GUIDELINES [ back to index ]

11. Identification of A/V fistula and graft problems

1. Changes in thrill or bruit
   
2. Oedema
   
3. Fistula
   
4. Pulsatility
   
5. Delivered dialysis dose low
   
6. Low access blood flow
   
7. Negative arterial pressure
   
8. Elevated dynamic venous pressure
   
9. Abnormal static venous pressure/static venous pressure ratio (SPRv)
   
10. Prolonged bleeding time
    
11. Indications for imaging
    
12. Colour-coded duplex-sonography
    
13. Angiography
    
14. Magnetic resonance angiography (MRA)
    
15. Exception: clinically clear thrombosis

• references

*1  Changes in thrill or bruit

• Vascular access flow is regarded as an important indicator of  future patency and can be screened by clinical examination. The patients themselves should either listen for bruit or palpate for thrill and/or pulse daily. Patients with a palpable thrill at the arterial, mid and venous segment of a graft usually have an access blood flow of > 450 ml/min ¹.

• The bruit should be assessed with regard to its continuous character and diastolic component. Shortening of the diastolic part is found in the presence of stenosis. Intensification of the bruit may indicate stenosis or stricture. A significant decrease in bruit after elevation of the arm may be a symptom of reduced arterial inflow, in some patients caused by hypotension without an underlying morphological alteration.

• In cases of graft outflow obstruction, Doppler investigation produces a “water-hammer” like sound. This is not audible in thrombosed grafts.  

*2  Oedema

• Oedema may be an indicator of  infection or venous outflow impairment. The pattern of the oedema and the type of collateralisation can reveal a relatively exact location of the venous obstruction.

• Isolated forearm oedema may occur in distal arteriovenous access with stenosis of the main draining forearm vein.

• A swelling of the total upper extremity (sometimes combined with reddening similar to cellulitis and ulceration) signals a central venous stenosis. Its location may be  in the subclavian vein, caused by previous catheterisation,  or particularly in  the left brachiocephalic (innominate) vein with a retrograde filling of the internal jugular vein, often with drainage via the contralateral internal jugular vein.

• Obstruction of the innominate vein or superior vena cava  may cause breast and face-swelling, sometimes even headaches due to benign intracranial hypertension.

• Any case of oedema requires careful clinical examination, in most cases supplemented by ultrasonography  and other radiological investigations.   

*3  Fistula

• Due to the superficial course of the access vein, thorough clinical examination alone is sufficient to plan the intervention in most cases of A/V fistula stenosis ². In anastomotic venous stenosis, the fibrotic segment of the vein can easily be palpated. Imaging by colour-coded duplex ultrasonography, angiography or MRA is indicated if an additional arterial stenosis is suspected, when arterial pulse at the anastomosis is reduced or lost. Stenosis within the punctured segment of the vein causes a strong pulse in the most peripheral segment of the vein without a murmur, a high-frequency thrill and a weak murmur proximal to the stenosis. Junctional stenosis, located in the superficial vein above the punctured segment and close to its junction with the draining deep vein, causes pulsation of the whole superficial vein and sometimes development of excessive superficial collaterals. On arm elevation, the prestenotic vein will not collapse. [Pre-treatment colour-coded duplex-ultrasonography or fistulography is necessary in junctional stenoses (see *11)].  

*4  Pulsatility

• In access graft stenosis, clinical examination is of limited value. Grafts with a thrill all over are unlikely to have a flow of less than 450 ml/min¹. Palpable graft pulsation, however, is predictive of venous anastomotic stenosis and consequently  reduced flow only in 28 % of cases ¹.

• Pulsatility is also be encountered in stenosis of  A/V fistulas.  

*5  Delivered dialysis dose low

• Decreases in delivered dialysis dose, as measured by Kt/V or URR, may  be indicative of either insufficient blood flow or high recirculation, possibly due to stenosis. However, recirculation is a relatively insensitive parameter, detecting stenosis very late ³.

*6  Low access blood flow

• Progressive stenosis due to intimal hyperplasia, usually at the site of the A/V anastomosis in native A/V fistulas and venous anastomosis in grafts, jeopardises  the blood flow with subsequent thrombotic occlusion. Therefore, it seems clear, that there is need to follow A/V fistulas and grafts in a standardised surveillance program with special emphasis on vascular access flow (Q_a) measurements. Several studies stress the importance of non-invasive or in-line flow measurements and correlate these with the outcome of native A/V fistulas and grafts.

