Association Between Blood Cobalt Levels and Arteriovenous Fistula Dysfunction in Patients Undergoing Maintenance Hemodialysis: A Multicenter Observational Study

Abstract

Rationale & Objective

Prior studies have suggested that increased cobalt ion concentrations in circulation provoke inflammation in endothelial cells, disrupting vascular function. In patients undergoing maintenance hemodialysis (MHD), elevated cobalt may result from various factors. Given the critical role of arteriovenous fistulas (AVFs) in the survival of MHD patients, we aimed to investigate the relationship between cobalt levels and AVF dysfunction.

Study Design

Multicenter, observational cohort study.

Setting & Study Populations

A total of 518 patients undergoing maintenance hemodialysis at 3 medical centers.

Exposures

Blood samples were collected before dialysis, and whole-blood cobalt levels were determined using inductively coupled plasma mass spectrometry.

Outcomes

Blood cobalt concentrations were measured at baseline. Participants were monitored for AVF dysfunction, defined by a blood flow reduction to ≤200 mL/min during dialysis, and followed up to 2 years.

Analytical Approach

Patients were divided into groups based on the occurrence of AVF dysfunction and analyzed using multivariable Cox regression, Kaplan–Meier curves, log-rank tests, restricted cubic spline analysis, and stratified analysis.

Results

Of the 518 patients, 126 (24.3%) experienced AVF dysfunction within a median 16-month follow-up. Patients with AVF dysfunction exhibited significantly higher blood cobalt concentrations than did those without AVF dysfunction. The highest AVF dysfunction rate was observed in patients in the upper cobalt quartile (P < 0.001). Multivariable analysis indicated a 2.90-fold increased risk of AVF dysfunction in patients in the highest cobalt level quartile (P < 0.001). No significant interactions were observed in the stratified analyses (P > 0.05).

Limitations

We only assessed the baseline cobalt levels at enrollment and did not perform serial measurements; we could not determine whether factors influencing trace element intake or absorption (eg, dietary habits or lifestyle) might have introduced temporal fluctuations in cobalt levels during follow-up, potentially influencing the results.

Conclusions

Elevated blood cobalt levels were independently associated with an increased risk of AVF dysfunction in patients undergoing MHD. Further research on cobalt metabolism and its effects on AVF functionality to guide clinical interventions is required.

Plain-language Summary

Previous studies have suggested that increased cobalt ion concentrations induce inflammation in endothelial cells (ECs) and disrupt vascular function. In patients undergoing maintenance hemodialysis (MHD), elevated cobalt levels may be caused by various factors. Given the critical role of arteriovenous fistulas (AVFs) for MHD patient survival, this study investigated the relationship between blood cobalt levels and AVF dysfunction. This multicenter observational cohort study included 518 patients receiving MHD from 3 medical centers. Divided into groups based on the occurrence of AVF dysfunction, data analysis involved multivariable Cox regression, Kaplan–Meier curves, log-rank tests, restricted cubic spline analysis, and stratified analysis. Of the 518 patients, 126 (24.3%) experienced AVF dysfunction within a median 16-month follow-up. Patients with AVF dysfunction exhibited significantly higher blood cobalt concentrations than did those without AVF dysfunction. The highest AVF dysfunction rate was observed in patients in the highest cobalt quartile (P < 0.001). Multivariable analysis indicated that elevated cobalt levels in the blood were independently associated with an increased risk of AVF dysfunction in patients undergoing MHD. This underscores the need for further research on cobalt metabolism and its effects on AVF functionality to guide clinical interventions.

Autologous arteriovenous fistula (AVF) remains the preferred route of vascular access for hemodialysis because of its stable blood flow, extended lifespan, reduced medical expenditure, and lower risk of complications and infections than alternative vascular access.¹ Despite these advantages, AVF immaturity and dysfunction pose ongoing clinical challenges.² Studies show that 20%-50% of AVFs are too immature for hemodialysis use,³ resulting in 1-year and 2-year patency rates of merely 64% and 51%, respectively.⁴ As AVF is a critical vascular access for patients undergoing maintenance hemodialysis (MHD), research exploring factors influencing AVF dysfunction has mainly focused on imbalances in macroelements such as calcium⁵ and phosphorus,⁶ the underlying disease status of patients,⁷ and inflammatory signaling pathways.⁸ However, few studies have investigated the effect of blood trace element levels on AVF function. This is because patients with end-stage kidney disease often exhibit alterations in blood, serum, or plasma concentrations of trace elements, which may result from reduced protein intake, changes in intestinal absorption, decreased renal excretion, and metabolic disturbances associated with kidney failure.⁹

