Blood osteoprotegerin is associated with arteriovenous access thrombosis in hemodialysis patients
Article information
Abstract
Background
A functioning arteriovenous (AV) access is essential for hemodialysis efficiency and the quality of life in hemodialysis patients. Blood osteoprotegerin (OPG) and soluble receptor activator of nuclear factor kappa B ligand (RANKL) have been linked to cardiovascular diseases and vascular calcification. This study investigated the relationship between blood OPG, RANKL, and the occurrence of AV access thrombosis.
Methods
This prospective cohort study was conducted from August 2016 to August 2021 and included patients undergoing prevalent hemodialysis in two hospital-based hemodialysis units. Cox proportional hazards models and Kaplan-Meier analysis were used to evaluate the association between blood OPG, RANKL, and AV access (AV fistula [AVF] and AV graft [AVG]) outcomes.
Results
A total of 333 hemodialysis patients were enrolled, with an AV access thrombosis rate of 22.2%. Cox regression identified several factors associated with AV access thrombosis: AV access type (AVF vs. AVG; hazard ratio [HR], 0.24; p < 0.001), C-reactive protein (HR, 1.07; p = 0.002), and log-transformed OPG (HR, 5.52; p = 0.005). Subgroup analysis revealed high log-transformed OPG and RANKL were associated with AVF thrombosis (HR, 10.77; p = 0.002 and HR, 3.26; p = 0.009, respectively), while high C-reactive protein increased the risk of AVG thrombosis (HR, 1.31; p < 0.001). Kaplan-Meier analysis showed that patients with AVF in the highest tertile of log OPG (>402 pg/mL) had the highest AVF thrombosis incidence (p = 0.03).
Conclusion
High blood OPG was associated with AV access thrombosis, particularly in the AVF.
Introduction
The prevalence of end-stage kidney disease (ESKD) continues to increase globally owing to improved patient survival, higher prevalence of risk factors for ESKD, and increasing availability of kidney replacement therapy [1]. In patients undergoing hemodialysis (HD), vascular access is recognized as a ‘lifeline’ and provides adequate blood flow for HD treatment. According to the current guidelines, arteriovenous (AV) access, including AV fistulas (AVFs) and prosthetic AV grafts (AVGs), is preferred over central venous catheters in prevalent HD patients due to fewer vascular access-related events (e.g., infections, thrombotic and non-thrombotic complications) [2]. Thrombosis is a leading cause of AV access failure, potentially leading to missed dialysis sessions, hospitalizations, the need for endovascular or surgical thrombectomy, and all-cause mortality [3,4].
Several risk factors for vascular access thrombosis have been identified, including older age, female sex, diabetes, obesity, peripheral vascular disease, atherosclerosis, pre-dialysis hypotension, smoking, the use of an AVG, a distal fistula site, hypoalbuminemia, and elevated inflammatory biomarkers [5–8]. Recent studies have highlighted the association between vascular calcification and AV access patency [9,10]. Vascular calcification is highly prevalent in patients with ESKD and is a well-known risk factor for cardiovascular events and all-cause mortality in this population [11,12].
Osteoprotegerin (OPG), a member of the tumor necrosis factor (TNF) receptor superfamily, functions as a soluble decoy receptor for the receptor activator of nuclear factor kappa B ligand (RANKL) and regulates osteoclast activation in bone remodeling [13]. Increasing evidence suggests the involvement of the OPG/receptor activator of nuclear factor kappa B (RANK)/RANKL axis in coronary artery disease, atherosclerotic disorders, and vascular calcification [14–17]. Some studies have demonstrated elevated OPG concentrations in patients with calcified vessels, suggesting an interacting pathway that increases OPG expression during the development of vascular calcification [18,19]. Additionally, many studies have observed a potential link between the OPG/RANK/RANKL axis and unprovoked venous thromboembolic disease and ischemic stroke [20,21]. A previous experimental study also showed that OPG can regulate platelet adhesion to von Willebrand factor secreted from endothelial cells [22]. However, studies exploring the relationship between blood OPG, soluble RANKL, and AV access outcome are limited. Hence, in this study, we prospectively evaluated the association between AV access thrombosis and clinical factors including OPG and RANKL blood levels in a cohort of stable HD patients.
