Hemodynamics of Distal Revascularization- Interval Ligation

Transcription

1 Hemodynamics of Distal Revascularization- Interval Ligation Karl A. Illig, MD, Scott Surowiec, MS, Cynthia IC. Shortell, MD, Mark G. Davies, MD, Jeffrey M. Rhodes, MD, and Richard M. Green, MD, Rochester, New York Distal revascularization-interval ligation (DRIL) empirically corrects steal after arteriovenous fistula (AVF) creation in most cases, but because there is no topologic alteration in anatomy, it is unclear as to why it is effective. To explore this issue, nine symptomatic patients underwent intravascular pressure and flow measurements before and after DRIL following upper arm autologous AVFs. Mean pre-dril systolic pressure (mmhg; mean * SD) in the proximal brachial artery (PROX) was 102 * 17, while that at the AV anastomosis (AV ANAST) was 47 * 38 (p < ). Flow (mumin) distal to AV ANAST was retrograde with the fistula open (-21 * 64) but became antegrade (58 f 29; p < 0.03) with occlusion of the fistula. Following DRIL, pressures at both PROX and AV ANAST sites did not change (104 * 24 and 51 * 43, respectively). However, pressure at the point at which the blood flow split to supply the hand or the fistula, now PROX, increased from 47 * 38 (pre-dril AV ANAST) to 104 f 24 (p < ). Pressure in the brachial artery distal to the ligature increased to 104 f 27 (p < O.OOOl), flow at this point (to the hand) became antegrade (51 * 39; p < 0.03), and occlusion of the fistula did not significantly change pressure at this site. We hypothesize that improvement in hand perfusion following DRlL is due to a higher pressure at the point at which the blood flow splits to supply both hand and fistula (pre-dril: AV ANAST; post-dril: PROX), allowing antegrade flow down the new bypass to the lower pressure forearm. This increased pressure must be due to the increased resistance of the fistula created by interposing the arterial segment between the original AV ANAST and new PROX ANAST. As such, DRlL is schematically equivalent to banding, but resistance is increased in a fashion that is physiologically and empirically acceptable. INTRODUCTION Steal is a significant problem following arteriovenous fistula (AVF) creation, leading to pain, loss of distal function, and tissue death. Estimates of its prevalence range from 1 to 20%, depending on definition and severity, and it may be more common after AV grafts (in contrast to fistulas). Operation is estimated to be required in only 1 to 6% of True steal, in which blood that has reached the tissue distal to the AV anastomosis via collaterals then flows retrograde into the fistula instead of the Division of Vascular Surgery, University of Rochester Medical Center, Rochester, NY, USA. Presented at the Twenty-ninth Annual Meeting of the Peripheral Vascular Surgery Society, Anaheim, CA, June 4-5, Correspondence to: Karl A. Illig, MD, University of Rochester Medical Center, Division of Vascular Surgery, 601 Elmwood Avenue, Box 652, Rochester. NY, 14642, USA, karllillig@urmc. rochester.edu Ann Vasc Surg 2005; 19: DO!: /~ y 0 Annals of Vascular Surgery Inc. Published online: March 14, 2005 tissues where it is needed (Fig. la), is subtly different from the situation in which not enough blood can be provided to supply both the fistula and the distal tissues (e.g., because of a proximal stenosis or distal occlusive disease), but apart from a clear proximal lesion has been difficult to distinguish clinically. Steal itself (the actual retrograde flow distal to the AV anastomosis) can be cured by simple ligation (interval ligation [IL]) of this artery (Fig. 1B). This may work well in situations where the distal tissue is still supplied by a native artery (i.e., a radiocephalic fistula), but the problem of distal ischemia remains if the fistula is based on the brachial artery. In this situation, residual ischemia has been treated by bypass from a more proximal artery to the artery distal to the ligature (distal revascularization [DR]; Fig. 1C). This procedure (DRIL) has been shown to significantly improve clinical status and hemodynamics, 5-9 and is increasingly accepted for treatment of steal for brachial- and more proximally based artery fistulae. When analyzed, however, a problem arises: the topologies of the original AVF and DRIL are no 199