Cobalt is integral to various physiological functions, existing in organic (such as vitamin B₁₂ [cobalamin]) and inorganic forms. In the clinical context of MHD, Chinese patients frequently use mecobalamin to alleviate symptoms of peripheral neuropathy, inadvertently increasing their cobalt intake. They also often suffer from iron deficiency, enhancing cobalt absorption owing to shared transporters in the upper small intestine.¹⁰ Cobalt is predominantly eliminated through renal excretion¹⁰; therefore, end-stage kidney disease patients often exhibit reduced cobalt clearance because of decreased glomerular filtration rates. Vitamin B₁₂ does not appear to be lost during MHD,¹¹ which suggests factors such as dietary patterns, including the Mediterranean diet, cobalt supplements, and contamination from dialysate or materials potentially contribute to elevated cobalt levels in patients undergoing MHD. Indeed, previous studies have documented higher blood cobalt levels in patients undergoing MHD compared with healthy controls.^12,13 Although an essential component of vitamin B₁₂, excessive amounts of cobalt and its inorganic form can adversely affect various body systems and ECs.¹⁰ The functional status of the AVF in patients undergoing MHD is closely related to that of vascular ECs. To date, no study has systematically evaluated the link between cobalt levels and AVF dysfunction. Therefore, this multicenter cohort study aimed to investigate the potential correlation between elevated blood cobalt levels and AVF dysfunction in patients undergoing MHD.

Materials and Methods

Study Design and Population

Between March 2021 and March 2023, patients undergoing hemodialysis with stable AVFs were recruited from Guangzhou Red Cross Hospital, Jiangmen Central Hospital, and Dongguan People’s Hospital. Participants meeting the following inclusion criteria were considered eligible: (1) age ≥18 years; (2) undergoing first-time AVF creation and initiating AVF use at the beginning of the study period; (3) receiving hemodialysis through a radiocephalic AVF; (4) undergoing stable dialysis for a minimum of 3 months, with treatments using bicarbonate dialysate conducted 3 times per week, each lasting 4 hours; and (5) provision of informed consent by the patients and their families. The exclusion criteria were as follows: (1) underwent a change in dialysis modality during the follow-up period, including kidney transplantation or peritoneal dialysis; (2) died, withdrew from the study, or were transferred; and (3) lost to follow-up or had incomplete data. The study was approved by the hospital ethics committees (number 2023-133-01), and all participating patients provided informed consent.

Laboratory Measurements

We collected demographic and clinical data pertaining to patients undergoing hemodialysis, including variables such as age, sex, smoking status, alcohol use, anthropometric measurements, end-stage kidney disease etiology, blood pressure, comorbid conditions, and medication profiles. Laboratory evaluations included a complete blood count; serum levels of albumin (Alb), uric acid, creatinine, cholesterol, triglycerides, and high-sensitivity C-reactive protein (hs-CRP); and measurements of calcium, phosphorus, parathyroid hormone, and ferritin. Blood cobalt concentrations were determined using inductively coupled plasma mass spectrometry, with a reference range of 0.1-0.7 μg/L. Blood samples were collected before dialysis, and information regarding the dialysis access type and modality was monitored. Data acquisition involved the review of medical records and clinical databases, which were routinely updated. Patient follow-up continued until the endpoint event of AVF failure or until March 31, 2025.

Outcome Measures

The 2019 Kidney Disease Outcomes Quality Initiative (KDOQI) 2019 refers to arteriovenous access (AVF or arteriovenous graft) flow dysfunction as clinically significant abnormalities in arteriovenous access flow or patency due to underlying stenosis, thrombosis, or related pathological defects.¹⁴ According to the 2018 European Society of Vascular Surgeons clinical practice guidelines, the blood flow available for hemodialysis should reach at least 300 mL/min, preferably 500 mL/min, depending on the arteriovenous modality to allow sufficient hemodialysis.¹⁵ Considering the generally lower health status of Chinese patients compared with international cohorts, the expert panel recommends that an AVF blood flow rate of <200 mL/min, an arterial pressure <−250 mm Hg and/or venous pressure >250 mm Hg at a blood pump flow rate of 200 mL/min, or a catheter recirculation rate >10%, along with the inability to achieve adequate dialysis, may indicate AVF dysfunction in the Chinese population.¹⁶ Thus, the primary endpoint in this study was the reduction of AVF flow to ≤200 mL/min during 3 consecutive hemodialysis sessions, which necessitated either surgical or medical intervention. In this study, surgical intervention was defined as AVF revision surgery or any endovascular procedure, including percutaneous transluminal angioplasty and thrombectomy. In this context, pharmacologic intervention refers to the intrafistular administration of thrombolytic agents via a needle or catheter.

