Modified MILLER banding procedure for managing high-flow access and dialysis-associated steal syndrome

Introduction

For effective hemodialysis, it is important to keep an adequate vascular access (VA) blood flow rate. However, creation of either an arteriovenous fistula (AVF) or an arteriovenous graft (AVG) results in a decreased peripheral resistance and then increased cardiac output (CO). The increased CO may induce structural and functional cardiac changes, including development of eccentric left ventricular hypertrophy (LVH) (1), an important factor for congestive heart failure. Furthermore, a high access blood flow (Qa) is postulated to increase CO and cause high-output cardiac failure (2-3-4). There are currently few criteria defining a high-flow AVF. Zanow et al (4) defined high flow as a flow volume more than 800 ml/min for an AVF and more than 1200 ml/min for an AVG. Schneider et al (5) defined a hyperfunctioning fistula when fistula flow was more than 1000 mL/min. Basile et al (6) showed the high predictive power for high-output cardiac failure of Qa cutoff values more than 2.0 l/min.

Dialysis-associated steal syndrome is also a serious complication. Steal syndrome is defined as signs and symptoms (pain, coldness, cyanosis, and necrosis) produced by an access resulting from the diversion of arterial blood flow into the fistula (7). The incidence of symptomatic ischemia has been reported to vary from 2% to 8% of the hemodialysis population (8-9-10-11). For patients with ischemic pain at rest, ulceration, necrosis, and gangrene are indications for interventional treatment (7, 10, 11).

To treat these complications, a flow reduction technique is necessary. The minimally invasive limited ligation endoluminal-assisted revision (MILLER) banding procedure was developed by Goel et al (12) and Miller et al (13) to treat dialysis-associated steal syndrome. Through a small skin incision, the VA is banded by using an endoluminal balloon to achieve and standardize the desired inflow reduction (13). Miller et al (13) extended the use of the MILLER procedure to also treat high-flow VA.

We adopted the MILLER banding procedure with some modifications to control the high blood flow and steal syndrome occurring during VA procedures and retrospectively assessed the outcome.

Materials and Methods

Between January 2011 and October 2012, seven patients underwent maintenance hemodialysis with high-flow access and five patients with dialysis-associated steal syndrome (with pain, coldness, or cyanosis) were referred from our hospital or affiliated outpatient clinics. Qa was measured by either a VA surgeon or a trained clinical engineer team by duplex Doppler ultrasound (DUS) using a Prosound Alpha 6 (Hitachi Aloka Medical, Tokyo, Japan) equipped with a broadband linear transducer (UST-5413, 411 MHz). Qa was determined by measuring brachial artery flow volume at our outpatient clinic, with high flow defined as at least 1500 ml/min. A diagnosis of steal syndrome was made by physical examination. Pale cold fingers with pain at rest that was exacerbated during dialysis or exercise was an inclusion criterion for banding. Banding was selected if the nephrologist or surgeon decided that a flow-reduction procedure was necessary.

Banding procedure

We followed the procedure of Miller et al (13) with modifications. