GUIDELINES [ back to index ]

Appendix

1. Diagnostics
2. Vascular access creation
3. Monitoring
4. Treatment of stenosis and thrombosis
5. Vascular access infections in dialysis patients

• references

1.    Diagnostics

1.1.  Physical Examination  

1.1.1. Examination of Veins  

Veins should be dilated during assessment. Therefore, examination should take place in a warm room with the patient feeling  comfortable (discomfort and cold lead  to vaso-constriction). Examination should include :

1.    Examination in general

Inspection of both upper limbs for scars since these may indicate former vascular accesses, other operations, central venous catheter, infection, venous collaterals around the shoulder.

2.    Examination for size and location of veins

A blood pressure cuff proximally placed on the arm and inflated to  60 mm Hg or a tourniquet can be used. Clenching and unclenching the hand  helps to engorge the veins. The  course of the veins should be palpated and they should be distensible up to 2 - 3 mm. The vein should feel soft with no narrowed or thickened areas.

3.    Examination for venous outflow

By light percussion  on the vein at the wrist, a transmitted wave should be felt at the vein in the elbow region proving the patency of the vein. Before the pressure cuff or tourniquet is released , the arm should be elevated. After releasing the cuff tourniquet, the veins should empty very quickly. Prolonged emptying may suggest  venous outflow stenosis.

1.1.2. Examination of Arteries

1.     Examination for central arterial stenosis

Blood pressure in both arms should be measured by the Riva-Rocci method. Differences in the systolic blood pressure of > 20 mm Hg, or relative differences of > 10 % (corresponding to an arm/arm index of < 0.90, i.e. systolic blood pressure of the arm intended for vascular access creation divided by the systolic blood pressure of the contralateral arm) is indicative of proximal arterial obstructive disease. In these cases, surgical or radiological intervention for the underlying cause should be performed before creating an A/V fistula.

2.     Examination of peripheral arteries

The pulse of the brachial artery is palpated at the medial side of the elbow, the radial and ulnar arteries are palpated 2 cm proximal of the processus styloideus radii and proc. styloideus ulnae, respectively. Calcified arteries are generally hard and non-pulsatile. A weak or absent pulse might indicate obstruction of the artery. Auscultation of the subclavian or axillary artery may  reveal a bruit suggesting stenosis .

3.     The Allen-Test

The patency of the distal radial and ulnar artery and the palmar arch can be examined by the Allen test ^{1 2}. The radial and ulnar arterial refilling of the hand is examined by the selective compression and decompression of the radial and ulnar artery at the wrist. In the modified version, the patient is asked to clench the fist, meanwhile both arteries are compressed. Then the fist is  unclenched followed by release of the radial artery.  Adequate hand perfusion results in a return of capillary refilling within 5 seconds. The test is then repeated but this time releasing the ulnar artery whilst keeping the radial artery compressed.  In 1976, Kamienski et al. pointed out that to obtain results correlating with Doppler ultrasound measurements, the proper performance of the Allen test is essential,  i.e. that there is no hyperextension of the wrist or the fingers as this may  reduce bloodflow ³.

However, the accuracy of the Allen test to detect distal radial or ulnar artery occlusion is doubtful. Jarvis et al. correlated clinical examinations by means of the Allen test with Doppler examination of the radial and ulnar artery. With cut-off values of 3, 5 and 6 seconds for capillary refilling-time the sensitivity was 100%, 76% and 54%, respectively, with a specificity of 27%, 82% and 92%. The diagnostic accuracy was 52%, 80% and 79%, respectively ⁴.

McGregor performed intra-arterial fluorescence angiography and found no correlation with clinical testing ⁵. Inter-observer variation concerning the Allen test is also reported ⁶.

On the other hand some investigators found a good correlation between objective measurements (Doppler ultrasound or angiography) and clinical assessment of the distal wrist arteries, and considered it as useful in the diagnosis of the patency of the palmar arch, despite the high percentage of false negative results (almost 20 %)  ⁷. Others recommend it only as a primary screening method ⁸, whose accuracy can be increased by the use of laser-Doppler measurement, but not by pulse oximetry  ⁹.

In conclusion, the Allen test is a subjective assessment of distal hand circulation with variable outcome. Therefore, it is not recommended in the assessment for AV fistula placement. Digital blood pressure monitoring has been described as an alternative to the Allen test ¹⁰.

1.2 Technical Methods

1.2.1  Iodine Venography

Iodine contrast media can cause a deterioration  in renal function. Despite its possible nephrotoxicity, highly diluted iodine contrast material is still recommended by several authors for angiography  and fistulography, as the amount needed (less than 5 ml ) has virtually no effect on renal function. Digital subtraction angiography is preferred ^{11 12}.

In patients with severe iodine allergy or with  poor renal function,  CO₂-Venography may  be an alternative . With CO₂ neither nephrotocixity nor allergic reactions are known. However the contrast between CO₂ and blood is lower than between iodine and blood. Thus, CO₂-Venography requires digital subtraction angiography.

The main risk in the use of CO₂-Venography is:

• Trapping of CO₂ in the right ventricle

Trapping of CO₂ in the right ventricle may  cause block (vapour lock) of the pulmonary outflow tract, thus leading to coronary ischaemia, bradycardia, hypotension and possible death. To avoid this complication, the injected volume of CO₂ should be restricted to 50-100 ml per single injection (100 ml  CO₂ at 4 bar is equivalent to 400 ml  CO₂ at 1 bar pressure, Boyle’s Law). Between each injection, 1 - 2 minutes should elapse to allow for CO₂ clearing by the lung.

If trapping of CO₂ in the  right ventricle  has occurred, the right side of the patient should be elevated in order to lower the pulmonary artery and to move the CO₂ into the right atrium ¹².

Gadopentetate dimeglumine, a gadolinum-based contrast medium, adequately visualises the  renal artery as well as veins of the upper extremity and central veins, if used with digital subtraction. It has been used safely in patients with severe pre-existing renal impairment and  is thought to be less nephrotoxic  than iodine contrast medium ^{13 14} although this remains controversial ¹⁵.

1.2.4.    Magnetic Resonance Angiography (MRA)

Peripheral and central veins have been successfully visualised by MRA, a non-invasive technique, which can also be performed with or without contrast media (Gadolinum) using the time-of-flight (TOF) technique. In the latter, veins with a diameter > 2 mm could be visualised ¹⁶. MRA has been proven to be accurate and even superior to contrast venography for thoracic vessels ¹⁷.

1.2.5.    Doppler Ultra sound

The Doppler ultrasound technique uses the Doppler effect - meaning that the frequency of sound waves change if they are reflected by a moving object. The change is dependent on the velocity of the object.  Circulating blood cells (predominantly erythrocytes) reflect the ultrasound. Ultrasound can either be transformed into  a sound that can be heard  or  visualised in a waveform.

Access flow can be measured by Doppler ultrasound with the caveat that Doppler techniques require laminar flow for accurate measurements of velocity whilst blood flow in a  fistula is usually turbulent.  Measurement must be performed at least 5cm away from the A/V fistula anastomosis.

1.2.6.    Duplex Colour-Coded Ultrasound

Duplex ultrasound is a combination of B-mode ultrasound, producing a two-dimensional picture of the anatomy, and of Doppler ultrasound. The velocity of the blood is shown using different colours for different velocities and different directions of blood flow. Blood flow in the vessel can also be calculated. Thus it is possible to visualise in a single screen both the anatomy and the blood flow in vessels.

Duplex Doppler flow measurements can be predictive  of thrombosis, but the disadvantages are the great inter-observer and inter-device variability, the adequacy being dependent on the anatomic location of the access, and the high costs ¹⁸.

Colour coded duplex ultrasonography is able to replace angiography except for hand arteries and central veins ¹⁹.

Normal Doppler waveform (measured by pulse wave Doppler) of a peripheral artery shows “high resistance”, with positive systolic velocity and negative diastolic velocity. The feeding artery of an A/V fistula on the other hand shows a flow of  “low resistance”, with positive systolic (SV) and positive diastolic (DV) velocity. The resistance index (RI) can be calculated by:

RI = (SV-DV)/SV

Resistance index of a peripheral artery is normally above 1 and in a feeding A/V fistula artery < 1. At reactive hyperaemia, provoked by clenched fist  for two minutes, “high resistance” flow of a peripheral artery (of the upper limb) is changed to “low resistance”, and RI changes from >1 to < 1. This could be used as a test for functional characteristics of the artery, which should be used for the creation of an A/V fistula ²⁰.

First choice are direct wrist A/V fistulas, second choice are autogenous A/V fistulas at the elbow - brachiocephalic, transposed brachiobasilic, and autogenous vein transfers ²¹.

2.1. Upper Limb – Forearm A/V Fistula

The most peripheral location of the A/V anastomosis provides an optimal  length of vein for cannulation.

2.1.1. Radiocephalic Fistula

Brescia et al. in 1966 originally described a side-to-side radial artery-to-cephalic vein A/V fistula. Today, the end-of-vein to side-of-artery anastomosis is preferred, as the patency rate is similar but there is a lower incidence of venous hypertension of the hand compared to side-to-side fistulas ²². The anastomosis is performed close to the wrist joint, if adequate vessels exist  at this site  (see Fig. 1).

Vein transposition, such as the forearm basilic or the long saphenous vein, increases the possibility  of  creating  a forearm A/V fistula, however, the transposition of the vein can  reduce its capacity to dilate and mature.

2.1.2. Other Forearm Fistulas

Less common are snuff-box radiocephalic fistulas (created in the anatomical snuff box between the tendons of extensor pollicis longus and extensor pollicis brevis) ²³ and ulnar basilic fistula (created between the basilic vein and the ulnar artery at the wrist) ²⁴.

2.2. Elbow/Antecubital Fistulas and Upper Arm A/V Fistula

If the creation of a forearm A/V fistula is deemed impossible, upper limb fistulas created at the elbow are the second choice. Controlled, randomised trials comparing upper arm A/V fistulas and prosthetic grafts are absent. Oliver et al. retrospectively analysed transposed brachiobasilic fistulas, upper arm grafts and brachiocephalic fistulas. They found higher thrombosis rates (50% vs. 23% at one year), higher infection rates (13% to 0 %) and increased requirement  for intervention (2.4 vs. 0.7 per access-year) in upper arm grafts compared to transposed brachiobasilic  fistulas ²⁵. Therefore, creation of an upper-arm A/V fistula should be preferred to graft placement.

Depending on the anatomy of the antecubital fossa several techniques are described to connect the brachial artery to the cephalic vein. The preferred method is the end-of-vein to side-of-artery configuration which reduces the risk of retrograde venous hypertension (see Fig. 2). This form of fistula may not be suitable for obese patients. Although technically possible, the vein may be located too deep to be needled easily in obese patients and the length of the arterialised vein may be too short. Resection of the fatty tissue can make needling easier. Reported patency rates are comparable to those observed in radiocephalic fistulas  and range between 70 - 75 % after one year ²⁶ to 80 % after 4.5 years ²⁷. The main complications are vein stenosis due to intimal hyperplasia, venous aneurysmal  dilatation and vascular steal syndrome ^{26 28 29}.

A special variant of elbow fistulae, the perforating vein fistula (PVF), was first described in 1977 ³⁰ (see Fig. 3). In many patients, the perforating vein seems to be the continuation of the  cephalic vein. In contrast to the original technical procedure with partial resection of the deep venous system, it is nowadays preferred to preserve the continuity of the deep veins. The perforating vein is transected before entering the deep veins. That means, that the length of the A/V anastomosis is limited to the diameter of the perforating vein, typically 3 to 5 mm. This procedure can advantageously be used for creation of an elbow vascular access in diabetic, female and elderly patients ³¹. A good arterialised cephalic vein provides for long-term  cannulation due to its subcutaneous position.

A further possibility is the creation of a brachiocephalic  bridge  graft fistula. A PTFE graft (usually 6 mm in diameter) is anastomosed to the brachial artery and the cephalic vein ³².

2.2.2.  Brachiobasilic Fistula with Transposed Vein

The anastomosis is constructed – if possible – at the level of the elbow using the median basilic vein, generally as an end-of vein to side-of-artery anastomosis (see Fig. 4). Special attention should be paid to the length of the anastomosis which should not exceed the diameter of the brachial artery  by more than 2 mm and should never be more than 7 mm in total, otherwise the risk of steal syndrome will increase substantially. Recent publications recommend a subcutaneous  superficialisation of the basilic vein up to the level of the axilla ²⁵. However, it might be advantageous to transpose only a segment leading to a cannulation area of 8 - 10 cm, in order to preserve the cephalad segment of the non-transposed basilic vein for future procedures.

f this technique is used without prior arterialisation of the vein, there may be a higher failure rate. A two-step procedure is recommended by performing the initial anastomosis and then transposing the basilic vein after an interval of 4-8 weeks, in order to allow sufficient time for maturation  and blood flow rates of the 500 or 600  ml/min may be measured by ultrasound. In a patient with a pre-dilated basilic vein,  transposition can be performed as a one stage procedure. 

brachiobasilic fistulas can also be performed at the level of the elbow in a side-of-vein to side-of-artery fashion. Thus, all the veins at the elbow, proximal and distal to the anastomosis are available for dialysis. In order to avoid venous hypertension and steal syndrome, the arterial anastomosis should not be more than 7 mm in length ³³.

2.3. Lower Limb Vascular Access

If all vessels in the upper limb are no longer suitable for vascular access creation, a vascular access in the lower limb serves as alternative. Multiple surgical approaches have been used to create a lower limb vascular access. We recommend saphenous vein and superficial femoral vein transposition and superficialisation.

Contra-indication for lower limb access are advanced peripheral arterial disease or critical ischaemia, including critical ischaemia stage III/IV (Leriche-Fontaine classification). Atherosclerosis will  be indicated by the absence of distal pulses in the foot and popliteal fossa. In these cases, the ankle-brachial pressure index should be determined. An index < 0.9  suggests peripheral vascular disease with the possibility of an elevated risk of limb ischaemia and steal-syndrome after creation of a vascular access.

Due to the greater diameter of the vessels, patency rates are in general better than in upper limb A/V fistula. On the other hand, the high flow rates, especially in combination with atherosclerosis of the vessels, carry the risk of steal syndrome and ischaemia. In addition, leg oedema might occur in case of venous obstruction.  Some authors find higher rates of infection in vascular accesses in the lower limb compared to those in the upper limb, whilst others have not ³⁴.

Primary use of prosthetic material for vascular access is not recommended because of its high frequency of complications. Such material should only be used if there is no possibility  of the creation of a autogenous A/V-fistula. Vascular access can be created using bio-grafts and synthetic grafts, to connect an adequate artery with an adequate vein. Figure 5 gives some examples of looped and straight grafts placed in the upper limb.

Bio-grafts can be divided into:

Autografts: Generally the  long saphenous vein or superficial femoral vein of the patient is used and either transplanted to the arm or transposed in the leg.

Allografts: Cryopreserved saphenous or femoral veins may become the graft material of the future. Veins obtained from patients after varicose  vein stripping are rarely used nowadays because of the risk of HIV infection although they have been used in the past.

Xenografts: Today,  bovine mesenteric veins (Procol®) and sheep ovine veins

(Omniflow®) are used as  xenograft. They are denaturated and (in case of Omniflow®) reinforced with a polyester mesh

Synthetic grafts are made from several different materials:

PTFE: Polytetrafluoroethylene is made from Teflon. The expanded PTFE (ePTFE) is composed of solid nodes, connected by thin fibres.

Dacron: Dacron may be a woven or  knitted porous graft made from polyester.

Leakage through the graft can be prevented by pre-clotting or coating with collagen or gelatin.

Polyurethane graft: Polyurethane has a spongy structure that is microporous. Due to its compliance, it may  inhibit intimal hyperplasia, thereby  reducing stenosis rates.

Catheters used for dialysis can be divided into:

1.     Tunneled or non-tunneled catheters
2.     Single or dual-lumen catheters
3.     Polyurethane or silicone catheters

Catheters for dialysis may be  used in cases of acute renal failure or when the catheter will be restricted to  2-3 weeks. Single and dual-lumen catheters can be used. Catheters which are relatively stiff have been associated with perforation of vessels and of the right atrium  and should be avoided ³⁵.

In dual-lumen catheters, the lines  ought not be reversed, as this increases the recirculation of blood, thus reducing the efficiency of dialysis , although sometimes this is unavoidable³⁶. Level et al. described an increase in recirculation from 2.9 +/- 5 to 12 +/- 9 % in reversing the blood lines ³⁷.

Catheters can also be made from different materials. While silicone is incompatible with tincture of iodine and may  be  degraded by povidone-iodine, polyurethane catheters are incompatible with alcohols including isopropyl alcohol, and ointments containing  polyethylene glycol (PEG) ³⁸.

Currently there is insufficient  data to recommend the use of ports as opposed to catheters. Ports ³⁹ or similar systems like Dialock® ⁴⁰ and LifeSite® ⁴¹ are  alternatives to permanent tunneled catheters. These  totally subcutaneous vascular access devices may  reduce the rate of  infection,  since  the skin barrier remains intact. They offer higher blood flow rates than catheters ⁴¹. A high acceptance by the patients has been reported, as they are less visible and allow bathing. However,  the devices  and the special cannulae needed for puncture are costly.

3. Monitoring

Early detection of stenosis of the vascular access and intervention increases patency rates of the vascular access ^{18 42   43 44}. A correct technique for the measurement of recirculation, static or dynamic venous pressure is essential to achieve reliable results:

3.1.     Urea Based Measurement of Recirculation (Slow-Flow-Technique)

Procedure ⁴⁵.

-       Measure during the first 30 min of dialysis;
-       Turn off ultrafiltration;
-       Draw blood samples for urea determination from the arterial and venous blood lines

A = concentration of urea in the arterial line with possible recirculation

V = concentration of urea in the venous line after the blood has been dialysed

-       Reduce blood pump to 50  ml/min;
-       Exactly 15 to 25 seconds later draw another blood sample S from the arterial line port

S = concentration of urea in the arterial line without recirculation

-       Resume dialysis;
-       Recirculation in % = (S – A) / (S – V) x 100

with    S – A = reduction in the urea concentration caused by recirculation

       S – V = reduction in the urea concentration caused by dialysis

3.2.     Static Intra-Access Venous Pressure (SVP)

Procedure ⁴⁶

-       Measure mean arterial blood pressure (MAP) in the contralateral arm

-       Open venous clamp manually and stop blood pump for 30 seconds

-       Determine the arterial and venous offset pressure (P_offa and P_offv) between access site and pressure sensor:

o     Measure the height difference A in cm between access and fluid level in the arterial pressure line. P_offa in mm Hg = A x 0.75

o     Measure the height difference V in cm between access and fluid level in the venous drip chamber. P_offv in mm Hg = V x 0.75

(leading to a P_off > 0 mm Hg, if access arm is below fluid level (normal case), or to a P_off < 0 mm Hg, if access arm is above fluid level (height is then negative))

-       Read the venous and arterial pressure (PA, PV) after pressure relaxation in the pressure display of your dialysis machine

-       Calculate the static arterial and venous intra-access pressure ratio SPR_a and SPR_v:

SPR_a = (PA + P_offa)/MAP

SPR_v = (PV + P_offv)/MAP

Procedure:

-       Use 15 gauge needle size;

-       During the first 5 min of each dialysis session raise the blood flow to 200  ml/min. At higher flow rates, there is excessive turbulence at the level of the dialysis needle and the pressure readings loses its predictive capacity.