Statistical Analysis

Data analysis was performed using FreeStatisticsV1.9, SPSS 26, and R software (version 4.3.1). Normally or approximately normally distributed continuous data are expressed as means ± standard deviations. Conversely, skewed data are represented as medians with interquartile ranges (25th, 75th percentile). Continuous data exhibiting a normal distribution and homogeneity of variance was compared using the independent samples t test and one-way analysis of variance. Nonnormally distributed data or data demonstrating unequal variances were analyzed using the nonparametric rank-sum test or the Kruskal–Wallis test. Categorical data are described as percentages and were analyzed using the χ² test. Statistical significance was set at P < 0.05.

Restricted cubic splines were used to investigate the potential curvilinear relationship between the whole-blood cobalt concentration and AVF dysfunction. Univariate Cox regression analysis was used to identify the factors associated with fistula dysfunction, whereas multivariable Cox regression analysis was used to determine the independent risk factors. Kaplan–Meier curves and log-rank tests were performed to evaluate the intervention-free survival of patients with AVFs, and stratified analyses were conducted to elucidate the relationship between cobalt levels and fistula dysfunction. The results are displayed as forest plots, with an interaction P > 0.05 indicating a lack of significant association.

Results

Clinical Characteristics of the Participants

This study included 518 patients (Fig 1), of whom 126 (24.3%) developed AVF dysfunction over a median follow-up period of 16 months (interquartile range, 6, 22). Patients with dysfunction had a mean age of 61.89 ± 13.82 years and included 75 men (59.5%) and 79 patients with diabetes (62.7%). The median blood cobalt level was 0.40 μg/L (interquartile range, 0.30, 0.70), with 403 patients within the normal range and 115 patients (22.2%) exceeding the upper reference limit. Among patients with AVF dysfunction, 37 (32.2%) had elevated cobalt levels. Table 1 presents the clinical characteristics and laboratory findings stratified by the occurrence of AVF dysfunction. Compared with patients without AVF dysfunction, patients with AVF dysfunction had higher age, body mass index, serum phosphorus levels, proportion of men, and diabetes, but lower serum Alb levels. A statistically significant difference in primary disease etiology was observed between the groups (P < 0.05). Patients with AVF dysfunction had higher blood cobalt levels than did those without AVF dysfunction, whereas no significant differences were observed between groups in serum calcium and hs-CRP levels or other clinical indices. Additionally, we compared the baseline clinical parameters according to quartiles of blood cobalt levels (Table S1), which revealed a higher incidence of AVF dysfunction in the fourth quartile, along with elevated serum potassium, hs-CRP, total cholesterol, and parathyroid hormone levels.

Figure 1.

Flowchart for inclusion of populations. ESRD, end-stage renal disease.

Table 1.

Comparison of Clinical Characteristics between the Patent AVF Group and the Dysfunctional AVF Group

