GUIDELINES [ back to index ]
6. Placement of graft
*1 Grafts only for patients with no possibility of an autogenous arm A/V fistula
In general, native A/V fistulas must always be preferred to grafts, as grafts are associated with higher rates of infection 1 2 3 and thrombosis 1 4. Even after successful thrombolysis of thrombosed grafts, the outcome in grafts is worse than in autogenous A/V fistulas 5 thus leading to high rates of intervention and revision 1 6 7. Thrombosis occurs at a rate of 0.5 2.5 events/patient year 8 compared to 0.2 events per patient year in native A/V fistulas. Failure rates are 20% and 10 % per year for grafts and native A/V fistulas, respectively (see appendix III for details).
In addition, the mortality risks in patients with grafts and especially with central venous catheters is increased when compared to patients with native A/V fistulas 9.
For those patients who start dialysis with a central venous catheter and who then develop life-threatening catheter-related morbidity, the lead time needed to perform, await development of or create a secondary access if the first one fails, may result in an unacceptably high risk. In such exceptional patients, a peripheral arteriovenous graft may serve as an alternative to an arteriovenous fistula, despite the poorer long-term result with grafts. This provision does not condone graft insertion except in very exceptional cases.
In elderly patients not suitable for autogenous proximal access, a proximal brachioaxillary prosthetic graft is considered a good alternative 10. Grafts have a low early failure rate of < 6 % (see appendix III), but face the risk of thrombotic occlusion due to impairment of flow caused by intimal hyperplasia-related stenoses at the venous anastomosis. Without a surveillance program and active interventional treatment of stenotic lesions, the patency rate of grafts is rather low with 29-60 % at 1 year 11 12 and approximately 40 % after 2 years 11. With surveillance and intervention with percutaneous transluminal angioplasty (PTA), the thrombosis rate can be reduced to less than 0.5 thromboses / patient-year 8 .
*2 Identify location of arterial and venous arm vessels with adequate diameter
An objective assessment of arm arteries and veins is obligatory especially if a lack of suitable vessels for primary radiocephalic A/V fistulas is suspected or if previous A/V fistulas have failed. A standardised duplex scan is performed with special emphasis on elbow and upper arm basilic/cephalic/antecubital veins and deep veins. Also, the diameter and the quality of axillary and brachial artery are examined. An outflow vein of > 3.5 - 4 mm is thought most appropriate for graft implantation 13 11, although there is no prospective study proving this.
*3 Graft Placement
Haemodialysis access grafts should be implanted only after exhaustion of superficial veins in both upper extremities.
Forearm straight grafts (from distal radial artery to antecubital or peripheral brachial vein) provide a sufficient graft length (20 cm or more) in normal sized adults. However, if a suitable vein for graft placement is found in the elbow, this vein might also be suitable for the creation of a native A/V fistula.
Upper arm grafts should be sutured to the larger veins found in the upper third of the arm. The first graft-to-vein anastomosis should not be performed high in the axilla, since surgical revision of anastomotic stenosis is easier and can be performed under local anaesthesia.
Different outcomes have been described in looped compared to straight grafts. Lazarides et al. described a higher risk of access failure in straight forearm grafts compared to looped forearm grafts, from antecubital artery to antecubital or peripheral brachial vein. However, there is a risk of steal syndrome when the graft is placed on the brachial artery but not when placed on the radial artery. In the upper arm, looped grafts performed worse than straight grafts, from distal brachial artery to axillary vein 14. The better patency rates of looped grafts in the forearm may be due to the larger diameter of the artery at the anastomosis and/or due to the haemodynamically favourable antegrade inflow. However, forearm grafts can preclude later creation of an elbow fistula. Upper arm grafts should be performed only after failure of forearm and elbow fistulas (transposed brachiobasilic fistulas included) 14.
When the subclavian vein is occluded, brachial artery to jugular vein grafts or axillary artery-to-internal jugular vein cross-over grafts should be considered (see Algorithm Management of Central Venous Stenosis (2), *7) 15 16 17 18. Patency and complication rates are comparable to those of upper arm grafts, although the number of such procedures is small.
Graft placement can be performed largely as an outpatient procedure 19, thus reducing costs, although the age and infirmity of the patient needs to be taken into account, and not cost alone.
*4 Venous anastomosis
For reasons of preservation of the venous tree, the venous graft anastomosis should be placed as peripherally as possible. Veins below the antecubital fossa, however, do not usually have a diameter of 4 mm or more needed to guarantee graft patency 11. Therefore, forearm grafts will end at the antecubital fossa or even above the elbow.