-       Read DVP from the venous pressure display in your machine;  

-       Proposed threshold to order angiography is dependent  on the machine. Pressures regarded as elevated are:

-       a DVP > 100 - 150 mm Hg in 3 consecutive readings ^{42 43} or

-       a DVP > 125 mm Hg with a blood flow of 200  ml/min ⁴⁵ resp. a DVP > 150 mm Hg in standard blood flow for the patient ⁴³.

-       Pressures must be consistently elevated in three consecutive dialysis to avoid errors caused by needle placement

3.4. Vascular Access Blood Flow Measurement with Reversed Lines

3.4.1.   Ultrasound dilution

-       Puncture the fistula with the arterial cannula against the flow direction of the vascular access
-       An injection port must be integrated in the venous blood line
-       Clip arterial line flow/dilution sensor onto arterial blood line and venous sensor onto the venous line
-       Reverse the lines for access flow measurement by cross connecting the venous blood line to the arterial needle tubing and vice versa
-       Choose an effective blood flow Qb between 200 and 300  ml/min
-       Extend the limits of the venous pressure to avoid pump stoppage during infusion
-       Draw 10 ml of physiologic saline into a sterile syringe and inject it within 3 to 5 seconds into the venous blood line
-       Determine recirculation RX from dilution data (e.g. a RX of 25% is equal to 0.25)
-       Calculate access blood flow Qa:

[ IMAGE ]

3.4.2.   Thermodilution

-       Puncture the fistula with the arterial cannula against the flow direction of the vascular access
-       Choose an effective blood flow QB of at least 200 ml/min and stop ultrafiltration
-       Measure recirculation RN in normal blood line position (e.g. a value of 5% is equal to 0.05)
-       Stop blood flow and reverse the lines for access flow measurement by cross connecting the venous blood line to the arterial needle tubing and vice versa
-       Restore effective blood flow Qb and measure recirculation RX with reversed blood lines
-       Calculate access blood flow Qa

[ IMAGE ]

The check-box below is an example and should be adopted to local requirements and to the vascular access used in the respective patient.

4.     Treatment of Stenosis and Thrombosis

4.1. Interventional Options

4.1.1.  PTA

4.1.1.1. Dilatation technique ⁴⁷

Regular dilatation is an outpatient procedure. An antegrade venous approach should be used for stenoses located far enough from the arterial anastomosis, and a retrograde approach should be used for stenoses close to the arterial anastomosis. An 18G needle and a 0.035 inch guide wire offer good support for placement of a 6 to 9 F introducer sheath. A short subcutaneous tunnel between the skin entry-point and the fistula or graft entry-point will facilitate the final compression and decrease the risk of   pseudo-aneurysm formation.

Once a guide wire has passed through the stenosis, heparin may be injected (2,000 to 3,000 units) but it is necessary only in small diameter or low flow fistulae.

Usually  the diameter of the dilatation balloon should be equal to or 1 millimetre greater than the diameter of the immediately upstream or downstream normal vessel (choose 1 mm above the smaller one in cases of discrepancy).

The balloon is inflated with a manometer filled with contrast medium diluted to 75%. Pressure is slowly increased to abolish the waist of the stenosis on the balloon, the edges of which must be completely parallel. The inflated balloon is left in place for 1 to 3 minutes. Dilatation is often painful locally. Local anaesthesia can be performed when stenoses are just under the skin. Neurolept analgesia may otherwise be necessary.

Haemodialysis access stenoses are often very hard and high pressure balloons with bursting pressures of over 25 atmospheres are often necessary. Immediate post-dilatation angiography, with the guide wire left in place through the dilated area, may show several possibilities:

• No residual stenosis and no wall damage: The procedure is then completed.

• Minor vein wall  damage: 3 to 5 minutes low pressure ballooning is performed to try to smooth out  the vessel wall.

• Rupture with clear extravasation of contrast medium and haematoma: The balloon must be re-inflated rapidly to 2 atmospheres for repeated periods of 10 minutes.

• Residual stenosis: If there is no  vein wall damage, a new dilatation is performed with a longer inflation time (3 min) with a balloon which is 1 mm greater in diameter. At this stage, a less than 30% residual stenosis is acceptable only if it results from the first dilatation ever performed in this stenosis. If there is greater than 30% residual stenosis, a new dilatation should be performed with a larger balloon but more than 2 mm over-dilatation is not recommended. Stenosis recoil is possible, especially in central veins, and should be treated by stent placement.

Once the dilatation has been performed, the catheters and introducer sheath are removed. The puncture site must be compressed as gently as possible to stop bleeding without stopping flow through the fistula. The procedure is thus completed  and the vascular access is immediately usable for haemodialysis.

4.1.1.2.       Specific cases ⁴⁷

• Stenoses resistant to 25 bar pressures are infrequent. Atherectomy catheters ⁶, cutting balloons ⁷ or the Redha-cut (Sheringmed, Switzerland) have been reported to be of some value in such cases.

• Arm oedema is due to stenosis or occlusion of a central vein (subclavian or brachio-cephalic)  and is the consequence  of previous central catheters. In chronic occlusion, it may be necessary to use a double approach via the femoral vein and via the fistula in order to traverse the occluded lumen successfully.

• Stenoses of the feeding artery and of the arterio-venous anastomosis may be impossible to traverse using a regular retrograde fistula approach. Antegrade puncture of the brachial artery and catheterisation of the feeding artery is feasible.

4.1.1.3.       Contra-indications  to dilatation ⁴⁷

Absolute contra-indications

Local infection

Concomitant arterial steal syndrome

Relative contra-indications

Surgical anastomoses of less than 6 weeks are at high risk of rupture. Gentle dilatation at low pressure with an exact size-matched diameter balloon can however be contemplated at venous anastomoses of grafts.

Immature (< 2 months) fistulas  are indications for surgery if the stenosis is located in the anastomotic area. However, when the underlying stenosis is located far from the anastomosis, gentle dilatation is the  simplest method to save the fistula.

Isolated stenosis within  5 cm of the wrist in a Brescia-Cimino fistula can be dilated but should preferably be treated surgically.

Long (> 5 cm) stenoses and chronic occlusions at the venous anastomoses of grafts and in upper arm cephalic veins usually provide poor results and surgical revision should be undertaken.

High flow is mainly a complication of upper arm fistulae and is a contra-indication for dilatation because the treatment of the stenosis would increase the already high flow.

4.1.1.4.       Stents⁴⁸

Indications for stent placement must be restricted, and only self-expandable stents should be placed in dialysis access. Covered stents are only recommended for rupture control. Stent diameter must be at least 1 to 2 mm greater than the diameter of the largest balloon used for dilatation, its length should be as short as possible.

Stent patency is limited because stenosis recurs either within the stent or at its ends. In order not to interfere with future access creation, stent placement is contraindicated at the venous anastomosis of a forearm graft if the stent would overlap the basilic vein. Equally, a stent placed in the final arch of the cephalic vein should  not protrude into the subclavian vein, a stent placed in the subclavian vein must not overlap the ostium of a patent internal jugular vein, and finally a stent placed in the right of left brachio-cephalic (“innominate”) vein must not protrude into the superior vena cava.

Percutaneous declotting of thrombosed autogenous  and prosthetic arteriovenous access can be accomplished by a variety of pharmacological, pharmaco-mechanical, and mechanical approaches.

4.1.2.1.      Pharmacological Thrombolysis

Pharmacological declotting basically means injection of a thrombolytic agent into the thrombosed access. The procedure should always be combined with treatment of the underlying stenosis, since  complete lysis and patency rates following  thrombolysis only have been  disappointing. 

Pharmaco-mechanical thrombolysis is composed of two phases.  First  pharmacological lysis of the thrombus takes place  followed immediately by mechanical maceration and removal of residual thrombus and by dilatation of residual stenosis. When combined with consequent treatment of all stenoses, immediate success rates of more than 90% can be achieved.

Purely mechanical methods include the balloon-based methods of Trerotola and Sharaffuddin, manual catheter-directed thromboaspiration, the saline pulse-spray technique of Beathard, the Gelbfish device, the rotating pigtail of Schmitz-Rode and several types  of declotting machines. Among these machines, some have a direct action (Arrow-Trerotola PTD®, Cragg brush®) but the majority work on the Venturi effect or on the creation of a vortex (Hydrolyser®, Amplatz-Thrombectomy Device®, Angiojet®, Oasis®), and many others are likely to be marketed . No method or device can claim to work better than another but there are difference in costs and in the size  of particles pushed into the lungs ⁴⁸. 

All surgical corrective procedures can be performed in an outpatient setting and under local or regional anaesthesia. Before clamping the access, 3,000 to 5,000 IU heparin are administered. At the end of the operation, after de-clotting  the access, protamine can be given to reverse  heparin effect  (beware of the possibility of  hypotension after protamine administration in insulin-dependent diabetics).  

Fistula vein stenosis is very often found close to the arterio-venous anastomosis in forearm fistulas. After ligation of the vein in its stenosed segment, a new side-to-end anastomosis immediately proximal to the stenosis can be performed using standard techniques as for primary access construction. In most cases, this secondary procedure is much easier than the primary one, because after some months or years of fistula function, artery and vein will be significantly dilated. Although published experience is limited, proximal re-anastomosis is believed to provide better patency rates than PTA ⁴⁷.