Methods
Study design and populations
This prospective study was conducted at two hospital-based HD units from August 2016 to January 2017. Patients were included if they 1) were over 20 years of age, 2) had been receiving maintenance HD treatment for at least 90 days, and 3) were using an AVF or AVG as vascular access for HD. The patients who had undergone vascular intervention within 3 months prior to enrollment were excluded. All of the patients received HD treatment three times a week, with each HD session lasting 3.5 to 4 hours, with a blood flow rate of 250 to 300 mL/min and dialysate flow of 500 mL/min. The patients were followed until the occurrence of AV access thrombosis, the end of the study on August 31, 2021, transfer to other HD units or death.
The study protocol was approved by the Institutional Review Board of Kaohsiung Medical University (No. KMUHIRB-E(I)-20160065), and all patients provided signed informed consent.
Demographic, medical, and laboratory data collection
We obtained information including demographics, comorbidities, and laboratory data from the medical records of the dialysis units. Blood samples were collected as soon as possible after enrollment and at the beginning of the week following an overnight fast from the patients via the AV access before the scheduled HD session and stored at –80 °C. Outcomes of the AV access were prospectively collected by reviewing the medical records. The baseline variables examined for correlations with thrombotic events of the AV access included the following: age, sex, smoking status, body mass index (BMI), duration of dialysis treatment, presence of diabetes, presence of hypertension, presence of cardiovascular disease (CVD), type of AV access being used, use of antiplatelet agents, use of warfarin, and levels of hemoglobin, albumin, potassium, calcium × phosphorus products, C-reactive protein (CRP), ferritin, Kt/V, OPG, and RANKL.
The definition of demographics and medical data
Diabetes was defined as a blood glycosylated hemoglobin level above 6.5% or the use of antidiabetic drugs. Hypertension was defined as the current use of antihypertensive medication. The definition of CVD is a patient having coronary artery disease (a history of percutaneous coronary intervention or coronary artery bypass grafting, or diagnosed by a cardiologist) or cerebrovascular disease (a history of cerebrovascular events or diagnosed by a neurologist). Smoking history was defined as current or past cigarette smoking. BMI was calculated as body weight in kilograms divided by the square of height in meters. The use of antiplatelet or warfarin was defined based on the patient’s current use of these drugs at the time of enrollment.
Measurement of osteoprotegerin and RANKL levels
Circulating levels of OPG and RANKL were measured with commercial multiplex enzyme-linked immunosorbent assay kits (MILLIPLEX MAP, Merck Millipore) following the manufacturer’s protocols. In brief, capture microspheres coated with monoclonal antibodies were applied to the wells, followed by standards and plasma samples. After incubation and washing, biotinylated secondary antibodies were added and then incubated with streptavidin-bound fluorescent protein. After washing, microsphere-containing precipitates were resuspended for analysis in a MAGPIX analyzer (Luminex Corp.). Results were processed by Milliplex Analyst software (MilliporeSigma) and reported in pg/mL.
Definition of arteriovenous access thrombosis
Routine monitoring and surveillance protocols for detecting AV dysfunction were conducted according to the KDOQI guidelines. All of the patients were referred for diagnostic angiography and intervention if clinical monitoring detected a sudden cessation of access function. AV access thrombosis was defined as the abrupt occlusion of AV access requiring an intervention. The outcome was defined as the time to the first AV access thrombotic event. We refer patients for vascular intervention only when a thrombotic event causes a complete cessation of blood flow in their AV access.
Statistical analysis
The baseline characteristics of the study patients were compared based on the occurrence of AV access thrombosis. Differences between the two groups were assessed using the chi-square test for categorical variables, the independent t test for continuous variables with approximately normal distribution, and the Mann-Whitney U test for continuous variables with skewed distribution. Data are expressed as percentages for categorical variables or mean ± standard deviation. Potential risk factors for AV access thrombosis were assessed using univariable Cox proportional hazards models. Significant risk factors for AV access thrombosis were selected by forward selection in multivariable Cox proportional hazards analysis. A forward selection procedure sequentially enters variables into the model if they meet the entry criterion. The first variable selected is the one with the largest positive or negative correlation with the dependent variable. The following variable chosen is the one with the largest partial correlation not yet included in the equation. The procedure terminates when no additional variables meet the entry criterion. To evaluate whether distinct variables contributed to AV access thrombosis according to the type of AV access being used, subgroup analysis categorized by AVF or AVG was conducted. The associations between tertiles of log OPG and log RANKL with AVF outcomes were analyzed using Kaplan-Meier curves, and the log-rank test was used to assess differences between the groups. All of the analyses were performed using IBM SPSS version 26 for Windows (IBM Corp.).
