Aortic Arch Calcification as a Predictor of Repeated Arteriovenous Fistula Failure within 1-Year in Hemodialysis Patients

Objectives The aim of the study was to identify the factors associated with repeated arteriovenous fistula (AVF) failure within 1-year, especially the impact of aortic arch calcification (AAC) on patency of AVF. Materials and Methods We retrospectively assessed chest radiography in hemodialysis patients who had undergone initial AVF. The extent of AAC was categorized into four grades (0–3). The association between AAC grade, other clinical variables, and repeated failure of AVF was then analyzed by binary logistic regression analysis. Results This study included 284 patients (158 males, mean age 61.7 ± 13.1 years). Patients with higher AAC grade were older, had more frequently diabetes mellitus and cardiovascular disease, had lower diastolic blood pressure, and had higher corrected calcium and lower intact parathyroid hormone levels. In multivariate analysis, the presence of higher AAC grade (odds ratio (95% confidence interval): 2.98 (1.43–6.23); p = 0.004), lower mean corrected calcium (p = 0.017), and mean serum albumin level (p = 0.008) were associated with repeated failure of AVF. Conclusions The presence of higher AAC grade, lower mean corrected calcium and mean serum albumin level were independently associated with repeated AVF failure within 1 year in hemodialysis patients.


Introduction
Clinical practice guidelines recommend the arteriovenous fistula (AVF) as the preferred form of dialysis access owing to good long-term patency and a low incidence of complications [1]. Nonetheless, the AVF also has the disadvantages of longer maturation time and higher primary failure rate, which is usually caused by early thrombosis [2,3]. A recent report pooled estimated primary AVF failure and 1-year primary patency to be 23% and 60% respectively [4]. Although percutaneous transluminal angioplasty (PTA) has generally replaced surgical procedures to be recognized as a standard approach to treat these stenotic lesions, its benefit is attenuated by a high re-stenotic rate within 6 months [1]. Of note, vascular access dysfunction is one of the leading causes of morbidity and mortality in hemodialysis patients and is also responsible for the high percentage of hospitalizations [5].
The main factors affecting primary patency loss of AVF are age, gender, the presence of cardiovascular disease, the preoperative diameter of the artery and vein, and so forth [3]. Besides, the major pathology cause in the development of AVF failure among hemodialysis patients is intimal hyperplastic stenosis [6]. Notably, vascular calcification including vascular access calcification and aortic arch calcification (AAC) are commonly seen in dialysis patients, and preexisting arterial calcification of the vascular access is also associated with worse AVF outcome in these populations [7][8][9][10]. Preexisting calcification may stiffen the artery and impair outward remodeling of the inflow artery, and thus limits its dilatation and resulting increase in blood flow after AVF operation [11]. A recent report even established that AAC, which is highly correlated with vascular access calcification, might predict primary patency loss of AVF [12]. Nevertheless, little has been published about whether repeated AVF failure is also affected by preexisting AAC.
The extent of AAC can be measured with electron beam computer tomography (EBCT) or multi-slice computer tomography (MSCT). However, these methods cannot be routinely performed due to the relatively high cost of examination and exposure to a high radiation dose [13,14]. On the contrary, chest radiography is a non-invasive and inexpensive tool for the identification of AAC. Moreover, compared with plain hand film, chest radiography is routinely performed in all hemodialysis patients for cardiothoracic ratio assessment. Besides, the grading of vascular calcification of hand film could only be categorized into the presence of calcification or non-calcification [10]. Thus, chest radiography is recognized as a cost-effective and precise screening tool that can be used in place of plain hand film in examining vessel calcification in hemodialysis populations.
On the basis of these views, we hypothesized that the grading of AAC could be related to future development of repeated AVF failure. Therefore, the objective of this study was to identify factors predictive of repeated AVF failure within 1-year, especially the impact of AAC on AVF patency after an initial AVF creation among hemodialysis patients.

Ethics Statement.
This study protocol was approved by the Institutional Review Board of Yuan's General Hospital, and the methods were performed in accordance with the Declaration of Helsinki. Requirement for patient consent was waived due to the minimal participant risk and retrospective nature of the study.

Study Population.
All consecutive end-stage renal disease (ESRD) patients over 20 years of age who underwent first AVF surgery and started hemodialysis at dialysis centers of Yuan's general hospital between January 2006 and June 2015 were initially included in this retrospective observational study. Among these enrolled patients, subjects with followup duration of less than 1 year ( = 39), those who had no chest radiography records ( = 5), and those had received arteriovenous graft (AVG) recreation and permanent catheter placement within 1-year after AVF surgery ( = 7) were excluded. Finally, 284 patients were included in this study.

Clinical Variables.
Clinical data including age, gender, comorbid diseases (diabetes mellitus, hypertension, and cardiovascular disease), body mass index (BMI), blood pressure (systolic and diastolic), medication, laboratory test and the AVF characteristics were extracted from medical records. Cardiovascular disease was defined as a history of coronary, cerebrovascular, or peripheral vascular disease: coronary disease was defined as a history of angioplasty, coronary artery bypass grafts, myocardial infarction, or typical angina and cerebrovascular disease as a history of cerebrovascular incidents such as cerebral bleeding or infarction, while peripheral vascular disease was defined as a history of prior revascularization procedure, amputation for ischemia or gangrene, or an ankle-brachial pressure index of <0.9. Diabetes was defined if fasting plasma glucose levels > 6.99 mmol/L, or glycated hemoglobin (HbA1c) > 6.5%, or if the patient was currently using hypoglycemic agents. Hypertension was defined if they had SBP ≥ 140 mm Hg, DBP ≥ 90 mmHg or a filled prescription for an antihypertensive medication. BMI was calculated as weight/height 2 (kg/m 2 ). Blood pressure was measured with an appropriate-sized cuff in the sitting position after a 5-min rest, using an automatic oscillometric monitor.
We evaluated baseline and mean laboratory tests including white blood cell, hemoglobin, uric acid, albumin, calcium, phosphorus, intact parathyroid hormone (IPTH), total cholesterol, and triglyceride. White blood cell and hemoglobin were measured on Beckman Coulter LH 750 (Miami, FL, U.S.A) using the electronic impedance method. Biochemical parameters (creatinine, uric acid, albumin, calcium, phosphorus, cholesterol, and triglycerides) were performed by standard laboratory procedure and measured on an Olympus AU2700 autoanalyzer (Beckman Coulter, Mishima, Japan) using an enzymatic assay. IPTH was assessed by a chemiluminescence assay. The mean laboratory levels were calculated from the time of the initial AVF occurrence to a follow-up period of 1 year. Meanwhile, calcium concentrations (mg/dL) were corrected for albumin concentrations (g/dL) using the following formula: corrected calcium = calcium + 0.8 × (4 − albumin).
All AVFs were created by three experienced surgeons (defined as A, B and C). With regard to our AVF placement strategy, normally our previous guideline was to start distally in the non-dominant arm and then proceed proximally and subsequently moving to the dominant arm for the purpose to preserve future vascular access options. However, the final decision regarding type and location of the initial access creation was individually based on clinical findings and surgeons' opinions.

Evaluation of Aortic Arch Calcification.
Two experienced medical doctors blinded to the patients' clinical data reviewed posterior-anterior plain chest X-rays taken at the time of AVF creation using a specific scoring system [15]. This scoring system, which divided the extent of calcification in the aortic arch into four grades, was as follows: grade 0, no visible calcification; grade 1, small spots of calcification or single thin calcification of the aortic knob; grade 2, one or more areas of thick calcification, but ≤50% of the circular area of the aortic knob; grade 3, circular calcification with >50% of circular area of the aortic knob. The extent of AAC in each chest radiography was evaluated as shown in Figure 1. Grades 0 to 1 and grades 2 to 3 were categorized as lower and higher AAC grade respectively. In all cases of disagreement between the physicians, consensus was eventually reached.

Clinical
Outcome. AVF patency was followed up by physical examination, measurement of dynamic venous pressure, and measurements of access flow by the Transonic machine for 1 year postoperatively. AVF failure was defined as occurrence of thrombosis or requiring endovascular intervention (angioplasty or thrombectomy) to restore blood flow, and repeated AVF failure was defined as patients having at least two episodes of AVF patency loss within 1 year after surgery. As such, the conditions of failing AVF were defined as inadequate AVF blood flow (<600 mL/min), We performed univariate and multivariate logistic regression analyses to determine associated factors of repeated AVF failure for all patients studied. Of this, all potential variables were included in the multivariate model (stepwise method). Test results were presented as odds ratio (OR) with 95% confidence intervals (CI), and two-sided < 0.05 was considered statistically significant.

Comparison of Clinical Characteristics between Lower
Grade AAC and Higher Grade AAC. The study included 284  higher, while IPTH (19.6 pmol/L versus 26.9 pmol/L) levels were significantly lower in the higher grade group. Other biochemical values (white blood cell count, hemoglobin, uric acid, phosphorus, albumin, total cholesterol and triglyceride) did not differ significantly between the two groups. On the other hand, there were also no significant differences in prevalence of sex and hypertension, BMI, AVF characteristics, and the use of medications between the two groups.   significantly associated with repeated AVF failure. Then, we performed multivariate logistic regression analysis to identify the independent associated factors. As shown in Table 3, only the presence of higher grade AAC (2.98 (1.43-6.23); = 0.004), lower mean corrected calcium ( = 0.017) and mean serum albumin level ( = 0.008) were associated with repeated AVF failure. Of note, the presence of comorbid conditions, AVF characteristics and surgeon factor were not associated significantly with repeated failure of AVF.

Summary of Repeated Failure of AVF according to AAC
Grades. Table 4 demonstrates the summary of repeated failure of AVF. A total of 54 patients (19% of study patients) experienced repeated failure of AVF within 1 year, and all subjects were transferred for PTA. The total episodes of thrombosis or intervention in AAC grade 0, grade 1, grade 2 and grade 3 were 12, 15, 58 and 44 respectively. Besides, the repeated AVF failure rate at 1 year in each AAC grades was as follows: grade 0 (10.2%), grade 1 (9.9%), grade 2 (24.5%) and grade 3 (30.8%).

Discussion
The results of the present study established a high prevalence of AAC (79.2%) in hemodialysis patients, and 54.2% of them had higher grade calcification. Among these patients, 19.0% of them had repeated AVF failure within 1-year. Furthermore, we found that preexisting higher grade AAC identified in plain chest radiography, lower mean corrected calcium and mean serum albumin level were all independent risk factors for repeated AVF failure. To our knowledge, this is the first article to evaluate the role of AAC grade on repeated failure of AVF.
In this study, higher grade of AAC was linked to cardiovascular risk factors such as older age, diabetes mellitus and cardiovascular disease. These findings were consistent with results from previous reports [15,16]. For this reason, these high-risk patients had increased rates of cardiac death as well as a higher overall mortality [17]. Otherwise, higher calcium and lower intact-PTH levels probably due to the decreased calcium-buffering capacity of bone were more prone to vascular calcification and thus associated with higher AAC grade [18]. Interestingly, patients with higher AAC grade had lower diastolic blood pressure. It was proposed that as the arteries became stiffer, the increased aortic pulse wave velocity (PWV) result in decreased diastolic blood pressure and high pulse pressure [19].
A previous report demonstrated that higher AAC grade predicted primary patency loss of AVF in ESRD patients [12]. The present study extends the previous finding that higher AAC grade is also a predictor of repeated AVF failure among hemodialysis populations. The reason of AAC-induced AVF failure is unclear, but we propose two possibilities. First, AAC could be a marker of systemic vascular calcification and arteriosclerosis. It has been shown that patients with AAC have linearly increased extent of calcification in the abdominal aorta and a trend toward more frequent arterial calcification of vascular access [20,21]. Of note, AVF failure is significantly more common in patients with preexisting arterial calcification than in those without calcification [7,8,10]. Georgiadis et al. established that macro-calcification in the radial artery identified on a plain arm radiograph was correlated with a worse outcome of radiocephalic AVF in diabetic patients [10]. In another study reported by Choi et al., the result indicated that micro-calcification of artery determined by pathological examination of arterial tissue during AVF surgery was also associated with AVF failure BioMed Research International 7 within 1-year in incident hemodialysis patients [8]. In fact, vascular calcification and arterial stiffness might represent a stage of arteriosclerosis disease. The classification of AAC, in particular, has been shown to provide a significant role in estimating cardiovascular events with higher PWV (a simple way to measure arterial stiffness) corresponding with increasing elevations of risk. Importantly, the development of arteriosclerosis disease is not exclusive to the aortic arch, but rather is a systemic process that develops in multiple vessel beds including artery of dialysis access [17,22]. As to our knowledge, it hinders the maturation process in AVF and subsequently contributes to access failure [23]. Second, AAC is linked to cardiovascular risk factors (such as older age, diabetes, etc.), which are also common predictor factors for AVF failure [3,15,16,24]. As expected, patients with these high risk factors are easily prone to AVF dysfunction [3,24]. Taken together, grading of AAC might indirectly indicate an increased risk of AVF failure and could serve as a noninvasive method for assessing overall AVF outcome in this population.
The issue regarding repeated failure of AVF remains one of the challenges for clinicians involved in vascular access care. Based on the findings of prior investigations, recurrent dysfunction of AVF within 6 months after initial PTA was seen in more than 25% of cases and most cases were recurrent stenosis [25]. Of note, AVF dysfunction is one of the leading causes of morbidity and mortality in hemodialysis patients and is responsible for their high percentage of hospitalizations [5]. Although the risk factors of repeated AVF failure particularly the factor of vascular calcification are rarely investigated, the relationship between vascular calcification and percutaneous intervention relatedrestenosis have been widely studied in cardiovascular diseases other than AVF [26]. In a pooled analysis of the HORIZONS and ACUITY trials that included 6855 subjects who underwent percutaneous coronary intervention (PCI), the presence of moderate/severe coronary calcification was a strong independent predictor of ischemic target lesion revascularization and 1-year definite stent thrombosis [26]. Likewise, stent expansion seemed to be inversely correlated with the arc of the calcium ( = −0.8; < 0.001) [27]. In another recent report with patients undergoing cervical carotid artery stenting, the authors also identified calcified plaque as one of the risk factors to be significantly associated with the incidence of in-stent restenosis [28]. Additionally, among patients who underwent endovascular treatment for femoropopliteal disease, a significantly greater total volume of calcified plaque was found in patients developing restenosis (>50%) compared with those who did not, and they had a significantly worse amputation-free survival rate [29]. On the basis of these findings, severe vascular calcification is indeed a strong predictor of restenosis in the vast majority of cardiovascular diseases following endovascular treatment.
There are several potential mechanisms for vascular calcification induce-repeated failure in AVF. First, impaired vascular dilatation and outward remodeling induced by arteriosclerosis are suspected of being a cause of AVF maturation failure and restenosis. Notably, inadequate arterial and venous dilatation associated with lesion calcium might decrease AVF blood flow, and subsequently lead to restenosis of AVF [30]. Second, a correlation between the blood thrombogenicity and severity of calcification has been reported. A previous investigation established that thrombin level could predict the degree of coronary artery calcification, and suggested that thrombin plays a certain role in the pathophysiology of vascular calcification [31]. Therefore, it is proposed that AVF occlusion may occur under severe vessel calcification owing to thrombus formation. Third, vascular calcification is commonly accompanied by other pathological changes of the artery and vein such as intimal hyperplasia, medial fibrosis and venous calcification, which could synergistically contribute to AVF thrombosis [32,33].
In our study, lower albumin and calcium level were related to the development of repeated AVF failure. It is known that both albumin and calcium as inflammatory factors impact on accelerated atherosclerosis and vascular calcification [34,35]. In addition, chronic inflammation has also been suggested to be implicated in the pathogenesis and progression of intimal hyperplasia and subsequent AVF failure [36]. Of note, these markers were measured based on mean laboratory data instead of baseline level, which represented the mean effect of these factors on AVF outcome. However, to answer the question of causality with certainty, it would be required to evaluate whether restenosis could be attenuated by regimens that increase these serum levels. Hence, it remains elusive whether these molecules are simply a bystander or a contributor to restenosis.
Some limitations should also be considered in the interpretation of our study results. First, this semiquantitative method using four grades to assess AAC is relatively crude and may miss trivial calcium deposition in aortic arch. However, this condition would have no effect on result interpretation since this type of calcification would be categorized as lower AAC grade. Second, most study subjects were followed up mainly by clinical signs suggesting AVF failure; therefore, we may miss the possible presence of silent stenosis in some fistulas. Third, certain potential factors like circulating markers of inflammation, various calcification activators and inhibitors and vascular calcification of AVF were not measured in this study.
In summary, preexisting higher AAC grade, lower mean corrected calcium and mean serum albumin level were associated with repeated AVF failure within 1-year in hemodialysis patients. In fact, routine follow-up by chest radiography could be a simple and cost-effective tool to stratify AVF outcome in these patients. Besides, identification of these risk factors would allow the clinician to implement more rigorous monitoring and planned intervention for some risky new AVFs.

Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this paper.