While it is obvious that the arterial graft anastomosis is to be sutured in an end graft to side artery fashion, comparative studies on venous end-to-side versus end-to-end anastomosis are absent. A vein cuff to widen the venous end-to-side anastomosis has not been shown to improve patency rates 11. Early graft failure is correlated with the initial diameter of the vein at the anastomosis. Suturing the graft to a vein smaller than 4 mm is not ideal 11.
A graft usually will not sustain a blood pressure < 100 mm Hg, so it should not be placed into a haemodynamically unstable patient.
*5 Graft length
The length of the graft should exceed 20 cm to provide at least 15 cm for rope-ladder puncture with two needles. Excessive graft length (> 40 cm) is believed to be associated with higher thrombosis rates.
*6 Graft diameter
The ideal graft diameter still remains to be defined. Small calibre grafts, make needling difficult, whereas large diameter grafts run the risk of high arterio-venous flow with consequent cardiac stress and/or peripheral steal syndrome. Therefore, 6 mm or 7 mm grafts are a widely accepted compromise.
In order to reduce midgraft stenosis observed with smaller graft sizes, Polo et al.. tapered an 8 mm graft to 6 mm at the arterial anastomosis. Primary and secondary patency rates at 1, 3 and 5 years were 73 %, 53 %, 41 % and 91 %, 80 % and 72 %, respectively 20. In a second study, Polo et al.. compared primary and secondary patency rates in grafts of 6 mm and 6-8 mm. Primary patency rates were 62 %, 58 % and 44 % in 6 mm grafts and 85 %, 78 %, and 73 % in 6-8 mm grafts at year 1, 2 and 3. Secondary patency rates remained stable during the first three years at 85 % for 6 mm and 90% for 6-8 mm grafts, respectively (unpublished data).
*7 Graft material
Expanded polytetrafluoroethylene (ePTFE) is the most frequently used graft material for prosthetic vascular access and recommended by the K/DOQI guidelines 21. In clinical trials ePTFE has been compared to other graft materials such as plasma tetrafluoroethylene 21 22, or polyurethane 23. In these studies, none of the alternatives performed better than ePTFE. Standard wall thickness ePTFE provides better patency rates than thin-walled ePTFE 24 (see appendix III for details).
Tordoir et al. found higher patency rates in stretch ePTFE grafts than in standard ePTFE grafts 12.
PTFE grafts have a high incidence of postoperative infection which is 6 % during the initial 30 days. Zibari et al... were able to reduce the incidence of infection down to 1 % with 750 mg of vancomycin i.v. given 6-12 hours before the operation 25. Others feel that this policy may induce resistance to this important antibiotic and prefer to use a cephalosporin. However, there has been no randomised controlled trial. With a preoperative dose of 2 grams of Cephalzoline, 30 days infection rates can also be as low as 0.9 % 26.
*8 Non-arm sites for graft placement
Only in those rare cases when the arterial and venous trees of both arms (including the axillary arteries and veins) are unsuitable for arterio-venous access, non-arm sites should be considered. Basically, a bridge graft can connect virtually any artery to any vein to form an arterio-venous access. Artery and vein should preferably have a diameter exceeding 4 mm 11, and the veins central to the graft have to be free from obstruction. Colour-coded duplex ultrasound is the method of choice to define the vessels best suitable. In questionable cases, angiography and/or venography need to be performed.
*9 Graft placement at non-arm site
The thigh is the next common site for fistula creation after the upper limbs.
Several options exist for autogenous AV fistulas in the thigh, such as transposition of the long saphenous vein (if not already translocated to the arm to create an AV fistula) or superficial femoral vein (SFV). Gradman et al. constructed AV fistulas using the transposed SFV, with primary and secondary patency at one year of 73 % and 86 % respectively 28. No infection occurred and 7 patients required PTFE graft insertion.
The PTFE graft fistula is the commonest type of vascular access created in the thigh. Patency and complication rates of lower limb grafts are comparable to those of upper arm grafts 27. A straight graft can be anastomosed to the popliteal artery or distal superficial femoral artery and Tunneled up to the long saphenous vein or common femoral vein in the groin, or, in reverse, from a femoral artery down to the origin of the superficial femoral vein in the distal thigh. Alternatively a femoro-femoral looped graft can be performed. The arterial in-flow can be from all three groin arteries the common femoral artery, profunda femoris or the superficial femoral artery (SFA) origin. The advantage of the profunda or superficial femoral artery is that in the rare instance when there is serious infection of the graft arterial anastomosis, the profunda or SFA can be ligated without compromising perfusion of the whole leg, as would happen in ligating the common femoral artery.
The femoral artery-to-femoral vein looped bridge graft is the most frequently performed fistula in the lower limb.