Nephrol. Dial. Transplant. 2002 Konner 376 9

8/13/2019 Nephrol. Dial. Transplant. 2002 Konner 376 9

1/4

Invited Comment

The anastomosis of the arteriovenous fistulacommon errors and

their avoidance

Klaus Konner

Dialysis Unit, Krankenhaus Koln-Merheim, Koln, Germany

Introduction

In 1966, Brescia, Cimino and colleagues described thesurgical creation of an arteriovenous fistula (AVF) toestablish a vascular access in haemodialysis patientsw1x. Despite the availability and the wide-spread useof grafts in some parts of the world, mainly ePTFE(expanded polytetrafluoroethylene), practically allexperts in the field strongly advised creating anAVF whenever possible because of its lower rate ofcomplicationsw2x.

Despite the generally low rate of complications ofnative AVF, early failure within 1 month has beenobserved in some series in up to 29% of patients w3,4x.Early failure, defined as non-function of the AVF, is

mostly caused by early thrombosis secondary to errorsin surgical technique. Thoughtful analysis of thetechniques of surgical creation of an arteriovenousanastomosis has identified a number of potential errorswhich may contribute to such early failure and whichwill be discussed in the following.

Haemodynamics of the AVF

In the long run, blood-flow rates in a well-functioningAVF can exceed 1000 mlumin or morea dramaticincrease when compared with a blood flow rate of2030 mlumin in a normal peripheral artery and theimmediate post-operative flow rate of 200300 mluminafter opening of the fistula w5x. This is accompaniedby a decrease in peripheral resistance w6x. In thepast, venous dilatation has been the focus of atten-tion, but the adaptive venous response is clearly theconsequence of increased flow which in turn dependson arterial dilatation which is the crucial and oftenlimiting step. Obviously, venous remodelling occurs

and some authors used the term venous arterializa-tion to characterize this process. But this terminology

does not take account of the fact that in order toaccommodate blood-flow rates increasing by morethan a factor of 2050 over baseline values, the entirevascular bed must undergo dramatic remodelling inorder to accommodate the extremely high blood-flowrates. Anything impeding the increase in blood flowand vascular remodelling will endanger the final out-come of a well-functioning AVF. It is for this reasonthat the surgeon has to carefully avoid any of the fol-lowing errors which may interfere with the mechanicsand the geometry of the anastomosis.

The arteriovenous anastomosis

Prerequisites

The first step is to select a suitable healthy artery andvein. This can be done by clinical and ultrasonographicinvestigation.

Most procedures creating a primary vascularaccess can be carried out under local anaesthesia.The first potential mistake concerns the incision: skinincisions (and the scar that develops) must nevercross the arterialized vein, particularly not close tothe anastomosis. Transverse skin incisions do notpermit exploration of more cephalad segments and

cause unnecessary destruction of lymphatic vessels.Consequently, we favour longitudinal incisions and usethis approach even in the elbow region.

Formation of post-operative haematoma mustbe avoided and this requires an extremely cleanpreparation of tissues and vessels using bi-polarelectrocoagulation.

The goal of atraumatic closure of the skin canbe achieved by using a few, but well adapted, sub-cutaneous sutures and sterile adhesive strips. Toachieve high quality surgery, microsurgical instru-ments, magnification glasses, or a microscope areindispensablew7x.

Correspondence and offprint requests to: Klaus Konner, MD, Med.Klinik I, Krankenhaus Merheim, Ostmerheimer Strae 200, D-51109Koln, Germany. Email: [email protected]

Nephrol Dial Transplant (2002) 17: 376379

#2002 European Renal AssociationEuropean Dialysis and Transplant Association


2/4

Technical errors

Whenever the arterial or venous vessels are pulled andstretched, the mechanical trauma will chronically leadto long segment stenosis.

Whenever the length of the venous segment isexcessive, there is a risk of kinking. The risk is com-pounded by the fact that during maturation of theAVF, the vein will further elongate secondary to theincrease in blood flow rate and vascular remodelling.

Artery-side-to-vein-side anastomosis

When an artery-side-to-vein-side anastomosis iscreated, the arterial and venous incision should bemade exactly on the lateral aspect of the vessels. Ifvessels that cross each other must be anastomosed,the incisions should be located on the top and thebottom of the respective vessel. In performing theseincisions, there is a great risk of causing lesions tothe posterior wall. If this occurs, immediate and

meticulous repair is mandatory, otherwise there is arisk of late aneurysm formation anduor perforation.

If a stiff arterial vessel has to be operated upon, it isadvisable to excise some tissue at the site of thearteriotomy, otherwise turbulences will develop andinterfere with the arterial inflow into the vein. Thelumen and the distensibility of the vein should be testedby gently compressing the vein and watching the fillingof the vessel. The use of a Fogarty catheter for theseprocedures should be avoided because of the risk ofinjury to the intima. It is indispensable to testthe venous run-off as described because optimal func-tion of the anastomosis requires, among other things,

optimal venous drainage.

Artery-side-to-vein-end anastomosis

It is more demanding to create an artery side-to-vein-end anastomosis. This type of anastomosis requiresmuch more care, experience, and power of imagina-tion, i.e. three-dimensional visualization of the finalresult. With this technique the stump of the vein isisolated and has to be approximated to the arteryacross a certain distance. Unavoidably, the anglebetween artery and vein at the site of the anastomosiswill differ from case to case. Each angle requiresan individual length of the arteriotomy and venotomy

w8x. If the vein approaches the artery at a rightangle, the length of the arteriotomy should beequal to the diameter of the vein; if the vein approx-imates the artery under an acute angle, the length ofthe venotomy and arteriotomy is defined by thelongitudinal axis of the artery as shown in Figure 1.

Another point requires attention. If the veinapproaches the artery at a right angle, outwarddirected rotation is necessary to avoid kinking(Figure 2). Such kinking may easily escape detectionbecause it takes place in the most proximal part of themobilized vein which is often hidden. It is thereforemandatory to inspect this area. If the smooth vein

yields the typical thrill upon palpation, torsion orkinking is unlikely. If, in contrast, a water hammerpulse is palpated, one should suspect stenosis, kinkingor torsion.

Handling of the vein

The vein must be mobilized and adapted to the artery.It is important, however, to limit the length of themobilized vein to an absolute minimum. Mobilizationimplies trauma to and devascularization of the venouswall secondary to interruption of vasa vasorum andremoval of the adventitia. Obviously this increases thelong-term risk of scar formation of the damagedvenous wall. This is particularly true since aftercreation of the anastomosis, high-flow rates and highpressure further increase the risk of scar formationand, in addition, aggravate the risk of torsion and

kinking. Because trauma to the vein must be avoided,we do not recommend the smooth loop technique ofKarmody w9x, although at first sight it appearsattractive because it provides a favourable haemo-dynamic configuration (Figure 3). For the samereason, other types of venous transposition shouldbe avoided whenever possible. More favourable con-ditions are found when a pre-dilated basilic vein in theupper arm will be placed in a subcutaneous position(superficialization).

It is an error to clamp the thin-walled vein duringthe procedure. Clamping causes injury and oedema,particularly to the intimal layer. This carries the

Fig. 1. Individually tailored length of arteriotomy in artery-side-to-vein-end anastomoses dependant on the angle between artery andvein.

Fig. 2. Outward directed rotation of the vein in a rectangularartery-side-to-vein-end anastomosis.

377The anastomosis of the arteriovenous fistula


3/4

long-term risk of stenosis. Instead of clamping oneshould prevent venous backflow during suturing bygently applying proximal digital pressure.

A known complication is arterial as well as venousspasm. Should this occur we advise to rinse the vesselfrom the outside using warmed 0.9% saline with orwithout addition of papaverine or nitroglycerine. Ifthis alone is not effective, mechanical tricks suchas proximal venous compression or, if everythingfails, introduction (with utmost caution) of a Fogartycatheter or olives can be successful. It is wrong tosuture the skin without making sure that vascularspasm is absent. Even a high blood-flow rate alone willnot be able to overcome and resolve venous spasm.

The suture

A standard technique should allow the placing of anyneedle stick to be extremely well controlled. Thetechnique of Tellis w10x is very advantageous and thisis true even for the very small vessels of paediatricpatients; the suture starts in the centre of the back wallof the arteriotomy and venotomy. Suturing is con-tinued passing the corners with excellent visualizationthroughout the procedure as shown in Figure 4. Thisvariant can be used in artery-side-to-vein as well as

side-to-side anastomoses.

End-to-end anastomoses

In the past, end-to-end anastomoses had beenfashionable. This has several disadvantages. The dia-meters of the artery and the vein differ, and this has tobe overcome by inserting a rhombus-like vein patchinto the suture. The suture itself is performed in threeindependent sections without any connection of theclosing knots. The main objections arise from the factthat one has to sever the distal radial artery. Thisprocedure is dangerous in diabetic and elderly patients

who constitute the majority of the patients that areseen to date. As an initial vascular access procedurethis technique should be abandoned today.