4.2.2.    Patch Angioplasty

Less frequently, stenoses of the needling segment of the fistula vein will compromise fistula function. In the case of a stenosed wrist access very proximal re-anastomosis can  result in significant loss of access length. Therefore, in short stenoses, surgical patch angioplasty will be the better alternative ⁹, because the complete access is preserved. Patch angioplasty can also be performed in anastomotic stenoses of prosthetic access.

The vein or graft-to-vein anastomosis is exposed through a longitudinal skin incision and opened longitudinally after clamping  proximal and distal to the stenosis. The incision must begin and end in a vein (graft) segment with “normal” calibre. A segment of the  long saphenous vein or a synthetic patch are used for closure of the venotomy or graft incision.

4.2.3.    Graft Interposition

In stenoses of the needling segment of autogenous  or prosthetic access longer than 4 or 5 cm, interposition grafting is easier and quicker than suturing a long patch ^{50 51}. The access is exposed through two separate incisions proximally and distally to the stenosis. After clamping, the vessel or graft is transected leaving the stenosed segment in place. In autogenous  access the long saphenous vein may serve as interposition graft provided it is of adequate diameter and length. In fistula veins dilated to 6 mm or more and in grafts,  graft interposition will probably give more satisfactory  results. The graft is positioned in a subcutaneous tunnel past the stenotic needling site and sutured end-to-end to both the proximal and distal end of the transected vein or graft.

4.2.4.    Graft Extension

Stenoses of the venous anastomosis of prosthetic access longer than 4 or 5 cm or complete occlusions of the post-anastomotic vein is bridged by graft extension. The graft is exposed close to its venous anastomosis as is the draining vein upward of the occlusion. After clamping and transecting the graft, an end-to-end anastomosis is sutured with the extension graft, which is then tunneled into the incision over the vein. When the new graft has to bridge a joint region to reach the draining vein, great care must be taken to find a position where it is not kinked during joint movement or a ringed graft used ⁵² . The anastomosis to the vein is sutured in an end-to-side fashion after a longitudinal venotomy.

4.2.5.    Surgical Thrombectomy

Simultaneous correction of the underlying access stenosis is an integral part of surgical thrombectomy. Therefore the thrombosed vein or graft is exposed through a skin incision at a location allowing for adequate access to the presumed site of the stenosis. This means that if open surgical correction is planned, the incision is made over or close to the suspected area of stenosis. When an endovascular procedure is performed, the incision is made at reasonable distance from the presumed stenosis to allow for easy handling of the introducer sheath and all necessary interventional equipment. The vein or graft is opened transversely, and the thrombus is extracted with an embolectomy  catheter of adequate diameter. In tortuous or aneurysmal  veins, remaining thrombus can be mobilised and even be expressed by digital massage. Treatment of stenoses is performed with standard surgical or endovascular techniques as described above. Completion on-table angiography is mandatory regardless of whether open surgical or endovascular correction of the access stenosis was performed.

Infection causes 15 to 36 % of all-cause mortality in haemodialysis patients, a 100 to 300 fold higher risk of death caused by sepsis in ESRD patients when compared to the general population ^{53 54}.

Vascular access infections are implicated as the cause of 48% to 73 % of all bacteraemia in haemodialysis patients ⁵³. The majority of bacteraemia is caused by staphylococcal organisms ^{53 55}.  Bacteraemia is associated with high rates of mortality (8 to 25 %) and recurrence (14 to 44 %) and metastatic complications (14 to 44 %, average 25 %) ⁵³, amongst  which are infectious endocarditis, septic arthritis, epidural abscess, septic pulmonary emboli and osteomyelitis. The risk is higher with S. aureus infection ⁵⁶.

5.1.       Infections in Different Types of Vascular Accesses

5.1.1.   Infections Related to Tunneled (Cuffed) Catheters

Bacteraemia is much more frequent in patients with central venous catheters than in patients with A/V fistula, with a relative risk (RR) of 7.6 ⁵⁷. Infection-related deaths are also more frequent in patients with central venous catheters, with a 2.3 RR  in patients with diabetes mellitus and 1.83 RR in patients without diabetes mellitus, compared to patients with A/V fistula ⁵⁸.

In 1988, Schwab et al. inserted eighty tunneled cuffed catheters. They reported only one case of bacteraemia in 4480 catheter-days, corresponding to 0.22 per 1000 catheter-days ⁵⁹. However, randomised studies comparing cuffed and non-cuffed catheters are rare. A  non-randomised study was performed by Jean et al., who followed up 62 double-cuffed catheters without lateral holes and 63 non-cuffed catheters with lateral holes, both inserted with creation of a tunnel. Bacteraemia occurred in 1.3 per 1000 catheter-days in cuffed and 1.08 per 1000 catheter-days in non-cuffed catheters. Similar, local infection rates were higher in cuffed than in non-cuffed catheters (1.3 to 0.77 per 1000 catheter-days respectively) ⁶⁰. Similarly, higher risk for access related bacteraemia per 1000 catheter-days was reported by the Center for Disease Control (Atlanta, USA) for cuffed (2.91) than for non-cuffed catheters (1.6) ⁶¹.

