Endovascular Management of Central Thoracic Veno-Occlusive Diseases in Hemodialysis Patients: A Single Institutional Experience in 69 Consecutive Patients

Materials and Methods 

Results 

Among 600 patients presenting with threatened dialysis accesses, 69 (11%) had evidence of central veno-occlusive disease. In the evaluation of initial angiographic data, central veno-occlusive disease was detected in 92 segments. The distribution of central veno-occlusive disease and the degree of stenosis are summarized in Table 1.

Table 1. Location of Lesions and Severity of Stenosis

	Location	No. of Lesions (n = 92)	Percentage of Stenosis
			Mean ± Standard Deviation	Range
	Subclavian vein	31	78±11	30–100
	Brachiocephalic vein	52	74±8	20–100
	Superior vena cava	9	86±14	50–100

Observer 1 detected 90 segmental disease including 18 segments with less than 30% stenosis, 60 segments with 30%–99% stenosis, and 12 segmental occlusions. Observer 2 detected 92 segmental disease including 16 segments with less than 30% stenosis, 64 segments with 30%–99% stenosis, and 12 segmental occlusions. There was no significant difference in the grading scores between the two readers (P = .8). The κ coefficient revealed an excellent overall interobserver agreement (κ = 0.84; 95% confidence interval: 0.67, 0.93) for all degrees of venous stenosis (grades 1–3).

The overall number of interventions per treated central venous segment averaged 2.5 (range, 1–8). Initial angioplasty alone was unsuccessful in treating 10 central venous segments (in seven patients), resulting in an initial technical success rate of 90% for PTA. Four segments (two subclavian occlusions, two brachiocephalic occlusions) could not be cannulated and crossed by the guide wire. In the other six segments (two subclavian occlusions, three brachiocephalic occlusions, and one superior vena cava occlusion), initial PTA was attempted and failed to restore sufficient venous flow (residual stenosis >30%), necessitating stent placement. All stent placements successfully achieved venous patency. Repeat interventions to maintain the patency of central venous segments included 120 PTAs and 14 secondary stent placements. The success rate of follow-up interventional procedures was 96% (129 of 134 procedures).

The deployed self-expandable stents (n = 20) in this study are summarized in Table 2. As we have described earlier, six central venous segments were treated with a stent immediately after initial PTA failed to restore sufficient venous flow (residual stenosis, >30%). Nine other venous segments were treated with stents (n = 14) during the follow-up due to restenosis or failed repeated PTAs. Among these, three segments received two stents and one segment received three stents. The 20 self-expandable stents were deployed in the subclavian vein (n = 6), brachiocephalic vein (n = 12), and superior vena cava (n = 2).

Table 2. Summary of 20 Self-expandable Stents in 15 Venous Segments

	Type of Stent and Segment	No. of Stents
	Wallstent
	Subclavian (left)	1
	Subclavian (right)	2
	Superior vena cava	1
	Brachiocephalic (right)	1
	Brachiocephalic (right)	1
	Brachiocephalic (right)	1
	Smart stent
	Subclavian (left)	1
	Subclavian (right)	2
	Superior vena cava	1
	Brachiocephalic (right)	1
	Brachiocephalic (right)	3
	Brachiocephalic (right)	2
	Brachiocephalic (left)	1
	Brachiocephalic (right)	1
	Brachiocephalic (left)	1

Note.—Seven Wallstents were placed in six venous segments, and 13 Smart stents were placed in nine venous segments. Six stents were placed in the subclavian vein, 12 in the brachiocephalic vein, and two in the superior vena cava.

On average, the initial procedure provided 6.8 months of further intervention-free access. Assisted procedures added an average of an additional 7.2 months to the functional status of the hemodialysis access. When evaluated with Kaplan-Meier analysis, dialysis access survival and patency rate was significantly increased by means of assisted procedures (P = .001, log-rank test). Figure 1 shows the overall performance and comparison of primary and assisted patency of hemodialysis access during follow-up.

View Large Image Download to PowerPoint

Figure 1. Graph shows the overall performance and comparison of primary and assisted patency rates of hemodialysis access during follow-up. When evaluated with Kaplan-Meier analysis, dialysis access survival and patency rates were significantly increased with use of assisted procedures (P = .001, log-rank test).

The mean follow-up was 14.5 months (range, 1–44 months) from the date of initial intervention. Forty-six of the 69 patients (67%) underwent at least 12 months of follow-up. When the patency rate and survival of hemodialysis access was further broken down to smaller intervals, primary patency rates of the hemodialysis access at 1-, 3-, 6-, 9-, and 12-month intervals were 82%, 65%, 49%, 38%, and 29%, respectively. The assisted patency rates were significantly (P =.0014) improved at all of these time points when compared to primary patency rates using Kaplan-Meier analysis and the modified peto test (Fig 2).

View Large Image Download to PowerPoint

Figure 2. Graph shows primary versus assisted patency and survival of hemodialysis accesses at 1-, 3-, 6-, 9-, and 12-month intervals. The assisted patency rates were significantly improved at all of these time points, as determined with Kaplan-Meier analysis and the modified peto test (χ2 value = 10.54 with 1 df; P = .001).

There was no significant difference between the behavior and/or patency of the venous segments treated with PTA alone and that of venous segments treated with additional stent placement (mean patency period, 12.5 vs 15 months; P = .1). However, this should be interpreted with caution because the sample size for the stent population (20 stents for 15 segments) is small. When we compared the Wallstent with the Smart stent, no significant difference was noted between the patency rates (P = .3), again considering the small sample size.

When hypertension, diabetes mellitus, and a history of smoking coexisted, the restenosis of the central veins was significantly higher (P < .001), requiring a significantly higher number of interventions (4.8 ± 2.1) when compared to a lower-risk population with only one or two of these risk factors. There was no significant difference in the laterality (right or left predominance) of the central venous segments involved or treated.

There were no procedure-related deaths and no periprocedural mortalities. Complications were, for the most part, minor, as we observed only two complications that required further intervention. One patient with right axillary and subclavian veno-occlusive disease developed a pseudoaneurysm subsequent to PTA that was successfully treated with a covered stent (Fig 3). In one patient with complete occlusion of the right subclavian and brachiocephalic veins, the angioplasty balloon (8 mm) ruptured at 20 atm. Several attempts to retrieve the ruptured balloon failed because the ruptured balloon catheter could not be pulled out through the previously inserted Smart stent in the right axillary vein. This patient was taken to surgery, and the ruptured balloon was retrieved with a direct cutdown of the axillary vein.

View Large Image Download to PowerPoint

Figure 3. Images from digital subtraction angiography in a 58-year-old man with end-stage renal disease and a right brachiocephalic fistula who presented with right upper extremity pain and swelling. Images reveal severe veno-occlusive disease in the junction of the right axillary and subclavian vein (arrow in a) with central filling via collateral vessels. The arrowhead in b shows the development of a pseudoaneurysm subsequent to PTA, which was treated with a covered stent (c).

Discussion 

Although its pathophysiology is not entirely clear, involvement of the central veins is usually associated with venous stress or trauma from previous central vein cannulation as well as high arteriovenous flows within the access circuit itself (1, 2). The growing population of patients with end-stage renal disease and their increased survival has substantially increased the burden associated with dialysis access–related complications.

Occlusive disease is common after repeated cannulation of the brachiocephalic veins. These stenoses can increase outflow resistance, leading to extremity edema and poor-quality dialysis. This often results in the abandonment of the fistula site and precludes placement of further access in the affected extremity. The site of previous vein cannulation is an important determinant of central venous occlusion. The placement of catheters in the subclavian veins have led to the development of stenoses in up to 50% of patients (3, 4), whereas catheterization of the right internal jugular vein is associated with the lowest frequency of central vein stenosis (5). We have adopted the policy of avoiding subclavian vein cannulation in dialysis patients and, as a result, have seen a relatively low incidence of central venous occlusion (11%) in our dialysis patients.

The management of central veno-occlusive disease is evolving. Open surgical treatment has yielded acceptable access preservation but is often associated with substantial morbidity in dialysis patients, who often have multiple medical problems (15). Beginning in the 1980s, endovascular methods for the treatment of central venous occlusive disease were being developed (2, 16). At present, treatment options include high-pressure balloon angioplasty, intravascular stent placement, and, most recently, cutting balloon PTA. However, the optimal treatment strategy remains unclear. Some have advocated primary stent placement for the treatment for central venous obstruction (9, 17), whereas others have advocated PTA, reserving selective stent placement for failed angioplasty or recurrent stenosis (10, 12, 18). The literature does not clearly support one treatment strategy over the other.

Since 1988, multiple reports have described the successful use of stents to primarily treat central veno-occlusive disease in dialysis patients. Oderich et al (17) presented 40 central venous obstructions that were treated with 50 stents. During a mean follow-up of 16 months, they reported a 1-year primary patency rate of 27% and a 1-year assisted patency rate of 71%. Another study, by Haage et al (9), evaluated 50 patients who underwent stent placement as the primary treatment for central venous obstruction. They reported 1-year primary and assisted patency rates of 56% and 97%, respectively. However, there are several studies that reported no additional advantages from primary stent placement in the treatment of patients with central veno-occlusive disease. In a prospective randomized trial, Quinn et al (12) showed no improvement in patency rates with stent placement. In that study, 1-year primary and assisted patency rates in patients undergoing PTA alone were 12% and 100%, respectively, and that in patients undergoing PTA with stent placement were 12% and 78%, respectively. Recently, Bakken et al (11) reported 1-year primary and assisted patency rates of 29% and 73%, respectively, in patients undergoing PTA alone and 21% and 46% in patients undergoing primary stent placement, promoting the conclusion that primary stent placement does not benefit long-term outcomes in the treatment of central veno-occlusive disease in hemodialysis patients. Literature discouraged the use of a subclavian stent due to its associated complications (eg, stent fracture and a relatively higher rate of restenosis). In this study, we only used subclavian stent placement for patients in whom adequate outflow could not be achieved with PTA alone despite several attempts with high-pressure balloon. Although we did not observe stent fracture, the required second stent placement in two subclavian segments and the relatively higher number of intrastent interventions (PTA) likely supports the literature.

Given the additional cost of stents and the potential negative consequences of stent placement—which include collateral vein obstruction, stent migration, and infection—one should only adopt primary stent placement as the preferred treatment option if a clear-cut benefit over PTA can be established. We have chosen to adopt the strategy of selective stent placement and have achieved results comparable to those seen in studies where primary stent placement has been employed.

Although recent advances in catheter technology and interventional procedures have prolonged the assisted patency rate, a review of the literature shows no significant improvement in assisted patency rates during the past 10 years, with most of the reports demonstrating a 12-month assisted patency rate of about 60%–70%. It is unclear how much of this stagnation can be further addressed by adjusting and/or fine-tuning the interventional procedures. This, in our opinion, demonstrates the complexity of this entity and the much-needed adjustment and/or improvement in diagnostic and therapeutic approach, which may require more clinical and basic research to bolster the pathophysiology of this entity in the growing population of patients with hemodialysis access.

Pre-existing comorbidities can influence the durability of angioplasty of the brachiocephalic veins. In a previous study (11), it was determined that congestive heart failure was associated with treatment failure. In our series, patients who had diabetes, hypertension, and a history of smoking were significantly more likely to experience restenosis after PTA than were patients who had only one or two of these risk factors. It remains to be determined whether optimal management of these factors will lead to improved durability of percutaneous interventions on the central veins.

Because of the limited sites available for access placement, it is crucial to preserve the functional patency of an existing access. By using percutaneous techniques, we were able to significantly prolong the life of existing access sites. This is reflected in the fact that 55% of our treated patients were able to remain free from placement of percutaneous dialysis catheters for at least 12 months.

Our study has several limitations. First, this is a nonrandomized retrospective study and a prospective randomized trial comparing PTA with stent placement may enhance and define the role of endovascular management in the treatment of central veno-occlusive disease. In addition, the results may have been skewed due to the retrospective nature of the study and potential recall bias. With prospective review of the images and measurement of the degree of stenosis by two independent observers, we attempted to minimize the recall bias, as shown by a high degree of interobserver agreement, which reflects the reproducibility of the study results. Finally, this study period does not include the more recent use of cutting balloon angioplasty within central veins, which may potentially enhance the current outcomes of threatened hemodialysis access in patients with central veno-occlusive disease.

We conclude that endovascular treatment of central veno-occlusive disease with angioplasty and selective stent placement is safe, with low technical failure rates and complications. Surveillance and multiple additional interventions are justified by low morbidity and complication rates, providing additional meaningful longevity for hemodialysis access sites. Further prospective studies in a broader clinical setting are required to establish the proper management of central veno-occlusive disease in patients with threatened dialysis access.