The cephalic arch is a major site of stenosis in proximal arteriovenous fistulas (AVFs), which leads to access dysfunction and thrombosis [1]. The incidence of cephalic arch stenosis (CAS) in proximal AVF has been reported to be approximately 19%-77% during follow-up [2,3]. CAS occurs more frequently in proximal AVFs, including brachiocephalic AVF (BC-AVF) and proximal radiocephalic AVF (RC-AVF), than in distal RC-AVF [4].

Although the exact cause of CAS is unknown, the unique morphology of the cephalic arch is likely a factor. Unlike distal AVFs, in which blood flow is distributed to many branches, the cephalic arch functions as the sole outflow of proximal AVFs, resulting in a higher hemodynamic inflow volume to the cephalic arch [5]. Anatomically, at the cephalic arch, the angle increases at the point where it joins the axillary vein, and more valves are present than at other locations. It also courses between the biceps brachii and pectoralis muscles in the deltopectoral groove [6]. Therefore, dilation of the cephalic vein (CV) after AVF creation is limited by anatomical constraints. Consequetly, turbulence forms at the cephalic arch, making it more susceptible to stenosis and ultimately causing CAS [7].

Several treatment options exist for managing clinically significant CAS. However, the endovascular approach has largely replaced surgical management in the recent era [8]. Endovascular options include balloon angioplasty using plain, cutting, or drug-coated balloons. Stents or stent grafts are used in refractory stenosis [9]. If the endovascular approach is ineffective, surgical approaches such as flow reduction, CV transposition, or interposition grafting can be considered [10]. The selection of a treatment strategy depends on the experience, available resources, and policies of the institution; however, there is no solid evidence as to which treatment is superior. This study aimed to identify and evaluate the outcomes of endovascular and surgical treatments for CAS in patients with proximal AVF.

1) Study design and data collection

This study was approved by the Institutional Review Boards of Kyungpook National University Hospital and Kyungpook National University Chilgok Hospital (No. 2023-12-019, 2023-12-014). The requirement for informed consent was waived because of the retrospective nature of the study. Between January 2018 and March 2023, 62 patients with proximal AVFs who underwent surgical or endovascular revision for CAS were enrolled and their data were retrospectively reviewed. Given the contemporary trend favoring an endovascular-first approach for CAS and the recurrent nature of CAS, which requires frequent reintervention, percutaneous transluminal balloon angioplasty (PTA) or stent angioplasty was commonly performed during the study period. However, repeated PTA may present different patency rates and could influence the outcomes; therefore, cases of repeated PTA were excluded from the analysis to mitigate bias, and first-time endovascular treatment (ET) for CAS after AVF creation was included. In contrast, surgical treatment (ST) was included regardless of previous ET for CAS because surgical intervention should be considered in cases of frequent and short-term recurrence after ET. ST procedures included CV transposition to the proximal basilic or axillary vein (transposition group) and prosthetic interposition grafting from the distal AVF to the veins in the axilla (interposition group).

Patient characteristics, imaging findings, surgical details, and follow-up results were obtained by reviewing medical charts and analyzing pre- and postoperative images. Patient demographics included age, sex, and the presence of other comorbidities. Lesion characteristics included the presence of thrombotic occlusion, history of ET for CAS, time interval between AVF creation and treatment, AVF type, and laterality.

2) Treatment indications and methods

After fistula creation, we exclusively addressed clinically significant stenoses that exhibited both hemodynamic changes and associated symptoms. We did not perform pre-emptive interventions in the absence of clinical symptoms. The hemodynamic changes considered included stenosis exceeding 50% of the vessel luminal diameter, volume flow less than 600 mL/min, venous pressure exceeding 200 mmHg, and recirculation surpassing 10% [11]. Symptoms included difficulty in needling, challenges in hemostasis at the puncture site, or acute thrombotic occlusion. The decision to pursue ET or ST was contingent on the operator’s clinical judgment, which included factors such as the frequency and intervals of previous ET for CAS, stenosis length, planned transposition segment diameter, and general condition of patients.

In cases of acute thrombotic occlusion, thrombus removal is typically performed via surgical cutdown, milking, or Fogarty thrombectomy. After successful thrombus removal, the decision to proceed with ET or ST for CAS was based on the surgeon’s judgment. ET encompassed PTA and additional stent placement in patients with severe recoil. ST includes CV transposition and interposition. Transposition involves dissection, division, translocation, and anastomosis of the proximal CV to the exposed proximal basilic or axillary vein in the axillary fossa, whereas interposition involves performing an anastomosis with the proximal basilic or axillary vein using a prosthetic graft. This is performed in patients for whom CV transposition is considered unsuitable because of the limited length of the CV resulting from stenosis in the planned transposition segment.