A comment concerning grafts

Although grafts should be avoided as the primaryvascular access whenever possible, I wish to commenton the graft-vein anastomosis. Following the recom-mendations of one of the manufacturers of PTFEgrafts, the graft should meet the vein at an angle of

approximately 608. This implies, however, that thehigh-velocity blood-flow bounces against the oppositewall of the vein. Intimal hyperplasia is then unavoid-able. I have found that the frequency of intimal hyper-plasia can be substantially reduced with a simpletrick, based on mechanical considerations. If the graftapproaches the vein in an almost parallel positionand the length of the graft-vein anastomosis isincreased up to 2030 mm by creating a sharp angletip (Figure 5), blood will flow through the anastomosiswithout a major change in direction, thus avoidingturbulence. Using this technique, we have had excellentresults for many years; but I admit that controlled

evidence is not available. I am convinced that thefrequent and costly complications of graft-vein ana-stomosis could be reduced by modification of theanastomosing technique when this simple mechanicalprinciple is respected. Whether this procedure allowselimination of the vibrations at the graft-vein ana-stomosis is not known, but is certainly an interestingscientific topic.

The suturing material

No reliable information is available on which suturingmaterial is best: polypropylene, polytetrafluoroethylene

Fig. 3. The smooth loop-anastomosis by Karmody w9x, an earlyexample for venous transposition.

Fig. 4. Anastomosis technique proposed by Tellis w10x.

378 K. Konner


4/4

and others are currently in use; some authorsrecommend absorbable material. Reliable studies arenot available to compare the different types of suturematerial and therefore the selection has to be based onindividual experience.

Non-surgical causes of early clotting

It is beyond the scope of this brief report to discussthe non-surgical causes of early AVF failure: Systemic

clotting disorders, hypotensive episodes (particularlyduring haemodialysis sessions secondary to excessiveultrafiltration), or haematoma formation. The lattermay be due to bleeding after over-heparinization orfollowing very early cannulation with injury to thevessel wall or following inappropriate compressionafter withdrawal of the cannula.

Comment

Creation of an arteriovenous anastomosisis it asimple procedure that can be delegated to the inexperi-

enced surgical beginner? The answer clearly is no.

In my experience, creation of the initial vascularaccess with the arteriovenous anastomosis as centralprocedure is more challenging than most revisions thatmay seem spectacular to the non-expert.

The surgeon has to be aware of the anatomical,physiological, haemodynamic, and mechanical prin-ciples underlying the procedure and this has to becombined with manual skill, experience, and creativ-

ity. There is no place for compromise. Even minimalerrors are not tolerated: the vulnerable venousvasculature is abruptly confronted with the unusuallyhigh-flow, high-pressure, high-velocity conditions andthis constellation punishes all surgical mistakes. Evenminor narrowing in the beginning will eventuallytranslate into late stenosis. So, not only the early, butalso the late failure rate reflects on the quality of thevascular surgeon.

Failure and success are close neighbours. Ourpatients will be grateful for the successes.

References

1. Brescia MJ, Cimino JE, Appel K, Hurwich BJ. Chronichemodialysis using venipuncture and a surgically createdarteriovenous fistula. N Engl J Med1966; 275: 10891092

2. National Kidney Foundation. KuDOQI Clinical practiceguidelines for vascular access, 2000. Am J Kidney Dis 2001;37wSuppl 1x: S137S181

3. Palder SB, Kirkman RL, Whittemore AD, Hakim RM,Lazarus JM, Tilney NL. Vascular access for hemodia-lysis. Patency rates and results of revision. Ann Surg 1985;202: 235239

4. Fernstrom A, Hylander B, Olofsson P, Swedenborg J. Longand short term patency of radiocephalic arteriovenous fistulas.Acta Chir Scand1988; 154: 257259

5. Wedgwood KR, Wiggins PA, Guillou PJ. A prospective study ofend-to-side vs. side-to-side arteriovenous fistulas for haemo-dialysis.Br J Surg 1984; 71: 640642

6. Sivanesan S, How TV, Bakran A. Characterizing flow distribu-tions in AV fistulae for haemodialysis access. Nephrol DialTransplant 1998; 13: 31083110

7. Bourquelot P, Cussenot O, Corbi Pet al. Microsurgical creationand follow-up of arteriovenous fistulae for chronic haemo-dialysis in children. Pediatr Nephrol1990; 4: 156159

8. Konner K. A primer on the av fistulaAchilles heel, butalso Cinderella of haemodialysis. Nephrol Dial Transplant1999;14: 20942098

9. Karmody AM, Lempert N. Smooth loop arteriovenous fistulasfor hemodialysis. Surgery 1974; 75: 238242

10. Tellis VA, Veith FJ, Sobermann RJ, Freed SZ, Gliedman ML.Internal arteriovenous fistula for hemodialysis. Surg Gynecol

Obstet1971; 132: 866870

Fig. 5. The parallel position graft-vein anastomosis.

379The anastomosis of the arteriovenous fistula

Documents

Nephrol. Dial. Transplant. 2002 Konner 376 9