Results
Comparisons of baseline characteristics between arteriovenous access outcomes in all patients
A total of 333 patients were included in this cohort study. The mean age of the patients was 59.4 ± 11.5 years, and 174 (52.3%) were male. Of these patients, 290 (87.1%) were using an AVF, and 43 (12.9%) were using an AVG. During a mean follow-up of 5.5 ± 2.1 years, 74 cases of AV access thrombosis occurred, accounting for 22.2% of the total cohort. The clinical characteristics and laboratory data of the study groups, categorized by the occurrence of AV access thrombosis, are summarized in Table 1.
Comparison of clinical characteristics between patients with or without AV access thrombosis
In the patients with AV access thrombosis, the thrombosis rate was higher in those who used an AVG (23 of 43 AVG users, 53.5%) than in those who used an AVF (51 of 290 AVF users, 17.6%). In addition, the patients who experienced an AV access thrombotic event were more likely to have CVD (32.4% vs. 21.2%, p = 0.046), use an AVG (31.1% vs. 7.7%, p < 0.001), use warfarin (8.1% vs. 2.3%, p = 0.03), and have higher levels of hemoglobin (11.0 ± 1.0 g/dL vs. 10.7 ± 1.3 g/dL, p = 0.04), CRP (3.5 ± 4.5 mg/L vs. 2.1 ± 4.2 mg/L, p = 0.02), and OPG (1,168.5 ± 588.9 pg/mLvs. 977.5 ± 442.4 pg/mL, p = 0.01).
Determinants of arteriovenous access outcomes in all patients
Risk factors for AV access thrombosis are presented in Table 2. Multivariable Cox regression model showed that the patients using an AVF had a significantly lower risk of AV access thrombosis than those using an AVG (AVF vs. AVG: adjusted hazard ratio [HR], 0.24; 95% confidence interval [CI], 0.14–0.41; p < 0.001). In addition, the patients with CVD (HR, 2.13; 95% CI, 1.08–4.21; p = 0.03), high CRP levels (HR, 1.07; 95% CI, 1.03–1.11; p = 0.002), high log OPG (HR, 5.52; 95% CI, 1.67–18.22; p = 0.005), and those who used warfarin (HR, 2.61; 95% CI, 1.03–6.63; p = 0.04) were associated with an increased risk of AV access thrombosis.
Determinants of AV access survival in HD patients
Comparisons of the clinical characteristics between arteriovenous access outcomes by the type of arteriovenous access
The relationships between risk factors and AV access thrombosis were explored through subgroup analysis, as shown in Table 3. Among the patients using an AVF, there were no significant differences in age, sex, or comorbidities between those with and without AVF thrombosis. However, the patients with AVF thrombosis had higher serum OPG levels than those without AVF thrombosis (1,177.2 ± 658.1 pg/mL vs. 979.7 ± 441.9 pg/mL, p = 0.045). In the patients using an AVG, those with AVG thrombosis had higher levels of hemoglobin (11.2 ± 1.1 g/dL vs. 10.2 ± 1.6 g/dL, p = 0.03) and CRP (4.6 ± 4.4 mg/L vs. 1.6 ± 2.0 mg/L, p = 0.006) compared to those without AVG thrombosis.
Comparison of clinical characteristics between patients with or without AV access thrombosis in different types of AV access (AVF or AVG)
Determinants of arteriovenous access outcomes by the type of arteriovenous access
Risk factors for AV access thrombosis categorized by AVF and AVG are presented in Table 4. Multivariable Cox regression model showed that patients with AVF thrombosis had higher levels of log OPG (HR, 10.77; 95% CI, 2.37–48.9; p = 0.002), log RANKL (HR, 3.26, 95% CI, 1.34–7.94, p = 0.009), and BMI (HR, 1.10; 95% CI, 1.02–1.20; p = 0.01) compared to those with AVF patency. In addition, the patients with AVG thrombosis had a higher CRP level (HR, 1.31; 95% CI, 1.15–1.50; p < 0.001), and higher rates of antiplatelet and warfarin therapy than those without AVG thrombosis.
Determinants of AVF and AVG thrombosis in HD patients
Association between blood osteoprotegerin and RANKL and arteriovenous fistula outcomes
Among the patients using an AVF, those in the highest tertile of log OPG (>402 pg/mL) had a higher cumulative incidence of AVF thrombosis compared to those in the lowest (<211 pg/mL) and intermediate (211–402 pg/mL) tertiles of log OPG (p = 0.03) (Fig. 1A). However, log RANKL levels across tertiles were not associated with AVF thrombotic events (Fig. 1B).