3) Study outcomes and definitions

This study aimed to evaluate the outcomes of the treatment options for CAS; therefore, we used CAS-specific patency. After ET, primary patency was defined as the duration of uninterrupted patency of the cephalic arch without endovascular reintervention or surgical revision for CAS. After ST, primary patency was defined as the duration of uninterrupted patency of the transposed CV and the interposition graft. Functional patency was defined as the duration of AVF patency with function after a successful endovascular or surgical revisions.

The primary outcome of interest was the CAS-specific primary patency rate after either ET or ST, and the secondary outcomes included functional patency and the number of patency loss events during follow-up. A patency loss event was defined as any endovascular or surgical revision of the treated segment, or loss of functional patency.

4) Statistical methods

For the comparison of the ET and ST groups, Student t-test or the Mann-Whitney U-test was performed for continuous variables after the normality test, whereas the chi-square test (for adequately sized samples) or Fisher exact test (for smaller samples) was performed for categorical variables. Kaplan-Meier plots were used to assess primary and functional patency, whereas the log-rank test was used to determine the statistical significance of the differences between the survival curves. All statistical results were analyzed using IBM SPSS statistics (ver 25.0; IBM), and significance was set at P-values of <0.05.

1) Characteristics of patients and fistulas

During the study period, ST was performed for 26 AVFs, including transposition in 16 and interposition grafts in 10. ET was performed for 36 AVFs involving balloon angioplasty in all patients; however, 3 patients required additional stent placement because of immediate recoil. Among the 26 patients who underwent ST, 15 underwent the procedure for the first time after AVF creation and 11 underwent ST after failed ET during follow-up (Fig. 1).

Figure 1. Flowchart showing the selection of the study population.

The baseline clinical and fistula characteristics are summarized in Table 1. When comparing the ET and ST groups, diabetes as an underlying disease was more prevalent in the ET group than in the ST group (80% vs. 54%; P=0.024). In contrast, the rates of preoperative AVF thrombosis and previous cephalic arch ET were higher in the ST group (39% vs. 19%; P=0.098 for preoperative AVF thrombosis, 42% vs. 0%; P<0.001 for previous cephalic arch ET). The location of the AVF and interval between AVF creation and CAS treatment did not differ between the groups.

Table 1 . Baseline characteristics.

	ET (n=36)	ST (n=26)	P-value
Age (y), mean±SD	65.5±12.9	66.9±14.4	0.699
Male	23 (64)	15 (58)	0.621
Diabetes mellitus	29 (80)	14 (54)	0.024
Preoperative AVF thrombosis	7 (19)	10 (39)	0.098
Previous cephalic arch ET	0	11 (42)	0.000
Fistula age, median (IQR) (mo)	30.1 (12.9-49.4)	31.2 (17.7-66.3)	0.429
Location of fistula			0.300
Brachiocephalic AVF	32 (89)	25 (96)
Proximal radiocephalic AVF	4 (11)	1 (4)
Laterality of fistula			0.124
Left arm	29 (80)	25 (96)

Values are presented as number (%), mean (IQR) or mean±SD deviation..

SD, standard deviation; AVF, arteriovenous fistula; IQR, interquartile range; ET, endovascular treatment; ST, surgical treatment..

2) Primary patency

The overall primary patency rates after CAS treatment were 67.9% at 6 months, 53.9% at 1 year, and 24.5% at 3 years.

When comparing ET and ST, the primary patency rates at 6 months, 1 year, and 3 years were 70.6%, 60.5%, and 41.9%, respectively, in the ST group and 65.7%, 48.9%, and 16.8%, respectively, in the ET group. There was no significant difference in the primary patency rate (P=0.264) (Fig. 2). However, in a subgroup analysis that divided the ST group into transposition and interposition groups, the transposition group exhibited significantly higher primary patency rates than the interposition and ET groups (P=0.001 for transposition vs. interposition; P=0.016 for transposition vs. ET) (Fig. 3). The primary patency rates at 6 months, 1 year, and 3 years were 86.7%, 86.7%, and 66.0%, respectively, in the transposition group, and 45.0%, 22.5%, and 11.3%, respectively, in the interposition group. There was no significant difference in the primary patency rate between the interposition and ET groups (P=0.190).

Figure 2. Primary patency data shown according to the treatment methods.

Figure 3. Primary patency data shown after subgroup analysis, in which surgical treatment was divided into transposition and interposition groups.

3) Functional patency and number of patency loss events

During the follow-up period, seven AVFs were abandoned, resulting in the loss of functional patency for various reasons. Three patients required new native or prosthetic AVFs following thrombotic occlusion, whereas three other patients continued hemodialysis with a catheter because of their poor general condition at the time of AVF occlusion. Additionally, one patient underwent AVF ligation and contralateral AVF creation because of central vein occlusion and severe arm swelling.

Overall, the functional patency rates after CAS treatment were 100% at 6 months, 97.2% at 1 year, 92.4% at 3 years, and 79.2% at 5 years. There was no significant difference in functional patency between the ET and ST groups (P=0.461) (Fig. 4A). Additionally, when the ST group was divided into transposition and interposition groups and compared with the ET group, there were no significant differences in functional patency among the 3 groups (P=0.108 for transposition vs. interposition; P=0.731 for transposition vs. ET; and P=0.141 for interposition vs. ET) (Fig. 4B).

Figure 4. Functional patency data shown according to the treatment methods. (A) Surgical treatment versus endovascular treatment. (B) After subgroup analysis, in surgical treatment was divided into transposition and interposition groups.

After CAS treatment, 27 recurrent stenoses and 30 occlusions requiring surgical or endovascular interventions, along with 7 abandonments, occurred in 34 patients during follow-up (Table 2). In the transposition, interposition, and ET groups, the numbers of patency loss events per person-year were 0.40, 0.52, and 1.42, respectively.

Table 2 . Outcomes of surgical and endovascular treatment for cephalic arch stenosis.

	Transposition	Interposition	ET
No. of patients	16	10	36
Follow-up duration, mean (mo)	47.4	32.5	19.7
Follow-up duration, overall (mo)	758.0	325.1	708.5
No. of stenoses	5	6	16
No. of occlusions	11	8	11
No. of abandonments	3	3	1
No. of overall events	19	17	28
No. of events/person-year	0.40	0.52	1.42

ET, endovascular treatment..

In patients undergoing dialysis with an AVF, long-term patency has a significant impact on quality of life and economic aspects. Proximal AVFs, including BC-AVF and proximal RC-AVF, offer advantages such as higher maturation rates and lower reintervention rates compared to distal AVFs [12]. They also provide cosmetic benefits and reduce pain during dialysis. However, the substantial flow volume after AVF creation and unique anatomical characteristics of the cephalic arch contribute to frequent stenosis, necessitating numerous interventions.

This retrospective study compared ST and ET outcomes in patients with CAS. The surgical approach was further analyzed by categorizing the patients into transposition and interposition groups. Notably, the incidence of AVF thrombosis at presentation was higher in the ST group (44% [7/16]) in the transposition and 30% (3/10) in the interposition group) than in the ET group (19%, 7/36). Although there was no significant difference in the primary patency rate between the ST and ET groups, a notable distinction emerged when analyzing the surgical methods separately, with transposition demonstrating superior primary patency compared with interposition or ET. Moreover, while no significant differences were observed in functional patency among the three groups, the transposition group exhibited the lowest number of events affecting patency. Therefore, based on our findings, we propose that CV transposition should be actively considered, particularly in patients requiring long-term durability, such as young individuals or those experiencing frequent restenosis after ET. In cases where transposition is not feasible owing to insufficient CV length, an interposition or endovascular approach can be considered based on hospital facilities and operator experience.

Similar to other AVF stenoses, CAS treatments can be categorized into surgical and endovascular approaches. However, the Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines, which are widely used for vascular access management, lack clear recommendations for CAS because of insufficient related research [13]. Several retrospective studies have indicated that CV transposition has superior outcomes compared to endovascular approaches [14-16]. For instance, the primary patency rates of AVFs after CV transposition at 1 year range from 66% to 85%, in contrast to 9.5%-50% after PTA [1,14,15,17,18]. Additionally, the primary patency of PTA in CAS is inferior to that of other stenoses [19], and repeat PTA in CAS is linked to a faster recurrence rate [20]. Despite these findings, a systematic review failed to identify a superior approach between open and endovascular options because the studies were based primarily on single-center retrospective trials with heterogeneous patient populations, interventions, and endpoints [21]. In general, the KDOQI guidelines recommend ST for AVF salvage when the problem recurs more than twice in 3 months despite PTA. Applying the “PTA more than twice in 3 months” criterion to CAS is difficult because of the scarcity of randomized clinical trials. However, transposition could be a proactive option considering the lower patency of CAS after PTA than other stenoses and the higher recurrence rate after repeated PTA. Furthermore, it is imperative to ascertain risk factors associated with short-term recurrent stenosis after PTA for CAS.

The outcomes of simple PTA for CAS are known to be unsatisfactory because of the frequent occurrence of rupture and recoil during PTA as well as poor long-term patency during follow-up [19]. Consequently, bare-metal stents or stent grafts have been used to overcome these challenges. In a recent systematic review of ET for CAS [18], bare-metal stents exhibited superior primary patency at 6 months compared with PTA; however, the primary patency at 1 year was similar in both groups. Notably, the stent graft cohort demonstrated superior primary patency at both 6 months and 1 year compared with the bare-metal stent and PTA groups. The reported pooled primary patency rates at 1 year were 44.0%, 12.9%, and 9.5% in the stent graft, bare-metal, and PTA cohorts, respectively [18]. Therefore, with the advancement of endovascular devices, stent grafts may offer significant benefits in CAS compared with bare-metal stents or PTA, with a higher primary patency rate. However, precise deployment of stents or stent grafts for CAS is often challenging in cases with stenosis at the cephalic to axillary vein junction, and improper deployment with protrusion into the axillary vein may compromise the patency of the axillary vein and preclude further revision options, including transposition or interposition grafting [22]. Additionally, considering the excellent patency of CV transposition in our series (86.7% at 1 year and 66.0% at 3 years), the placement of an excessively long stent in the upper CV should be avoided because it can limit vein availability and hinder future CV transposition for recurrent and frequent stenosis after intervention.

Furthermore, the patency of interposition grafts was poor in our series. In a study comparing ST with ET, both transposition and interposition showed higher patency rates than ET, including PTA and stent placement [15]. In addition, transposition and interposition did not result in any differences in patency [15]. However, in our study, the outcomes of interposition using prosthetic grafts were inferior to those of transposition and were similar to those of ET. Several factors could explain this discrepancy. First, this may be related to the inherently low patency associated with prosthetic graft use. Vascular access creation using a prosthetic graft has been linked to lower patency than native AVFs, particularly due to a compliance mismatch, resulting in intimal hyperplasia [13]. Second, although the exact diameters of the CV, proximal basilic vein, and axillary vein were not determined in this study, interposition grafts are usually performed in cases deemed small in the planned CV transposition segment. Therefore, the diameter of the CV in the proximal anastomosis of the interposition graft has the possibility of being smaller compared with that in transposition. The status of the inflow and outflow vessels, including CV diameter, should be assessed to accurately compare the outcomes of CV transposition and interposition with a prosthetic graft.

This study had several limitations. First, the study had a single-center retrospective design with a small number of patients; therefore, the statistical power may be low, and bias may be present owing to the reliance on the selection of treatment methods by attending physicians. Although the primary patency and number of reinterventions during follow-up were better in the transposition group than in the ET group, functional patency was similar. This finding may raise questions about the exact benefits of transposition as overall survival of AVF is more important than the primary patency for patients requiring long-term dialysis. However, cost effectiveness is a critical consideration in the management of dialysis access. Although we did not conduct a cost analysis for repeated procedures and hospitalizations after the initial ET or ST, it may be more favorable in the transposition group than in the ET group considering the longer primary patency and lower reintervention rates. Therefore, further studies of the cost of each treatment during the follow-up period are necessary. Additionally, it would be beneficial to develop and incorporate the results of patient-reported outcome measures into future studies. Second, the diameter or hemodynamic values of the CV and axillary vein where the surgery was planned were not evaluated, potentially affecting outcomes such as patency. Third, the number of patency loss events per person-year after CAS treatment was lower than that reported in other studies, particularly for PTA. The number of reinterventions per person-year after PTA for CAS has been reported as 2.5 to 3.5 [15,16]; however, in our series, it was 1.42 in the ET group. This could be because the primary patency in this study used the CAS-specific patency for analysis, excluding reinterventions performed in other segments, such as the juxta-anastomotic segment. Additionally, the policies of our center may have influenced this result, as the intervention for CAS was performed exclusively in cases with clinically relevant hemodynamic changes and associated symptoms, rather than pre-emptively revising for subtle hemodynamic changes without associated symptoms. These factors may have contributed to the relatively low rate of events observed after CAS treatment in this study.

Currently, the KDOQI guidelines do not favor one strategy over other strategies for managing CAS, and a systematic review failed to identify a superior approach among open and endovascular options. However, this study suggests that CV transposition should be actively considered in suitable patients with CAS due to the higher primary patency and fewer events affecting patency during follow-up. Additional large-scale prospective studies with long-term follow-up are necessary to validate these findings.

None.

Hyung-Kee Kim has been the editor-in-chief of the VSI since 2023. Woo-Sung Yun has been the senior editor of the VSI since 2023. They were not involved in the review process. Otherwise, no potential conflict of interest relevant to this article was reported.

Conception and design: JHJ, SH, HKK. Analysis and interpretation: YRP, JHJ, DH, WSY. HKK. Data collection: YRP, JHJ, DH. Writing the article: YRP, JHJ, HKK. Critical revision of the article: WSY, SH, HKK. Final approval of the article: all authors. Statistical analysis: DH, HKK. Obtained funding: none. Overall responsibility: HKK.