In tunneled cuffed catheters Beathard et al., Marr et al. and Saad reported 3.4, 3.8 and 5.5 catheter related bacteraemia per 1000 catheter-days, respectively

^{56 62 63}. These higher rates compared to those mentioned above may  be explained by different catheter handling procedures. Beathard et al. described a decrease in catheter - related bacteraemia from 4.7 to 1.6 per 1000 catheter-days after changes in the protocol for catheter management ⁶².

Among temporary catheters, femoral catheters are more susceptible to infections than those in thoracic location, and internal jugular have higher infection rates than subclavian ^{64 65}. However, these results do not justify use of the subclavian vein catheterisation, since subclavian vein catheters carry the highest risk of central vein stenosis ⁶⁶.

Colonisation of catheters occurs most frequently through the lumen at the time of connecting the hub to dialysis blood lines, and not by migration of bacteria down the outer surface of the catheter ⁶⁷.

The presence of a biofilm on the inner and outer surface of the catheter may play an important role in catheter-related bacteraemia. Bacteria adhere and become embedded in the glycocalyx of the biofilm making them more resistant to antibiotics than those floating in the circulation ⁶⁸.

Frequent destruction of the endoluminal biofilm by means of a fibrinolytic agent and local instillation of antibiotics left in situ (antiseptic lock) may be an effective means for the prevention of a blood stream infection.

5.1.2.   Infections of the PTFE Graft and the A/V Fistula

PTFE grafts have a high risk of infection beginning at surgical placement, with a 6 % initial 30 days infection rate or even higher in the femoral location. The risk of infection in grafts is much higher than in A/V fistulas ^{69 70}.  Likewise, the relative risk of infection-related deaths is 2.47 for patients with diabetes, and 1.27 for patients without diabetes, compared to patients with an A/V fistula ⁵⁸.

Antibiotic Treatment

Empirical antibiotic treatment must cover both gram positive, responsible for up to 75 % of catheter-related infections, and gram negative organisms. Resistance to antibiotics is also a problem in catheter- related bacteraemia in dialysis patients. About 40 - 75 % of gram positive bacteraemia is methicillin resistant. Some species of S. aureus are not only resistant to methicillin but also show a decreased susceptibility to glycopeptides ⁷¹.

Aiming to cover both gram positive and negative organisms in empirical treatment for suspected infection of haemodialysis vascular access, and, considering the high incidence of MRSA, and the convenient pharmacokinetics of vancomycin in ESRD, it has become common practice to combine vancomycin 1g i.v. (or 20 mg/kg) every 5 to 7 days plus gentamycin 80 mg (1 to 2 mg/kg) every 24 to 48 hours, thus allowing outpatient prolonged parenteral antibiotic treatment. The use of high flux or large surface area dialysers may require vancomycin to be dosed after each treatment to maintain effective blood levels. In this setting a dosing schedule of 500 mg of vancomycin after each dialysis session may be more appropriate ⁷¹.

The empirical use of this combination is still justified in severe infections with bacteraemia, fever or blood pressure instability.

With the emergence of vancomycin-resistant enterococci and the justifiable concern of induction of vancomycin resistant staphylococci, empirical use of vancomycin for the febrile patient on hemodialysis has recently been challenged by several authorities ^{72 73}. The Center of Disease Control and Prevention (CDC) suggested vancomycin to be reserved for b -lactam allergy or for serious infection where b -lactam resistant gram-positive bacteria (MRSA, Staphylococci epidermidis) are likely.

The European Best Practice Guidelines on Haemodialysis recommend the empirical use of methicillin, in order to avoid development of glycopeptide resistance. They recommend vancomycin in settings with increased MRSA. In severely ill patients, a third or fourth generation cephalosporin should be added to cover gram-negative bacteria including Pseudomonas aeruginosa.

In dialysis units with a low rate of MRSA, or when antibiotic sensitivity is available, vancomycin can be safely and effectively substituted by cefazolin 20 mg/kg at the end of each dialysis treatment ⁷⁴.

Most authorities, however, still recommend initial treatment with broad spectrum vancomycin plus aminoglycoside. Rapid conversion to cefazolin regimen or other appropriate antibiotics based on culture and sensitivities is needed, not only to prevent emergence of resistant organisms, but also to avoid ototoxicity.

Metastatic complications are more common with short courses of antimicrobial treatment (2 weeks or less). Therefore, in the case of bacteraemia, four weeks of adequate antibiotic treatment is advocated for S. aureus and a minimum of three weeks for all the other organisms ^{62 75}. The EBPG recommend four weeks of antibiotics  in all cases of bacteraemia.

Blood cultures should be repeated 1 week following the end of therapy to ensure that the infection has been eradicated ⁶².