2 200 Illig et al. Annals of Vascular Surgery A B C 2 supplled by collaterals Steal cured and hand well perfused \ Fig. 1. Schematic of arteriovenous fistula (AVF). A Original situation with steal (collaterals supplying forearm not shown). B Interval ligation, correcting steal, al- though distal ischemia persists. C Completed distal revascularization-interval ligation (DRIL). different (Fig. 2). In other words, there is still a topologically identical AVF, just one that arises from a more proximal point in the arm (i.e., the new proximal anastomosis of the distal revascularization component). We hypothesized that the clinical benefit seen after DRIL must be due to the interposition of the arterial segment added between the original AV anastomosis and the new proximal anastomosis of the bypass. In other words, by adding the resistance of this arterial segment to the fistula, pressure at the point at which blood flow splits to supply the fistula and the hand is increased, leading to antegrade flow down the bypass to the distal tissues. METHODS Nine patients undergoing DRIL for symptomatic steal were prospectively studied. Several patients had undergone fistulography, but all nine underwent conventional transfemoral arteriography prior to DRIL. Angiography was performed to precisely delineate anatomy proximal and distal to the arteriovenous anastomosis and also to document improvement in perfusion with AVF compression (in several cases actual retrograde flow was seen in the brachial artery distal to the anastomosis on delayed images). We did not use angiographic findings to determine the site of the proximal anastomosis. After operative exposure of the proximal brachial artery near the axilla (PROX) and the original AV anastomosis (AV ANAST), measurements were recorded as below. The artery was then ligated just distal to the AV ANAST, bypass constructed from PROX to the artery distal to this, and measurements again recorded (Fig. 3). Data collected included pressures, flows, and velocities. Pressures were measured by direct puncture using a 23 g needle attached to a pressure transducer at the head of the table. The system was zeroed and after puncture allowed to equilibrate for 10 sec before mean, systolic, and diastolic pressures were recorded. Pressures at the AV ANAST were measured as close to the suture line as possible for standardization. Flows were directly measured with a 4-mm directional coronary flow probe and flowmeter previously calibrated for human arterial flow (CardioThoracic Systems, Inc., Portola Valley, CA). These probes can be reused sparingly but must be disposed of after a limited number of uses to ensure maintenance of calibration. Finally, maximal velocities were measured with conventional intraoperative duplex imaging (ATL HDI 5000; Phillips Medical Systems, Bothell, WA). While directionality was usually possible to assess with duplex, the amount of turbulence occasionally made this difficult to assess, and direction of flow for purposes of this study was solely determined by the flowmeter. Velocities themselves were not found useful for this analysis and are not further reported. Data were recorded and analyzed using Statview (SAS Institute Inc., Cary, NC). Changes at each point were compared be means of paired t-tests, whereas overall means were otherwise compared using unpaired t-tests. Unless otherwise

3 Vol. 19, No. 2, 2005 Hemodynamirs of DRIL 201 DRIL ACTUAL ANATOMY TOPOLOGIC ANATOMY PROX AV ANAST DlST > Fistula interposed artery Vein bypass Fig. 2. Original AVF and final URTL, showing identical topology. Labels defined in text (see also Fig. 4). Left Original AVF. Right Arm following DRIL. A AVF B DRIL AV ANAST \\ AVANAST PROX (102) FOREARM (67) PROX (104) DlST (67) FLOW = -2lcdmin ". 1 (RETROGRADE) I DlST Fig. 3. Schematic of measurement sites used in thi\ 51ridy and systolic preswres at each site. A\teri\ks (DIST and FOREARM pressures and DIST flow) denote significant differrriccs. A Original AVF. B Arm following DRIL. noted, pressures given are systoiic and arc ex- RESULTS pressed in mmhg, and flows (negative sign indicating retrograde direction) are expressed in ml/ Between December 2002 and November 2003, min. nine patients presented to our service with 5ymp-

4 ~ ~~ 202 Iliig et ai. Annals of Vascular Surgery Table I. Details of operation and outcome Final clinical Interval to Follow-up outcome, including Age DRIL Short-term interval status at last Patient (years) Fistula s ymp t omsa (months)b outcome (months) follow-up AR 78 WB 50 JW 78 CW 72 EP 77 MM 49 LD 70 JW 62 WJ 73 Brachiocephalic P, gangrene 8 Brachiocephalic P, weakness 1 Radiocephalic' P 15 Basilic P 4 transposition Basilic P, ulcer 5 transposition Brachiocephalic P, weakness 50 Brachiocephalic P, acute ischemia 0 Ulnar-cephalicc P, ulcer 13 Brachiocephalic P, ulcer 9 Pain relief, healed 12 amputation Residual pain 13 Pain relief 2 Pain relief 6 Pain relief, healing 1 ulcer Occluded bypass, 9 redo, residual pain Residual pain 5 Pain relief, 6 healed ulcer Pain relief, 3 healed ulcer Pain relief, healed amputation Residual pain despite improved hemodynamics (feels better) Pain relief, died (cardiac) Pain relief, died (cardiac) Pain relief, died (MVA) Residual pain despite improved hemodynamics (no clinical benefit) Residual pain despite improved hernodynamics (feels better) Pain relief, healed ulcer Pain relief, healed ulcer MVA, motor vehicle accident. "All patients had continuous distal pain (P). 