The effect of a venous anastomosis Tyrell vein collar on the primary patency of arteriovenous grafts in patients undergoing hemodialysis☆☆☆
Article Outline
Abstract
Purpose: Vein collars and patches are used at the distal anastomoses of infrainguinal prosthetic grafts to improve graft patency. We initiated a randomized, prospective study to determine whether a Tyrell vein collar at the venous anastomosis of forearm loop arteriovenous grafts (AVGs) would improve patency. Methods: Patients who required new forearm AVGs were randomized to (1) a standard end-to-side graft-vein anastomosis (control group) or (2) a Tyrell vein collar between the graft and the vein (study group). End points were (1) graft thrombosis, (2) graft removal and ligation, or (3) inadequate graft function. Randomization of 75 subjects was planned. The study was terminated early for ethical reasons. Results: Seventeen patients (eight men, nine women) with a mean age of 52.8 years (range, 31-79 years) had 17 grafts placed (control group, n = 10; study group, n = 7). Comorbidities were not different between the groups (P > .05). Six (86%) of seven study grafts failed by 9 months (mean, 4.6 months). Four (66%) failed study grafts had venous outflow tract stenosis from intimal hyperplasia. This was confirmed at surgery in three and by angiography in one. The 9-month primary patency was 80% for the control group versus 17% for the study group (P = .015). Smaller outflow vein diameter in the study group (P = .048) did not account for this inferior graft patency. Conclusion: A Tyrell vein collar at the venous anastomosis of a forearm AVG resulted in premature graft failure. The use of a Tyrell vein collar may accelerate venous anastomosis intimal hyperplasia. (J Vasc Surg 2000;32:1149-54.)
Please see related commentary by Dr Julie A. Freischlag on pages 1235-6.
Patients undergoing hemodialysis for end-stage renal disease (ESRD) require constant attention to maintain vascular access. Failed access is the most frequent cause of hospitalization for these patients, and it necessitates expensive interventions to restore function.1 Although mature, autologous arteriovenous fistulas (AVFs) have the best patency rates and the least complications of the various access techniques, many patients are not candidates for an AVF.1 Polytetrafluoroethylene (PTFE) arteriovenous grafts (AVGs) are currently constructed almost twice as often as autologous fistulas1 despite a discouraging mean patency of only 13 months.2 Any means to improve AVG patency and decrease the frequency of revisions would improve the quality of life of patients with ESRD and decrease health care expenditures.
Venous anastomosis intimal hyperplasia is the most common cause of late AVG thrombosis1 and a common cause of PTFE graft failure in infrainguinal reconstructions. Compliance mismatch between a rigid PTFE graft and a pliable outflow vessel has been blamed for this hyperplasia. Vein collars and patches have been used to decrease compliance mismatch at the distal anastomoses of lower extremity PTFE bypass grafts with improved patency reported.3, 4 The goal of this study was to determine if a Tyrell vein collar could improve primary AVG patency by decreasing compliance mismatch and venous outflow tract intimal hyperplasia (Figure).
Diagram of a Tyrell vein collar with and without a PTFE graft.
Methods
From August 1996 to March 1998 (20 months), all patients referred for hemodialysis access were considered for this study. The study was approved by the Institutional Review Boards of the participating hospitals. Patients were study candidates if they needed new PTFE forearm loop grafts and gave informed consent. Patients requiring AVG revision or an AVG in the upper arm or leg were excluded. All data were recorded on a standard reporting form. The preoperative history and physical examination of patients focused on events and findings that might predispose to early graft failure because of arterial insufficiency or prior venous occlusive disease.
The patients were prospectively randomized to have a conventional end-to-side venous anastomosis (control group) or an end-to-side anastomosis with a Tyrell vein collar between the graft and the outflow vein (study group). The vein used to create the Tyrell collar in the study group was obtained from either normal-appearing ipsilateral superficial forearm veins (n = 5), or the greater saphenous vein or its branches (n = 2). Three surgeons created 10 control AVGs (A, n = 6; B, n = 3; C, n = l), and two surgeons created the seven study AVGs (A, n = 6; D, n = l). Details of the operative procedure were recorded (Table). Outflow vein diameter was measured with a ruler after it was distended by the AVG. Patients did not use their AVGs for 2 weeks after surgery to allow for incorporation.