Variables	Overall	Patients Without AVF Dysfunction	Patients With AVF Dysfunction	P
	N = 518	N = 392	N = 126
Sex				0.005
Women	195 (37.6%)	144 (36.7%)	51 (40.5%)
Men	323 (62.4%)	248 (63.3%)	75 (59.5%)
Age (y)	59.89 ± 13.56	59.25 ± 13.43	61.89 ± 13.82	0.02
BMI (kg/m2)	22.43 ± 3.75	22.37 ± 3.83	22.46 ± 3.73	0.02
Drug use	290 (56.0%)	225 (57.4%)	65 (51.6%)	0.40
Smoking	74 (14.3%)	57 (14.5%)	17 (13.5%)	0.29
Alcohol consumption	16 (3.1%)	13 (3.3%)	3 (2.4%)	0.14
Diabetes	316 (61%)	237 (60.5%)	79 (62.7)%	0.03
CVD	172 (33.2%)	124 (31.6%)	48 (38.1%)	0.18
Primary disease				0.02
PGD	154 (29.7%)	120 (30.6%)	34 (27.0%)
DN	130 (25.1%)	99 (25.3%)	31 (24.6%)
HN	33 (6.4%)	24 (6.1%)	9 (7.1%)
Others	134 (25.9%)	105 (26.8%)	29 (23.0%)
Unknown	67 (12.9%)	44 (11.2%)	23 (18.3%)
WBC (109/L)	6.38 ± 1.85	6.39 ± 1.87	6.34 ± 1.78	0.61
Hb (g/L)	106.14 ± 15.06	106.60 ± 14.31	104.71 ± 17.18	0.26
PLT (109/L)	191.41 ± 69.79	189.51 ± 69.95	200.24 ± 68.70	0.56
NEUT (109/L)	4.69 ± 3.92	4.76 ± 4.41	4.48 ± 1.60	0.29
Potassium (mmol/L)	4.74 ± 0.78	4.71 ± 0.80	4.84 ± 0.72	0.07
Calcium (mmol/L)	2.18 ± 0.24	2.18 ± 0.25	2.17 ± 0.21	0.86
Magnesium (mmol/L)	1.07 ± 0.57	1.05 ± 0.22	1.14 ± 1.08	0.40
Phosphorus (mmol/L)	2.13 ± 0.63	2.12 ± 0.63	2.16 ± 0.63	0.01
BUN (mmol/L)	27.70 ± 9.94	27.69 ± 9.77	27.73 ± 10.48	0.40
cr (μmol/L)	1034.69 ± 289.81	1036.48 ± 292.39	1029.12 ± 282.69	0.80
UA (μmol/L)	461.82 ± 118.85	457.23 ± 117.82	476.10 ± 121.38	0.13
Alb (g/L)	38.55 ± 3.42	38.60 ± 3.41	38.38 ± 3.48	0.02
Cobalt (μg/L)	0.40 [0.30, 0.70]	0.40 [0.30, 0.60]	0.50 [0.30, 0.85]	0.001
Kt/V	1.25 [1.10, 1.43]	1.25 [1.10, 1.43]	1.25 [1.11, 1.42]	0.85
hs-CRP (mg/L)	8.28 [2.20, 36.00]	8.02 [2.30, 35.10]	9.89 [1.90, 36.95]	0.18
TG (mmol/L)	1.45 [0.89, 2.22]	1.45 [0.92, 2.26]	1.44 [0.82, 2.00]	0.50
TC (mmol/L)	3.92 [3.30, 4.74]	3.93 [3.30, 4.70]	3.90 [3.26, 4.80]	0.99
HDL (mmol/L)	1.02 [0.84, 1.25]	1.02 [0.82, 1.25]	1.03 [0.86, 1.24]	0.38
LDL (mmol/L)	2.28 [1.80, 2.93]	2.27 [1.80, 2.94]	2.30 [1.76, 2.88]	0.83
PTH (pg/mL)	322.80 [99.54, 621.10]	322.80 [86.80, 613.30]	322.80 [158.60, 663.20]	0.41

Note: Data are presented as n (%), mean ± standard deviation, or median [interquartile range].

Abbreviations: Alb, albumin; AVF, arteriovenous fistula; BMI, body mass index; BUN, blood urea nitrogen; cr, creatinine; CVD, cardiovascular disease; DN, diabetic nephropathy; Hb, hemoglobin; HDL, high-density lipoprotein; HN, hypertensive nephropathy; hs-CRP, high-sensitivity C-reactive protein; LDL, low-density lipoprotein; NEUT, neutrophil; PGD, primary glomerular disease; PLT, platelet; PTH, parathyroid hormone; TC, total cholesterol; TG, triglycerides; UA, uric acid; WBC, white blood cell.

Univariate Cox Proportional Hazards Regression Analysis Identified Factors Associated With AVF Dysfunction

Univariate Cox regression analysis revealed that male sex, older age, diabetes, and serum phosphorus, blood cobalt, and Alb levels were associated with the incidence of AVF dysfunction (Table 2). Patients undergoing MHD with blood cobalt levels exceeding the normal range (>0.7 μg/L) were at a 1.63-fold increased risk of AVF dysfunction (95% confidence interval [CI], 1.11-2.39). Compared with patients in the first quartile of blood cobalt levels, patients in the fourth quartile had a 2.82-fold higher risk of AVF dysfunction (95% CI, 1.62-4.89).

Table 2.

Univariate Cox Proportional Hazard Regression Analysis of Risk Factors Associated With AVF Dysfunction

Variables	Univariable Analysis
	HR (95% CI)	P
Sex	1.441 (1.069-1.943)	0.02
Age	1.018 (1.007-1.030)	0.001
BMI	1.003 (0.954-1.048)	0.99
Calcium	1.115 (0.574-2.167)	0.44
Phosphorus	1.331 (1.059-1.672)	0.01
Alb	0.936 (0.897-0.976)	0.002
hs-CRP	1.003 (0.995-1.011)	0.45
TG	1.021 (0.940-1.110)	0.62
TC	0.997 (0.873-1.140)	0.88
LDL	0.983 (0.823-1.174)	0.63
Diabetes	1.57 (1.16-2.12)	0.003
CVD	1.36 (0.95-1.95)	0.09
Cobalt	1.53 (1.22-1.93)	<0.001
Cobalt range
0.1–0.7 μg/L (normal)	1 (Ref)
>0.7 μg/L (high)	1.63 (1.11-2.39)	0.01
Cobalt quartile
Q1	1 (Ref)
Q2	2.02 (1.14-3.58)	0.02
Q3	2.04 (1.15-3.62)	0.02
Q4	2.82 (1.62-4.89)	<0.001
Trend.test	1.34 (1.14-1.57)	<0.001

Abbreviations: Alb, albumin; AVF, arteriovenous fistula; BMI, body mass index; CI, confidence interval; CVD, cardiovascular disease; HDL, high-density lipoprotein; HR, hazard ratio; hs-CRP, high-sensitivity C-reactive protein; LDL, low-density lipoprotein; Ref, reference; TC, total cholesterol; TG, triglyceride.

Association Between Cobalt Levels and AVF Dysfunction Based on Multivariable Cox Regression Analyses

We performed multivariable Cox regression analysis to further investigate the relationship between blood cobalt levels and AVF dysfunction (Table 3). After adjusting for potential confounders such as sex, age, smoking, alcohol consumption, diabetes and cardiovascular disease (CVD) status, pre-dialysis calcium, pre-dialysis phosphorus, total cholesterol, low-density lipoprotein, and hs-CRP, the results indicated that patients undergoing MHD with blood cobalt levels exceeding the normal threshold (>0.7 μg/L) had a 1.66-fold increased risk of AVF dysfunction (95% CI, 1.12-2.45; P = 0.01). Compared with patients in the first quartile of blood cobalt levels, those in the fourth quartile had a 2.90-fold higher risk of AVF dysfunction (95% CI, 1.65-5.10; P < 0.001). Additionally, older age, higher serum phosphorus levels, and the presence of diabetes were identified as independent risk factors for AVF dysfunction in patients undergoing MHD (P < 0.05).

Table 3.

Multivariable Cox Regression Analysis of Cobalt and AVF Dysfunction

Variables	Model 1		Model 2		Model 3
	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P
Age	1.018 (1.007-1.030)	0.001	1.02 (1.01-1.03)	0.005	1.02 (1.01-1.03)	0.004
Phosphorus	1.331 (1.059-1.672)	0.014	1.46 (1.16-1.84)	0.001	1.52 (1.17-1.93)	0.02
Diabetes	1.57 (1.16-2.12)	0.003	1.48 (1.09-2.01)	0.012	1.58 (1.15-2.18)	0.003
Cobalt	1.53 (1.22-1.93)	<0.001	1.56 (1.24-1.97)	< 0.001	1.56 (1.23-1.99)	<0.001
Cobalt range
0.1-0.7 (normal)	1 (Ref)		1 (Ref)		1 (Ref)
> 0.7 (high)	1.63 (1.11-2.39)	0.013	1.67 (1.13-2.45)	0.01	1.66 (1.12-2.45)	0.01
Cobalt quartile
Q1	1 (Ref)		1 (Ref)		1 (Ref)
Q2	2.02 (1.14-3.58)	0.017	2.01 (1.13-3.59)	0.018	2.05 (1.14-3.66)	0.02
Q3	2.04 (1.15-3.62)	0.015	2.01 (1.12-3.61)	0.019	2.07 (1.15-3.74)	0.02
Q4	2.82 (1.62-4.89)	<0.001	2.81 (1.61-4.91)	< 0.001	2.90 (1.65-5.10)	<0.001
Trend.test	1.34 (1.14-1.57)	<0.001	1.34 (1.14-1.57)	< 0.001	1.35 (1.14-1.59)	< 0.001

Note: Model 1: Unadjusted. Model 2: Adjusted for sex, age, smoking, alcohol consumption, and diabetes status. Model 3: Adjusted for sex, age, smoking, alcohol consumption, diabetes status, pre-dialysis calcium, pre-dialysis phosphorus, TC, LDL, hs-CRP, medications (antiplatelet agents, lipid-lowering drugs, and phosphate binders), and CVD.