'Interval (in months) from original AV and DRIL. 'Both forearm-based fistulae included in this study were in patients in whom the other forearm artery was occluded. tomatic steal following brachial artery-based AVF (7) or a radial or ulnar artery-based AVF in the setting of occlusion of the other forearm artery (one each). Patients ranged in age from 49 to 78 (mean, 67) years. Four patients had antecubital fossa brachiocephalic fistulas, two had basilic transpositions, one had a cephalic-to-brachial loop fistula with vein, and two had fistulae based off of a single patent forearm vessel. All had highly symptomatic steal: all nine had continuous pain, four had active ulceration or gangrene, two had motor dysfunction, and one was acutely ischemic immediately following AVF creation (operated on the next day). Diagnosis of steal was made by the combination of (a) a patent, high-flow AVF; (b) no clinical problems prior to creation of the fistula; (c) flat digital waveforms significantly increased by temporary occlusion of the fistula; (d) retrograde flow suspected by duplex examination distal to the open fistula, again corrected by occlusion; and (e) no proximal obstructive lesion by angiography. Interval from AVF creation to DRIL was surpris- ingly long, with a mean of 1 year (ranging from 0 to 50 months). This may reflect (in all but the patient who presented acutely) the time needed for dilation of the fistula and true steal to be created, but also reflects delay in diagnosis and treatment options offered to the patient (Table I). Seven DRIL procedures were performed with saphenous vein and two with polytetrafluoroethylene (PTFE). No morbidity was encountered apart from superficial wound infection (arm, 3, vein harvest site is], 2), although one prosthetic bypass occluded at 3 months and required redo with lesser saphenous vein (redo not included in this series). During follow-up, six of the nine patients experienced significant acute improvement, and the remaining three had improvement in forearm pressures despite reporting no pain relief. Ulcers healed or were healing at death in all affected patients. Three of the nine patients (all with pain relief) died within 6 months of DRIL from causes unrelated to the DRIL itself. The three surviving patients with symptomatic relief have had sus-

5 ~~ Vol. 19, No. 2, 2005 Hemodynamics of DRIL 203 Table 11. Original (pre-dril) hemodynamics Status of AVF Proximal brachial artery AV anastomosis Artery distal to AV anastomosis Fistula Flow Open 574 & rf: k 114 Occluded 89 f SBP Open 102 f k 38 67* 15 Occluded i If: 15a Flow and pressure measurements in the original AVF, during steal situation, before DRIL performed (see Fig. 3A). Flow is measured as mlimin, with a minus sign if retrograde, and systolic blood pressure (SBP) as mmhg. All measurements are mean f standard deviation. ap < 0.05 by paired t-testing (comparing identical measurements with AVF open and fully compressed). tained benefit (mean follow-up, 7 months) whereas two of the three with continued pain feel improved and consider their quality of life acceptable; all three (including the patient with no symptomatic improvement) exhibit dramatically improved hemodynamics (Table I). All AVFs remained open at death or last follow-up. Interestingly, although all bypasses (apart from the occlusion reported above) have remained open, all surviving patients have experienced a slow decline in their forearm pressures. Hemodynamics Sites of measurement and results in the pre-dril state, with active steal, and the final, post-dril situation are shown in Figure 3 and indicated tables. Before DRIL (Fig. 3A), forearm cuff pressures with the fistula open were 67/41 (mean systolic/ diastolic, in mmhg) but increased to 127/52 with the fistula occluded. Flow in the artery distal to the fistula (DISTAL) was retrograde (-21 * 64 ml/min) but became antegrade (58 f 29) with the fistula occluded. Systolic pressure at the AVF anastomosis itself (AV ANAST) was very low at rest (47 f 38) but increased to systemic levels with AVF occlusion. Finally, systolic pressure in the brachial artery distal to the AVF (DISTAL) was also very low at rest (67 f 15), but higher than that at the AV ANAST, explaining the retrograde flow; DISTAL pressure expectedly rose with fistula occlusion (Table 11). All of these changes were