Comparison of perioperative comorbidities and technical data
| Control group N(%) | Study group N(%) | P value | |
|---|---|---|---|
| Preoperative data | |||
 Diabetes mellitus | 3 (30) | 2 (29) | NS |
| Hypertension | 10 (100) | 7 (100) | NS |
| CAD/CHF | 2 (20) | 3 (43) | NS |
| Smoking | 3 (30) | 0 (0) | NS |
| Prior ipsilateral AVF | 2 (20) | 2 (29) | NS |
| Prior ipsilateral CIV | 1 (10) | 2 (29) | NS |
| Operative data | |||
| Donor artery | |||
| Brachial | 5 (50) | 7 (100) | NS |
| Radial | 3 (30) | 0 (0) | |
| Ulnar | 2 (20) | 0 (0) | |
| Outflow vein | |||
| Cephalic | 2 (20) | 1 (14) | NS |
| Basilic | 3 (30) | 3 (43) | |
| Median antecubital | 3 (30) | 1 (14) | |
| Brachial | 2 (20) | 2 (29) | |
| Vessel control | |||
| Clamp/vessel loop | 7 (70) | 6 (86) | NS |
| Tourniquet | 3 (30) | 1 (14) | |
| Graft across AC fossa | 1 (10) | 2 (29) | NS |
| Mean (± SD) | Mean (± SD) | ||
| Graft length (cm) | 30.7 ± 3.5 | 33.4 ± 3.8 | NS |
| Arteriotomy length (mm) | 11.0 ± 2.7 | 13.3 ± 1.5 | NS |
| Venotomy length (mm) | 17.8 ± 5.8 | 15.4 ± 6.5 | NS |
| Outflow vein diameter (mm) | 5.2 ± 0.6 | 4.4 ± 0.9 | .048 |
| Operative time (min) | 173 ± 56 | 213 ± 65 | NS |
Postoperatively, the patients were examined for complications and graft patency by the surgical staff at regular intervals. Graft failure was defined as (1) graft thrombosis, (2) graft removal and ligation, or (3) inadequate graft function that required revision to maintain patency and effective dialysis. Determination of inadequate graft function was made by the staff of the dialysis center. Graft thrombosis was confirmed by the attending surgeon.
We planned on randomizing 75 patients for this study to have a statistical power of 0.89 and a β value of .11. When the study was terminated for ethical consideration, the number of subjects in the study and control groups was not equal because they were not block randomized. Statistical analysis was performed with the Student t test for continuous variables and the Fisher exact test for nominal variables.
Results
Seventeen patients consented to participate in this study. Subsequently, the recruitment of patients was stopped for ethical concerns. Early graft failure due to venous outflow obstruction was extremely high in the first seven patients entered into the study group. For this reason, we considered further randomization in the study group to be inappropriate.
Nine women and eight men with a mean age of 52.8 years (median, 54 years; range, 31-79 years) were randomized (control group, n = 10; study group, n = 7). Patient demographics are compared in the Table. Comorbidities were similar between the two groups. No patients had abnormal or unequal brachial blood pressures that required evaluation for arterial insufficiency. Three patients (one from the control group, two from the study group) had prior ipsilateral central venous lines. Central vein stenosis was not evident by means of preoperative venography or duplex scan examination in the two patients studied. Six-millimeter regular PTFE grafts were used in 16 (94%) of the 17 patients in the study. One patient in the control group had a thin-walled, 4- to 7-mm tapered PTFE graft (primary patency, 14 months).
Comparison of the operative data for the control and study groups identified outflow vein diameter and total operative time as the only significantly different variables (Table). The increased operative time in the study group reflected the extra time to form a Tyrell vein collar. The outflow vein diameters were statistically smaller (P = .048) in the study group (4.4 mm) than in the control group (5.2 mm). This difference was largely due to a patient in the study group with a 3-mm outflow vein. The size of the outflow vein ranged from 4 to 6 mm for all other patients in the study and control groups. There was no statistical difference between the two groups if this one 3-mm vein in the study group was excluded from the analysis (mean diameter: control group, 5.2 mm; study group, 4.58 mm; P = .12). This patient with a 3-mm outflow vein had a well-functioning graft for 5 months before intimal hyperplasia in the venous outflow tract led to graft thrombosis.
