Office of Academic Affiliations, the Veterans Affairs Maryland Health Care System, and the Department of Veterans Affairs Baltimore Geriatric Research, Education, and Clinical Center (GRECC). The APC was funded by the Geriatric Research, Education, and Clinical Center, Veterans Affairs Maryland Health Care System, Baltimore. The views expressed in this article are those of the authors and do not necessarily represent the position or policy of the U.S. Department of Veterans Affairs of the United States Government.

Disclosure

The funding institution had no role in the design and conduct of the study, analysis and interpretation of data, preparation, review, or approval of the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: jgs70258‐sup‐0001‐supinfo.pdf.

Rekant J. S., Malik H. B., Giffuni J. E., et al., “The Role of Brain Structure in Explaining Physical Functioning in Male Veterans With Impaired Kidney Function,” Journal of the American Geriatrics Society 74, no. 2 (2026): 470–478, 10.1111/jgs.70258.