Kaplan-Meier curve of the occurrence of AVF thrombosis in hemodialysis patients.
(A) According to the tertile of log-transformed osteoprotegerin (OPG) levels (log-rank p = 0.03). (B) According to the tertile of log-transformed receptor activator of nuclear factor kappa B ligand (RANKL) levels (log-rank p = 0.43).
AVF, arteriovenous fistula.
Discussion
This study suggested that higher blood OPG is associated with an increased risk of AV access thrombosis. Additionally, our results revealed that different types of AV access correspond to distinct risk factors for developing thrombotic events. Specifically, in patients using an AVF, elevated blood levels of OPG and RANKL were strongly associated with AVF thrombosis, whereas increased levels of CRP were associated with thrombosis in patients using an AVG.
A key finding of the study is that the patients with higher levels of OPG and RANKL were at a higher risk of developing AVF thrombosis, while these associations were not observed in the patients using AVG. Furthermore, the highest log OPG tertile was linked to an increased risk of an AVF thrombotic event. Most previous studies highlight that blood levels of OPG are associated with arterial calcification, such as coronary arterial calcification, and mortality in HD patients [16,23,24]. Clinical studies link serum OPG levels with AV access dysfunction or thrombosis, but the results have been inconsistent. Kim et al. [25] reported that serum OPG levels were positively correlated with the degree of AVF stenosis using ultrasonography. On the other hand, Morena et al. [26] found no association between OPG and RANKL levels with AVF thrombosis in a study with 128 HD patients. In a prospective study of 727 patients with HD, Lyu et al. [9] investigated markers of vascular calcification and vascular access complications and suggested that OPG was not significantly associated with the risk of an AVF intervention. Instead, they found that other vascular calcification markers, including fetuin-A, osteopontin, and bone morphogenetic protein 7, were associated with an AVF intervention. There are various possible explanations for the link between OPG and AVF thrombosis. As a marker of vascular calcification, higher OPG levels are more likely to be associated with preexisting AV access calcification, which has been suggested to predict AVF failure in HD patients [10,27]. In addition, prior research has shown that OPG is expressed in normal vasculature and is more predominant in advanced atherosclerotic plaques [19], suggesting that circulating OPG may originate from these vascular plaques. The presence of atherosclerosis has been linked to an increased risk of AVF failure [7,8], establishing OPG as a potential biomarker associated with AVF thrombosis [15]. Despite the strong correlation between OPG/RANKL and vascular calcification, there is currently no literature focusing on the direct relationship between OPG and AVF or venous calcification, possibly because AVF calcification is difficult to detect, thus limiting research on this topic. Noteworthy, OPG has been associated with many risk factors that cause vascular calcification and AVF thrombosis, such as old age, diabetes mellitus, and chronic inflammation [5,8,28]. These risk factors may also contribute to eventual AVF thrombosis. Among these risk factors, inflammation, indicated by elevated inflammatory markers and cytokines such as CRP, interleukin-6, and TNF-α, is particularly relevant to AV access thrombosis [29,30]. OPG, belonging to the TNF receptor family, is stimulated by pro-inflammatory agents such as TNF in vascular smooth muscle and endothelial cells [31,32]. Furthermore, emerging evidence has revealed the complex interplay between inflammation and vascular thrombosis [33]. Therefore, OPG levels may link to AVF thrombotic events under inflammatory conditions. However, the association was not observed in patients using AVGs in the present study, which is consistent with Lyu et al. [9], who reported no significant correlation between calcification markers, including OPG, and AVG intervention. Generally, stenosis induced by neointimal hyperplasia at the venous anastomoses is the primary cause of AVG thrombosis. The histology of neointimal hyperplasia comprises myofibroblasts, extracellular matrices, pro-inflammatory cells, neovasculature, and various growth factors and cytokines [30]. Because of the strong correlation of neointimal hyperplasia-induced stenosis with AVG thrombosis, which is less related to vascular calcification, this may partly explain the insignificant association between OPG and AVG thrombosis. Moreover, the small sample size of AVG users in our study may have resulted in reduced statistical power. Further research is needed to investigate this association in AVG users.