Six (86%) of the seven patients in the study group had graft failure within 9 months of surgery. A single study group AVG remains patent at 12 months, with no evidence revealed by duplex ultrasound scan of venous outflow stenosis. At 9 months, only 20% of the grafts in the control group had failed. The mean primary patency for the control group was 13.4 months (range, 1-25 months) compared with 4.6 months (range, 1-12 months) for the study group. Primary patency at 9 months for the grafts with a Tyrell vein collar (study group, n = 7) was significantly inferior to that of the control group (n = 10) (P = .015). The two AVGs with vein collars from the greater saphenous vein or a large saphenous branch vein failed at 1 and 2 months. Although the small number (n = 7) of patients with vein collars in the study group limited any statistical comparison, AVGs with arm vein collars appeared to remain patent longer (mean, 5.6 months; range, 1-12 months). If the one graft with the 3-mm outflow vein was excluded from the statistical analysis, the primary patency of the remaining study group AVGs (n = 6) remained inferior (P = .035) to the control group AVGs (n = 10).
A month after surgery, one patient in the study group had graft failure subsequent to a prolonged period of postdialysis hypotension. Her graft was salvaged with thrombolysis, and angiography revealed no stenosis in the graft or outflow tract. This graft has functioned without further problems for a total of 9 months (secondary patency). A second patient in the study group had a steal syndrome 2 weeks after her AVG was placed. Because of worsening hand ischemia, the graft had to be ligated 1 month after it was placed. Physical examination, segmental pressures, and pulse volume recordings revealed no proximal arterial occlusive disease.
We evaluated four thrombosed study group AVGs with post-thrombolysis angiography (n = l) or surgical exploration (n = 3). These grafts, which failed 1.25, 2, 5, and 9 months after surgery, had findings consistent with intimal hyperplasia at the venous anastomosis. Angiography showed a smooth tapered stenosis in the toe of the venous anastomosis involving the first 1 to 2 cm of the venous outflow tract. Surgery identified a smooth, shiny, stenotic intraluminal lesion at the toe of the venous anastomosis and on the back wall of the outflow vein. Of the six AVGs with Tyrell vein collars that failed, four (66%) had venous stenoses from intimal hyperplasia, one was ligated for arterial steal, and one failed from hypotension without a stenosis. Overall, four (57%) of seven patients in the study group had graft thrombosis within 9 months of surgery because of anastomotic intimal hyperplasia.
Discussion
Almost a quarter of a million Americans undergo dialysis each year for ESRD.1 Most require long-term hemodialysis, and one third of these patients receive inadequate dialysis, which may decrease their survival.5 Failed vascular access contributes to the problem of inadequate hemodialysis,5 and it also accounts for the greatest number of days spent in the hospital by these patients.1 Reliable, well-functioning vascular access is essential for improving the quality and length of life for patients with ESRD.
Autologous AVFs are the vascular access of choice for patients undergoing hemodialysis.1 They are associated with a low risk of infection, and when mature, they are extremely durable. Unfortunately, many patients lack adequate autologous conduits for an AVF, or they have an AVF that is inadequate for effective hemodialysis. As a result, the use of PTFE AVGs for hemodialysis access is increasing.1 Unfortunately, the mean primary patency of PTFE AVGs is reported to be only 13 months and may be decreasing.1, 2
The most common cause of late failure (> 1 month) of a well-functioning PTFE AVG is venous outflow tract stenosis due to intimal hyperplasia. Circulating factors, hemodynamic stresses, and compliance mismatch have all been described as causing venous anastomosis intimal hyperplasia.6 Compliance mismatch describes the differences in the deformation properties of prosthetic grafts and native vessels. These differences create a region of stress at anastomoses that increases graft thrombosis.7 Tyrell et al8 have noted that PTFE anastomosed to small arteries also causes anastomotic distortion. Clark et al9 have postulated that chronic compliance mismatch and anastomotic distortion lead to chronic irritation of the native vessel, intimal hyperplasia, and early PTFE graft failure.