• Thrombosis rates in grafts were 92.8 % and 26.3 % when the access flow was < 800 or > 1600  ml/min, respectively ⁴. The mean flow in grafts, that did not thrombose, compared to grafts, that thrombosed, was 1193 vs 875  ml/min as measured by ultrasound dilution technique (1171 ml/min vs 762 ml/min measured by  Doppler ultrasound). The relative risk was 1.23, 1.67, and 2.39 with a Q_a of 950, 650 and 300  ml/min ⁵. Half of the accesses clotted, when Q_a was less than 750  ml/min, while only 2 out of 27 accesses clotted with Q_a > 750  ml/min ⁶. Depner documentated high failure rate from stenosis and thrombosis in accesses with Q_a < 600  ml/min, while those with flows > 800  ml/min had few failures ⁷.

• With a cut-off value of 600  ml/min the rate of thrombosis was 7.2 for the patients having a Q_a < 600  ml/min compared to those patients having a Q_a > 600  ml/min⁸. Grafts with Q_a < 500  ml/min had a relative risk of clotting within 8 weeks of 3.12, compared to grafts with Q_a of 1100 – 1400  ml/min ⁹.

• Tonelli et al. investigated 177 native A/V fistulas ¹⁰. They used ultrasound dilution technique to determine blood flow Q_a. In fistulas with a blood flow rate < 500  ml/min, a subclinical stenosis was found in 81 % of cases (defined as a > 50 % reduction of vessel diameter determined by angiography). Thus, the authors suggested  use of a blood flow rate of 500 ml/min as a cut off value, corresponding to the recommendation of the Canadian Society of Nephrology ¹¹.

• As these different results indicate, it is difficult to determine a fixed cut-off value, because the blood flow rates of accesses that thrombose overlap with the blood flow rates of the accesses that did not thrombose. In addition, native A/V fistulas have in general lower blood flow rates than grafts, despite a lower risk of thrombosis ¹². Wang et al. found that grafts had a 5.6 greater risk on thrombosis compared to native A/V fistulas ⁹.

• A single measurement of reduced intra-access blood flow appears to be a relatively poor predictor for thrombosis ¹³ the percentage of decrease in flow did, not correlate with the likelihood of access failure.

• Serial measurements with a documented decrease of blood flow over time seems more accurate than single measurements. In a prospective study, Neyra ¹⁴ observed a 13.6-fold increase in the relative risk of thrombosis for accesses with more than 35 % decrease in vascular access blood flow when compared to those accesses with no changes in blood flow.

• Nevertheless, the DOQI guideline consider an absolute Q_a of < 600  ml/min or a decline in Q_a of > 25 % in vascular accesses with a Q_a > 1,000  ml/min between two monthly measurements as indicative of an enhanced risk of thrombosis ¹⁵.

• Gallego et al. concluded, that early detection of stenosis, indicated by increased venous pressure or by difficulties in cannulation, and subsequent treatment of the stenosis can prevent thrombosis ¹⁶.

• Schwab et al. initiated an active intervention program in grafts and A/V fistulas with flows < 600  ml/min or a decrease in flow of > 20 % compared to the last measurement. The overall thrombosis rate for grafts and A/V fistulas declined from 25 to 16 %. For A/V fistulas alone, the thrombosis rate decreased from 16 to 7 % ¹⁷.

• In a meta-analysis of 12 studies, Paulson et al. found a wide range of Q_a values that were used as indicators for impending failure ¹³. In combining these studies, the predictive value of flow measurement for thrombosis was 0.70 and for access failure 0.76. Serial Q_a measurements are a better indicator than a single measurement.

*7  Negative arterial pressure

• With a standard 15-gauge needle set, the development of excessive negative pressures below –200 mm Hg in the arterial blood line at a blood flow rate of 400  ml/min indicates a failure of the fistula to provide  the flow demanded by the blood pump. This means, that the fistula-flow from the anastomosis is inadequate ^{18 19}.

*8  Elevated dynamic venous pressure

• Dynamic venous pressure is considered abnormal when it is greater than 125 mmHg with a blood flow of 200  ml/min  with the use of a 15 gauge needle on three repeated treatments ²⁰ (for details: see appendix). Bosman et al. measured the venous drip chamber pressure, while the vascular access of the patient was placed at the same level as the venous drip chamber (thus avoiding false measurements due to differences in height between the drip chamber and the vascular access). Venous drip chamber pressure did not discriminate between patients with graft flow > or < 600  ml/min ⁸.

• Access flow (see *6) and static venous pressure (see *9) are more sensitive indicators of access dysfunction then dynamic venous pressure ^{8 21 22}.

*9  Abnormal static venous pressure / static venous pressure ratio (SPR_v)

• Static venous pressure (SVP) is measured with the blood flow stopped (for procedure see appendix II). It should be normalised to mean arterial blood pressure (MAP), thus yielding the quotient SVP/MAP, the “Static venous pressure ratio” (SPR_v). However, intra-access venous pressure is not identical with venous drip chamber pressure but must be corrected for the differences in height between the drip chamber and the access. A vertical difference of 1 cm corresponds to a pressure change of approximately 0.75 mm Hg ²³ (see appendix for calculation of SVP).