In this study, a positive correlation was found between blood RANKL and thrombosis of the AVF. However, the tertiles of blood RANKL were not significantly associated with AVF survival in the Kaplan-Meier analysis. Prior research has suggested the pathogenic effects of the RANKL/OPG/RANK axis in atherosclerosis [15,34], and vascular calcification has been linked to RANKL activity [35]. Nevertheless, the use of serum RANKL as a biomarker for CVD remains controversial. Lieb et al. [36] observed no association between serum RANKL and coronary artery calcification, incident CVD, or mortality in the general population. Similarly, in another cohort study conducted by Spartalis et al. [16] involving 80 HD patients, RANKL levels did not correlate with the presence or progression of vascular calcification in the abdominal aorta and muscular arteries. However, the Bruneck Study, including 909 participants, found that baseline serum RANKL levels strongly predicted CVD [37]. Generally, OPG levels are significantly higher than RANKL levels in circulation, and previous research suggests that OPG is a more stable and reliable indicator of the RANKL/OPG/RANK axis than soluble RANKL [15].
Another important finding in this study is that the patients with elevated CRP levels had a higher risk of developing AV access thrombosis, and the risk was more pronounced among the patients who used an AVG. In general, patients using AV access experience localized and systemic inflammation due to factors such as oxidative stress, wall shear stress, endothelial dysfunction, needle cannulation, AV access surgery, graft material biocompatibility, and uremic status [30,38,39]. Several clinical studies have reported an association between inflammation in HD patients with AV access thrombosis. Chou et al. [29] found that higher serum CRP levels in chronic HD patients were linked to a higher risk of vascular access thrombosis. Notably, compared to AVF users, those with an AVG have been reported to be more likely to have low-grade inflammation at baseline [40], which may lead to an increased risk of AV access thrombosis. These findings are consistent with our study, as suggested by the association of elevated CRP levels with AVG thrombosis. Although the timing of blood sample collection in the present study may not have aligned with that of AV access thrombosis, potentially limiting its ability to reflect the inflammatory status at the time of the thrombosis, our findings still indicated that the HD patients with elevated baseline CRP levels, suggesting chronic inflammation, were at an increased risk of AV access thrombosis, and particularly those with AVG.
The present study has several limitations. First, the small sample size may have decreased the statistical power, particularly in stratified analyses of the patients using AVGs. Second, an observational study may not adequately adjust for all potential confounders. For example, we didn’t have detailed data on AV access vintage, which may have a potential link with AV access thrombosis. Third, the absence of serial measurements of serum OPG and RANKL meant that we could not analyze associations between their fluctuations and the development of AV access thrombosis. Fourth, while we observed an association between higher OPG levels and AVF thrombosis, it is still unclear whether increased OPG contributes directly to the development of a thrombosed AVF or is simply a compensatory response to atherosclerotic disease. In addition, we didn’t perform other imaging studies, such as sonography, to confirm the presence of AV access calcification, despite the significant correlation between OPG and calcification found in previous research. Further investigations are needed to understand the underlying mechanisms of this observation.
In conclusion, our results identified high blood OPG levels associated with AV access thrombosis, particularly in the AVF, suggesting that OPG may be a potential biomarker in clinical use. Further research may be required to investigate the underlying pathophysiological mechanisms for preventing AV access thrombosis.
Notes
Conflicts of interest
All authors have no conflicts of interest to declare.
Funding
The study was funded by grants from the Ministry of Science and Technology, Taiwan (MOST 111-2314-B-037-032-MY3), Kaohsiung Medical University Hospital, Taiwan (KMUH111-1M60, KMUH111-1R73, KMUH111-1M09, KMUH110-0M13, KMUH110-0M73, KMUH110-0M12, and KMUH-DK(C)112001), and Kaohsiung Medical University, Taiwan (KT113P006, KT112P012, NYCUKMU-112-I006, NHRIKMU-111-I003, and NHRIKMU-111-I001). This study is supported partially by the Kaohsiung Medical University Research Center Grant (KMU-TC112B04), KMUH-DK(C)112001, and KMUH-DK(C)113003. The funding sources did not play any role in the design or conduct of the study, collection, management, analysis, interpretation of the data, or preparation, review, or approval of the manuscript.
Acknowledgments
The authors express gratitude to all the patients and healthcare providers who participated in this study, as well as to Kaohsiung Medical University Hospital and Kaohsiung Municipal Siaogang Hospital for their support in conducting this research.
Data sharing statement
The data presented in this study are available from the corresponding author upon reasonable request.
Authors’ contributions
Conceptualization: TLC, YHL, PHW, SCC
Formal analysis: TLC, PHW, SCC
Funding acquisition: PHW
Methodology: MCK, YWC, PHW
Writing–original draft: TLC, PYW, JCH, YCT
Writing–review & editing: YLH, PHW, SCC
All authors read and approved the final manuscript.