Intimal hyperplasia at the distal anastomosis of infragenicular PTFE grafts results in premature graft failure and poor primary patency rates.10, 11, 12, 13 To counter this, several authors have interposed a vein collar or patch between PTFE bypass grafts and infragenicular arteries.3, 4, 14 Theoretically, a vein collar or a patch decrease anastomotic intimal hyperplasia by (1) decreasing the chronic injury caused by PTFE–native vessel compliance mismatch8; and (2) shifting intimal hyperplasia from the native outflow vessel to the vein collar or patch, thereby limiting its hemodynamic impact.4 In a multicenter, randomized trial, Stonebridge et al15 showed improved patency (50% vs 19% at 2 years) using a Miller vein collar at the distal anastomosis of infragenicular PTFE bypass grafts. A Taylor vein patch has also yielded improved patency (58%4 and 81%16 at 3 years) when used at the distal anastomosis of infragenicular PTFE bypass grafts.
In this study, our hypothesis was that a venous anastomosis Tyrell vein collar would improve AVG patency for the same reasons it improves infragenicular PTFE bypass graft patency. We expected that decreased compliance mismatch and anastomotic distortion at the venous anastomosis would incite less intimal hyperplasia and venous outflow tract stenosis. We are experienced with the Tyrell collar and chose to use this technique rather than other published configurations for a variety of theoretical and technical reasons (Figure).17, 18 The Tyrell collar may be easier to construct than a Linton patch in small vessels, it has a more favorable angle to the native vessel than a Miller cuff, and the vein is interposed between the graft and native vessel along the entire circumference of the anastomosis, unlike with the Taylor patch.3, 4, 16, 17, 18
In contrast to our expectations, this study showed that the use of a venous anastomosis Tyrell vein collar actually diminished AVG patency. The Tyrell collar appeared to accelerate graft failure by promoting intimal hyperplasia in the venous outflow tract. Although the number of patients in the study was small, the marked difference in patency rates between the AVGs in the control and study groups and the recurrent finding of significant venous intimal hyperplasia in the failed study group grafts led us to accept these findings as clinically meaningful, rather than as statistical chance. For this reason, we decided that randomizing more patients to the study arm of this trial was unethical.
Similar results have recently been reported by Pipinos et al.19 In a prospective, randomized trial, they found that although a Taylor patch at the venous anastomosis of an AVG improved early (1 month) patency, at 6 months the grafts with a Taylor patch had patency rates significantly inferior to conventional AVGs (25% vs 47%). Interestingly, intimal hyperplasia in the venous outflow tract was a prominent cause of AVG failure in their vein patch group as well. Unexpectedly, vein collar and patch adjuncts at the venous anastomoses of AVGs seem to predispose to intimal hyperplasia rather than prevent it.
From our study, it is not possible to explain why a vein collar at the distal anastomosis of an arterial bypass graft promotes graft patency, whereas a vein collar at the venous anastomosis of an AVG decreases graft patency. Hofstra et al,20 using vessel-wall Doppler scan tracking techniques, have shown that PTFE-arterial anastomoses respond differently than PTFE-vein anastomoses under arterial pressure. They found that although venous anastomoses increased their cross-sectional area and changed their capacity during the cardiac cycle, arterial anastomoses showed a decrease in cross-sectional area and no capacity change. This increase in the area and capacitance of venous anastomoses during the cardiac cycle may cause flow disturbances and abnormal wall shear stresses that stimulate intimal hyperplasia.20 A vein collar at the venous anastomosis of an AVG may compound hemodynamic stresses that overcome any benefit derived from decreasing compliance mismatch. This may explain, in part, why a vein collar and patch at the venous end of an AVG seemed to promote intimal hyperplasia.