• Static venous pressure ratios in grafts are abnormal, if SPR_v > 0.5 or if an increase of 0.25 from baseline is seen within one month, for A/V fistulas, the threshold is SPR_v > 0.35 ^{21 24}. Besarab et al. found a sensitivity of 91 % and 48 % and a specificity of 86 % and 100 % in grafts and A/V fistulas, respectively, for 50 % luminal diameter reduction with a SVP/MAP > 0.4 ²¹. Static venous pressure is more effective than dynamic venous pressure for monitoring grafts ²⁵.

• For A/V fistulas, venous pressure is less effective in detecting stenosis, as collaterals permit blood-outflow without dramatic increase in venous pressure. In addition, stenosis in native A/V fistulas most commonly occur in the inflow or in the body of the fistula, thus the venous needle is usually placed after (or downstream of) the stenosis. In contrast, stenosis in grafts are frequently located in the outflow or at the venous anastomosis, thus the needle is placed before the stenosis. In the latter, the intra-access pressure increases and can be detected ²⁶.

*10 Prolonged bleeding time

• Measurement and documentation of bleeding time after needle withdrawal (without excessive anticoagulation) has been mentioned as a further method (but not method of choice) to detect graft or fistula stenosis ¹².

• Prolonged bleeding post dialysis for more than 10 minutes or a change from the patient’s baseline in the absence of a change in anticoagulation may indicate outflow stenosis particularly in PTFE grafts, but can also occur in skin atrophy. Repeated prolonged bleeding is an indication for evaluation of the vascular access, to rule out venous outflow stenosis.

• Bleeding after removal of the needle is also dependent on the needle size and the degree of anticoagulation ²⁷.  

*11  Indications for imaging

• In cases of A/V fistula thrombosis, clinical examination may reveal the presence and location of an underlying stenosis in most patients.

• When a stenosis of an A/V fistula or graft is suspected, adequate and timely assessment should be performed to initiate  early treatment before thrombosis ²⁸. The decision whether clinical examination alone is sufficient or additional ultrasound examination must be performed before stenosis treatment, depends on local custom and practice.  In cases of endovascular treatment, pre-, intra-, and post-operative angiography take place as  a matter of course.  When open surgical revision is undertaken, completion angiography should also be performed whenever possible. Purely diagnostic angiography without concomitant treatment should be avoided in order to reduce inconvenience for the patient and treatment delay.

• Once thrombosis has occurred, immediate surgical or interventional radiological clot removal is necessary to allow for haemodialysis through  the arterio-venous access without the need for central venous catheter insertion. Correction of the underlying stenosis is an integral part of any declotting procedure.  

*12  Colour-coded duplex-sonography

• Whenever stenosis is suspected, colour-coded duplex ultrasonography can be performed to locate and to quantify the stenosis ^{29 30}. Colour-coded duplex ultrasonography is able to replace angiography except for hand arteries and central veins ³¹.

• In cases of low flow or clinically obvious tight stenosis, duplex examination should be performed only if it does not delay PTA or surgical revision.

• Angiography is not necessary if colour-coded duplex-ultrasonography shows evidence of an isolated stenosis close to the anastomosis in forearm fistulas amenable to surgical revision by proximal re-anastomosis.

• Colour-coded duplex-ultrasonography can be helpful  in defining  the extent of thrombosis.

• Colour-coded duplex examination is especially valuable in detecting stenoses and flow measurements in immature fistulas of predialysis patients in whom iodine injection  ought to be avoided.

• An Italian team has reported  that it perform graft dilatations only under colour Doppler guidance ³².  

*13  Angiography

• Isolated diagnostic angiography without concomitant dilatation or surgical revision must be avoided. Angiography is performed before, during and after dilatation or percutaneous declotting and after surgical thrombectomy in order to guide the treatment and  document the final results.  

*14  Magnetic resonance angiography (MRA)

• When central (mediastinal) vein stenosis or occlusion is suspected, angiography of the complete venous drainage up to the cavo-atrial junction is mandatory, because only the lateral half of the subclavian vein can be visualised by ultrasound ³⁰. Magnetic resonance angiography of the central chest veins is accurate and even superior to contrast venography, which fails to show all patent thoracic vessels ^{33 34}. However, it is not yet, and may never be, possible to perform recanalisation procedures under MRA guidance.  

*15  Exception: Clinically clear thrombosis

• Native A/V fistula and graft thrombosis must be treated without preceding diagnostic work-up. The underlying stenosis must be demonstrated (and treated) in combination with surgical or interventional clot removal.

• See: Management of A/V fistula thrombosis and/or management of graft thrombosis for details.