highly significant unless otherwise specified, indicating that distal tissue perfusion had a high degree of dependence on the patency of the fistula in the pre-dril, actively stealing state. By contrast, hemodynamics following DRIL (Fig. 3B) were very different. Forearm cuff pres- sures remained somewhat AVF-dependent (128/67 open, 149/77 occluded), but the magnitude of increase was much less than before DRIL. Flow in the proximal brachial artery (PROX) remained high and fistula-dependent, but flow in the artery distal to the AVF (DISTAL) now became antegrade (51* 39) and did not change with AVF occlusion (Table 111). Pressure at this location showed a similar pattern (DISTAL: 104 f 27 open, 122 * 26 oc- cluded, not significant [NS], indicating that perfusion in this segment was much less dependent on the status of the fistula than prior to DRIL. Pressures at the proximal brachial artery (PROX: 104 * 24 open, 126 f 34 occluded) and at the AV anastomosis (AV ANAST: 51 k 43 open, 118 * 30 occluded) were very similar to their pre-dril counterparts, indicating no hemodynamic changes produced by DRIL (Table 111). Comparison of certain critical corresponding points before and after DRIL (all with AVF open) is revealing (Table IV). First, systolic pressure at the point at which blood flow splits to provide flow to the fistula or arm increased significantly from 47 * 38 (AVF pre-dril) to 104 * 24 (PROX post- DRIL; p < O.OOOl), Second, systolic pressures in the artery just distal to the AVF (and ligature post- DRIL) increased significantly from 67 * 15 (pre- DRIL) to 104 * 27 (post-dril; p < O.OOOl), and flow at this point went from retrograde (-2 1 * 64 pre-dril) to antegrade (51 f 39 post-dril; p < 0.03). Finally, forearm cuff blood pressures increased from 67/41 pre-dril to post-dril (p < 0.001) (Table IV). Finally, although the significance of this value is as yet unknown, the resistance of the interposed segment is calculable (resistance = AV/Flow). Mean length interposed (PROX to AV ANAST) was 19.6 & 4.2 (range, 15-24) cm, leading to a mean resistance of * mmhg-min/ml. More experience will be needed to understand the clini-

6 ~~ ~~ 204 illis et al. Annals of Vascular Surgery Table 111. Final (post-dril) hemodynamics Status of AVF Proximal brachial artery AV anastomosis Artery distal to AV anastomosis Fistula Flow Open Occluded 84 * 62a SBP Open 104 * 24 Occluded f * * 30a 122 k * Flow and pressure measurements in the final situation after DRIL finished (see Fig. 3B). Flow is measured as ml/min, with a minus sign if retrograde, and SBP as mmhg. All measurements are mean * standard deviation. ap < 0.05 by paired t-testing (comparing identical measurements with AVF open and fully compressed). Table IV. Pre- and post-dril hemodynamic comparison Split point Artery distal to AVF/ligature Site Pressure Pressure Flow Forearm blood pressure Pre-DRIL AVF 47 k k /41 Post-DRIL PROX 104 * * k < < ~0.03 <o.oo 1 Flow (ml/min; minus sign if retrograde) and pressure (mmhg) measurements, as indicated, comparing certain points before (pre) and after (post) DRIL (i.e., comparing Fig. 3A with 3B). Split point refers to the point at which the blood flow splits to provide perfusion of the fistula or arm; note that this point is changed from AVF prior to DRIL (Fig. 3A) to PROX following DRIL (Fig. 3B). All measurements are mean * standard deviation (or systolic/diastolic in the case of cuff pressures). Statistical comparison by means of unpaired t-tests. cal significance of this calculation and its role in decision making. DISCUSSION As discussed above, steal is a significant problem following AV access for dialysis. Its prevalence is difficult to pin down because of widely varying definitions and types of fistulas (and grafts) examined. For example, it has been shown that actual retrograde flow (steal) is present after as many as 80% of brachial artery-based fistulas. Symptoms, however, seem to occur in 10% or less of patients, while symptoms of enough severity to warrant repair seem to occur in only 1-6% of patients.2*6-8 Steal may be more common after prosthetic grafts (in contrast to autologous fistulas) and is clearly more common after upper arm artery-based ac~ess.l-~ DRIL has clearly been shown to improve distal perfusion, allowing relief of ischemic symptoms and maintenance of fistula f~nction.