In conclusion, we found that a Tyrell vein collar placed at the venous anastomosis of AVGs failed to improve primary graft patency when compared with conventional graft-vein anastomoses. A Tyrell collar at the venous anastomosis of an AVG appeared to increase intimal hyperplasia, venous outflow tract stenosis, and early AVG thrombosis. A vein collar at the venous anastomosis of an AVG seems to promote a different perianastomotic response than a vein collar at the distal anastomosis of an arterial bypass graft. This suggests that the interactions of prosthetic grafts with veins may be different than the interactions of prosthetic grafts with arteries. This may be an important consideration in future attempts to improve primary AVG patency.
References
- . NKF-DOQI clinical practice guidelines for vascular access. Am J Kidney Dis. 1997;30(Suppl 3):s140–s191
- . Long-term patency of polytetrafluoroethylene graft fistulas. Am J Surg. 1987;155:644–646
- . Interposition vein cuff for anastomosis of prosthesis to small artery. Aust N Z J Surg. 1984;54:283–285
- . Improved technique for polytetrafluoroethylene bypass grafting: long-term results using anastomotic vein patches. Br J Surg. 1992;79:348–354
- Barriers to adequate delivery of hemodialysis. Am J Kidney Dis. 1998;31:593–601
- . An overview of intimal hyperplasia. Surg Gynecol Obstet. 1990;171:433–447
- . Effect of compliance mismatch on vascular graft patency. J Vasc Surg. 1987;5:376–382
- . Experimental evidence to support the use of interposition vein collars/patches in distal PTFE anastomoses. Eur J Vasc Surg. 1990;4:95–101
- . Mismatch of mechanical properties as a cause of arterial prostheses thrombosis. Surg Forum. 1976;27:208–210
- Six-year prospective multicenter randomized comparison of autologous saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J Vasc Surg. 1986;3:104–114
- Results of revascularization and amputation in severe lower extremity ischemia: a five-year clinical experience. J Vasc Surg. 1985;2:174–185
- . What is the proper role of polytetrafluoroethylene grafts in infrainguinal reconstruction?. J Vasc Surg. 1989;10:299–305
- Correlation of operative findings with angiographic and noninvasive hemodynamic factors associated with failure of polytetrafluoroethylene grafts. J Vasc Surg. 1984;l:136–148
- . Linton patch angioplasty—an adjunct to distal bypass with polytetrafluoroethylene grafts. Ann Surg. 1984;199:684–692
- . Randomized trial comparing polytetrafluoroethylene graft patency with and without a Miller cuff. Br J Surg. 1995;82:548–563
- . PTFE bypass grafting to isolated popliteal segments in critical limb ischaemia. Eur J Vasc Surg. 1993;7:26–30
- . New prosthetic venous collar anastomotic technique: combining the best of other procedures. Br J Surg. 1991;78:1016–1017
- . Anastomotic intimal hyperplasia: mechanical injury or flow induced. J Vasc Surg. 1992;15:708–717
- . Early experience with Taylor and notched vein patching in the construction of arteriovenous hemodialysis access procedures: results of a prospective study. In: 6th ed. Henry ML editors. Vascular access for hemodialysis. Chicago: : WL Gore and Associates, Inc, and Precept Press; 1999;p. 213–222
- . Mismatch in elastic properties mound anastomoses of interposition grafts for hemodialysis access. J Am Soc Nephrol. 1994;5:1243–1250
☆ Competition of interest: nil.
☆☆ Reprint requests: Paul J. Gagne, MD, NYU Medical Center, 530 First Ave, Suite 6F, New York, NY, 10016 (e-mail: paul.gagne@ med.nyu.edu ).
PII: S0741-5214(00)19479-0
doi:10.1067/mva.2000.109204
© 2000 Society for Vascular Surgery and The American Association for Vascular Surgery, a Chapter of the International Society for Cardiovascular Surgery. Published by Elsevier Inc. All rights reserved.
Refers to article:
- Regarding “The effect of venous anastomosis Tyrell vein patch collar on the primary patency of arteriovenous grafts in patients undergoing hemodialysis” and “Effects of a venous cuff at the venous anastomosis of polytetrafluoroethylene grafts for hemodialysis vascular access”