Abbreviations: AVF, arteriovenous fistula; CI, confidence interval; CVD, cardiovascular disease; hs-CRP, high-sensitivity C-reactive protein; HR, hazard ratio; LDL, low-density lipoprotein; Ref, reference; TC, total cholesterol.

We also conducted a restricted cubic spline analysis using R to explore the dose-response relationship between blood cobalt and AVF dysfunction (Fig S1). However, no significant curvilinear relationship was observed between cobalt levels and fistula dysfunction (nonlinear, P = 0.231).

Survival Analysis of Patients Stratified by Cobalt Quartiles in Relation to AVF Dysfunction

Kaplan–Meier survival analysis of patients categorized into different quartiles based on blood cobalt levels revealed statistically significant differences in the incidence of AVF dysfunction among the quartiles. Patients in the fourth quartile of blood cobalt levels exhibited the highest cumulative incidence of AVF dysfunction (Fig 2), consistent with the findings of the univariate and multivariable Cox regression analyses.

Figure 2.

Cumulative incidence of AVF dysfunction by cobalt quartiles. AVF, arteriovenous fistula; CI, confidence interval; Co, cobalt; HR, hazard ratio.

Stratified Analysis

We explored the potential heterogeneity of the relationship between blood cobalt levels and AVF dysfunction across different populations using a stratified analysis. After adjusting for confounding factors such as sex, age, smoking, alcohol consumption, diabetes and CVD status, pre-dialysis serum calcium, pre-dialysis serum phosphorus, total cholesterol, low-density lipoprotein, and hs-CRP, subgroup analyses were conducted based on sex, age (≥ 65 years), body mass index, smoking status, diabetes, serum Alb level (≥35 g/L), and medication use (including antiplatelet agents, lipid-lowering drugs, and phosphate binders). The results (Fig 3) indicated that the effect of blood cobalt levels on AVF dysfunction remained consistent across all subgroups, with P values for the interaction tests exceeding 0.05, indicating no statistically significant interaction effects.

Figure 3.

Forest plot of blood cobalt and arteriovenous fistula dysfunction. Alb, albumin; BMI, body mass index; CI, confidence interval; DM, diabetes; hs-CRP, high-sensitivity C-reactive protein.

Discussion

Cobalt is an essential trace element that is widely distributed in nature. Cobalt metabolism in the body mainly occurs through absorption in the digestive tract; cobalt is then transported by the blood to the liver for metabolism. Cobalt is primarily excreted in urine and feces, regardless of the route of exposure.¹⁷ Despite theoretical predictions suggesting cobalt accumulation in patients undergoing MHD, empirical evidence remains inconclusive. Stojsavljević et al¹⁸ observed lower serum cobalt levels in patients undergoing MHD compared with healthy individuals, contrasting with findings from 2 studies that reported markedly higher blood cobalt concentrations in this patient group (P < 0.001).^12,13 Moreover, studies by Finley et al¹⁹ and Paustenbach et al²⁰ on oral cobalt supplementation, conducted in the general population and in patients with kidney failure, demonstrated that serum cobalt concentrations significantly exceeded those within the blood, a phenomenon attributed to the limited absorptivity of cobalt in red blood cells. The clearance rate of serum cobalt also surpasses that of red blood cells, with serum cobalt levels displaying substantial fluctuations during supplementation. These variances likely stem from individual differences in gastrointestinal cobalt absorption and timing of cobalt intake relative to blood sampling. Given these complexities, blood cobalt concentration has been proposed as a superior measure for assessing long-term cobalt exposure, whereas serum cobalt levels are indicative of recent or acute exposure.²¹ Consequently, this study used blood cobalt levels as a metric to investigate its relationship with AVF dysfunction in patients undergoing MHD.

In this study, we aimed to investigate the relationship between cobalt levels and AVF dysfunction and found that male sex, older age, diabetes, serum phosphorus, blood cobalt, and Alb levels were associated with the incidence of AVF dysfunction. After adjusting for potential confounding variables, elevated blood cobalt levels were identified as an independent risk factor for AVF dysfunction in patients undergoing MHD (adjusted hazard ratio, 1.56; P < 0.001). Notably, this association was more pronounced in patients in the highest cobalt level quartile (adjusted hazard ratio, 2.90; P < 0.001). Additionally, age, diabetes, and serum phosphorus levels were identified as independent risk factors of AVF dysfunction in this patient population, consistent with previous studies.^6,22

Cobalt is a crucial element in the treatment of anemia and alleviation of peripheral neuropathy symptoms in patients with chronic kidney disease because of its role as an essential component of vitamin B₁₂.²³ Despite its clinical benefits, research investigating cobalt levels in patients with chronic kidney disease who are undergoing MHD remains limited. Data from the National Health and Nutrition Examination Survey (2003-2012)²⁴ showed a positive correlation between urinary cobalt excretion and estimated glomerular filtration rate, suggesting that elevated cobalt levels might correlate with diminished estimated glomerular filtration rate. Subsequently, experimental studies²⁵ on male Wistar rats revealed that cobalt chloride exposure could impair glomerular podocyte morphology and slit diaphragm function by stabilizing hypoxia-inducible factor-1α, leading to acute kidney injury and the progression of chronic kidney disease. Excessive cobalt intake has been linked to toxic reactions and an elevated risk of CVDs and cerebrovascular diseases, particularly in the context of compromised renal function. A cross-sectional study identified a nonlinear positive correlation between serum cobalt levels and CVDs, where higher levels were associated with an increased incidence of these conditions.²⁶ Another study indicated that exposure to multiple metals, including cobalt, was positively associated with CVD risk in US adults.²⁷ Cobalt compounds can also inhibit mitochondrial function, enhance reactive oxygen species production, induce oxidative stress, and subsequently cause endothelial damage and atherosclerosis development.²⁸ However, no study has investigated the specific effect of cobalt on vascular function in patients undergoing MHD, particularly with respect to AVF dysfunction.

AVF dysfunction results from a multifaceted interplay of factors. The hallmark pathological lesions are neointimal hyperplasia and impaired outward vascular remodeling,²⁹ the development of which is intricately linked to underlying patient conditions, such as diabetes mellitus, hypertriglyceridemia, and a prothrombotic state.³⁰ The initial damage to ECs and vascular smooth muscle cells is precipitated by upstream factors such as surgical trauma, hemodynamic shear stress, and uremic toxins. This damage subsequently engages downstream processes, including oxidative stress, inflammatory responses, and cell migration and proliferation, cumulatively leading to luminal stenosis, neointimal hyperplasia, and thrombosis.³¹ Patients undergoing MHD commonly exhibit CVD and widespread use of antithrombotic agents. The fundamental pathophysiology of CVD is chronic inflammation and endothelial dysfunction,³² a systemic state that exacerbates the local inflammatory response during arteriovenous anastomosis. Consequently, inflammatory mediators such as C-reactive proteins and interleukins inflict damage on the vascular endothelium.³³ EC activation enhances the expression of growth factors that drive the translocation of vascular smooth muscle cells from the medial to intimal layers, rapidly contributing to the formation of a new intimal layer.³⁴ At the same time, dysfunctional endothelium diminishes the release of nitric oxide while increasing the production of inflammatory cytokines and vasoconstrictors.³³ This imbalance inhibits outward remodeling of the AVF, thereby leading to maturation failure. Moreover, CVDs such as heart failure and severe arrhythmias can reduce cardiac output and cause systemic hemodynamic instability. The resulting hypotension diminishes the blood flow through the AVF, and this sluggish flow predisposes the circuit to thrombosis.³⁴ However, the European Vascular Access and KDOQI guidelines do not recommend the routine use of antiplatelet drugs (including aspirin and clopidogrel), statins, and fish oil. The effects of these drugs are not precise, and their risks outweigh their potential benefits.^14,15 A prospective study in South Korea enrolled 881 patients receiving hemodialysis using internal fistulas (including AVFs and arteriovenous grafts), among whom 241 (27.4%) received aspirin. The median follow-up period was 30 months. During follow-up, 180 patients (20.4%) experienced the main outcome event (vascular access failure). Multivariable analysis showed that the hazard ratio for the incidence of vascular access failure among aspirin users was 0.89 (95% CI, 0.62-1.27; P = 0.51). Aspirin has no protective effect on internal fistulas (including AVFs and arteriovenous grafts).³⁵ A meta-analysis of the use of adjuvant drug therapy to increase AVF patency (including AVFs and arteriovenous grafts) included 15 clinical trials involving 2,230 patients who underwent internal fistula surgery. The adjuvant drugs used in the clinical trials were aspirin, ticlopidine, and dipyridamole combined with aspirin, warfarin, fish oil, clopidogrel, thiopyrazolone, and recombinant human type I pancreatic elastase (PRT-201). The authors compared studies of aspirin and placebo, ticlopidine and placebo, fish oil and placebo, and clopidogrel and placebo and discovered that there was insufficient evidence to determine whether there was a difference between the effects of placebo and other treatments (such as aspirin, fish oil, clopidogrel, PRT-201, dipyridamole, dipyridamole plus aspirin, warfarin, and thiopyrazolone) on AVF patency.³⁶

In summary, AVF dysfunction may occur due to various reasons, including elevated blood cobalt levels as an independent risk factor for AVF dysfunction. Several potential mechanisms have been proposed. First, cobalt exposure induces cytotoxic effects, including apoptosis, necrosis, and inflammation.^37,38 This inflammation is mainly driven by an extensive release of macrophage migration inhibitory factor, leading to the recruitment of inflammatory cells to the neointima and subsequent vascular wall thickening. Second, cobalt may induce reactive oxygen species formation through Fenton-like reactions, thereby increasing oxidative stress.³⁹ This is supported by elevated markers of oxidative damage in the venous tissue of the AVF outflow tract,⁴⁰ and persistent excessive superoxide anions can rapidly oxidize and inactivate nitric oxide, further leading to the uncoupling of endothelial nitric oxide synthase to produce superoxides, which are a primary cause of endothelial dysfunction due to oxidative stress.⁴¹ Additionally, cobalt ions may act as angiogenesis stimulators, mimicking hypoxic conditions through hypoxia-inducible factor-1α, thereby promoting the expression of angiogenic genes such as vascular endothelial growth factor and fibroblast growth factors, which aid in EC polarization, migration, and homing.^42,43 The interaction between vascular endothelial growth factor and its 2 receptors plays a significant role in the inflammatory and proliferative processes in AVF.^44,45 However, the mechanisms linking elevated blood cobalt levels to AVF dysfunction in patients undergoing MHD remain unclear. This necessitates further verification through rigorous basic research using cellular and animal models to validate these hypotheses.

This study substantiated that advanced age, diabetes, and elevated serum phosphorus levels are additional independent risk factors for AVF dysfunction in patients undergoing MHD. These observations were consistent with those of previous studies.^6,22 As patients age, the proliferative capacity of their vascular ECs diminishes and oxidative stress increases, thus fostering apoptosis and cellular aging. This series of events causes local inflammation and free radical damage, undermining the integrity and functionality of ECs.⁴⁶ Diabetes intensifies inflammatory responses by augmenting the expression of proinflammatory cytokines such as interleukin-6 and tumor necrosis factor-α, which are stimulated by hyperglycemia. The resultant inflammation may be attributable to uremia-related systemic or localized inflammation secondary to surgical intervention.^29,47 Both inflammatory mechanisms are implicated in the formation of scars, stenosis, and thrombosis within the AVF, thereby impairing its function and patency.⁸ Hyperphosphatemia frequently occurs in patients undergoing MHD and acts as a precipitant of vascular calcification. Elevated levels of extracellular inorganic phosphate promote phenotypic transformation of vascular smooth muscle cells, triggering matrix vesicle secretion and activating apoptotic pathways, ultimately leading to vascular smooth muscle cell dysfunction and vascular calcification.⁶ In turn, vascular calcification accelerates vascular injury and thrombosis and plays a pivotal role in AVF dysfunction.

This study has some limitations. First, it is based on a retrospective observational design where trace element levels were measured only at the point of enrollment, without subsequent dynamic monitoring. This limitation precluded determination of whether factors such as residual kidney function, dialysis treatment, dietary habits, and lifestyle changes during the observation period could have led to variations in trace element levels, potentially affecting the outcomes of the study. Second, other significant variables known to influence AVF patency, including the vascular condition, surgical technique, and vascular histological characteristics, were not considered. This omission restricts the ability to adjust for these potential confounding factors related to variations in blood trace element levels. Third, although there is a hypothesized association between cobalt levels and AVF dysfunction, it remains speculative. Detailed research at both the cellular and tissue levels is imperative for a thorough understanding of the underlying mechanisms.

Supplementary Materials

Supplementary File (TIF). Figure 1.