~ ~ The major conceptual problem, invariably noted during discussion following presentation of these data publicly, is that the topology of the fistula and limb vessels is not altered (Fig. 2). Therefore it remains unclear why this procedure works as well as it does. If it is granted that topology does not change and assumed that DRIL does not affect inflow proximal to it, it must be accepted that the only alteration is the addition of an arterial segment (the brachial artery between the proximal portion of the new bypass [PROX] and the original AV anastomosis [AV ANAST]; Fig. 3). We hypothesized that this segment acts to increase the resistance of the fistula enough to allow a greater volume of blood to reach the arm, and our data suggest that this is correct. The AV anastomosis prior to DRIL and PROX after DRIL are the points at which the blood flow splits to supply both fistula and hand. Forearm pressure prior to DRIL was higher than that at the anastomosis, which explains why flow preferentially passes into the lower-resistance AV fistula rather than to the forearm. In addition, blood that has reached the forearm by collaterals will also flow retrograde, away from the hand, down the pressure gradient to the AV anastomosis. In other words, DRIL does not change the pressure at the AV anastomosis or proximal brachial artery-there is still a pressure gradient from the proximal brachial artery to AV anastomosis (Fig. 3). However, what it seems to do is increase the pressure at the split point to a level higher than that in the forearm, thus creating a pressure gradient that favors blood

7 Vol. 19, No. 2, 2005 Hemodynamics of DRIL 205 DRIL ACTUAL ANATOMY TOPOLOGJC ANATOMY STEAL NO STEAL LOW RESISTANCE (Fistula) HIGH RESISTANCE (Artery) LOW RESISTANCE (Vein) Fig. 4. A copy of the schematic views of the AVF, with steal, and final DRIL, as originally shown in Figure 2. The segments have been relabeled to emphasize changes in resistance and resultant change in the direction of flow in the forearm. Left Original AVF. Right Arm following DRIL. flow to the hand. In turn, this pressure is not increased by altering local hemodynamics but by altering the location of the split point (originally the AV anastomosis but now the proximal anastomosis of the bypass; Fig. 3). In turn, the increase in pressure at the newly proximal AV anastomosis means that we have increased the resistance of the anastomosis by adding an additional arterial segment to it. In other words, DRIL seems to be hemodynamically analogous to banding, but seems to increase resistance in a physiologically and empirically acceptable fashion. It should be noted that pressures at both end of the distal revascularization were equal (both 104 mmhg; Table 111). This implies that the resistance of the saphenous vein bypass (seven of the nine conduits) is extremely low. If we hypothesize that autologous arteries have an intrinsically high resistance while saphenous veins have very low resistance, the mechanism of effect of this procedure becomes very clear in a schematic sense (Fig. 4). It is not surprising that perfusion to the distal arm can be improved without degradation of fistula performance. At the range of blood flow needed to supply the arm (50 ml/min in our experience) versus within these fistulas (4-600 ml/ min), a small alteration in extremity perfusion will not decrease fistula flow enough to cause problems. These results and our interpretation raise many interesting questions. For example, is this increased resistance a passive property of length or an active property of the artery itself? To what degree does the length itself matter (the University of Arizona group s experience seems to suggest that it may not7)? What is the significance of our resistance calculation, and can this be used to

8 206 lllig et a[. Annals of Vascular Surgery predict the optimal configuration in an priori sense? Why do clinical results seem to degrade over time (at least in our experience)-is this due to dilatation of and loss of resistance within the interposed arterial segment, or some other alteration in either proximal perfusion, the venous portion of the fistula, or the bypass itself? Finally, where do the original collaterals supplying the area arise from, especially in relation to the new proximal anastomosis? A very interesting question, in fact, is whether the distal bypass component is needed at all. In other words, after ligation of the artery, is collateral flow adequate to supply the arm as is? This issue was recently explored by Balaji and colleagues in a case report describing a patient with steal following an upper arm bridge graft. Their plan was conventional DRIL, but both sapheni were found to be unusable. Pressure measurements documented expected hemodynamics (brachial artery pressure just distal to the AV anastomosis, 22 mmhg). After interval ligation alone, however, pressure distal to the ligature rose to 60 mmhg, they halted the procedure at this step, and the patient subsequently did very well with relief of ischemic symptoms and maintenance of access. A similar improvement in distal artery pressure (prior to bypass) was also noted by Schanzer s group. 5r8 We have not had the courage to eliminate distal revascularization yet, but also document a modest increase in pressure distally after ligation. Whether simply eliminating retrograde flow and increasing forearm pressure to only this level will be adequate is unknown. It should be noted that the opposite technique-bypass alone without ligation-has also been suggested (see discussion following knox et al. ), although no results have been published. The benefit of this method is to decrease the resistance to the hand (by substituting vein for artery). The site of ligation is variable in different authors reports. We ligate the brachial artery just distal to the AV anastomosis, proximal to its bifurcation, to eliminate steal altogether, as do Balaji et al. Schanzer et a1.58 and Knox and colleagues from the University of Arizona in their more recent report.7 However, an interesting variation, not specifically commented on, is shown in an earlier article by the Arizona group.6 Although they state that they perform DRIL according to Schanzer, their Figure 1 shows ligation of and bypass to the ulnar, rather than brachial artery, with the radial (and interosseous) arteries still in continuity to the Low-pressure AV anastomosis. It is unclear whether this will be of benefit, but is an interesting variation. Finally, Minion and colleagues (unpublished data) describe three patients in whom steal was treated by essentially reimplanting the fistula into a more distal artery. Whether this is effective because of the increase in arterial resistance proximal to the fistula or through other mechanism is as yet unknown. A fundamental question is that of differentiation of steal from ischemia due to either inadequate total inflow versus distal vessel disease. How do these really differ in a hemodynamic sense, how can they be predicted, and what is the optimal treatment for each? These are questions that should be answered by prospective data collection in a much larger group of patients than have been previously studied. CONCLUSIONS Our data demonstrate that the clinical benefit seen after DRIL is matched by hernodynamic improvement (increase in distal pressure and conversion of forearm flow from retrograde to antegrade), and that this improvement is created by the interposition of an arterial segment at the beginning of the fistula, increasing the resistance of the fistula while decreasing that of the path to the arm (Fig. 4). This implies that the length between the beginning of the new bypass and existing AV anastomosis will potentially determine success, and suggests a mechanism to explain the degradation of hemodynamics seen over time. REFERENCES 1 Nicholson ML, Murphy GJ. Surgical considerations in vascular access In: Conlon PJ, Schwab SJ, Nicholson ML, (eds). Hemodialysis Vascular Access: Practice and Problems. Oxford: Oxford University Press, 2000, pp Wilson SE. Complications of vascular access procedures In: Wlison SE, Vascular Access: Principles and Practice. (3rd ed)st. Louis: Mosby, 1996, pp Zibari GB, Rohr MS, Landreneau MD, et al.complications from permanent hemodialysis vascular access. Surgery 1988; Morsy AH, Kulbaski M, Chen C, Isiklar H, Lumsden AB. Incidence and characteristics of patients with hand ischemia after a hemodialysis access procedure. J Surg Res 1998;74: Schanzer H, Schwartz M, Harrington E, Haiinov M. Treatment of ischemia due to steal by arteriovenous fistula with distal artery ligation and revascularization. J Vasc Surg 1988;7: Berman SS, Gentile AT, Glickman MH, et al.dista1 revascularization-interval ligation for limb salvage and maintenance of dialysis access in ischemic steal syndrome. J Vasc Surg 1997;2 6: Knox RC, Berman SS, Hughes JD, Gentile AT, Mills JL. Distal revascularization-interval ligation: a durable and

9 Vol. 19, No. 2, 2005 Hernodynarnics of DRIL 207 effective treatment for ischemic steal syndrome after hemodialysis access. J Vasc Surg 2002;36: Schanzer H, Skladany M, Haimov M. Treatment of angioaccess-induced ischemia by revascularization. J Vasc Surg 1992;16: Sessa C, Pecher M, Maurizi-Balzan J, et al.critical hand ischemia after angioaccess surgery: diagnosis and treatment. Ann Vasc Surg 2000;14: Kwun ICB, Schanzer H, Finkler N, Haimov M, Burrows L. Hemodynamic evaluation of angioaccess procedures for hemodialysis. Vasc Surg 1979;13: Balaji S, Evans JM, Roberts DE, Gibbons CP. Treatment of steal syndrome complicating a proximal arteriovenous bridge graft fistula by simple distal artery ligation without revascularization using intraoperative pressure measurements. Ann Vasc Surg 2003; 17: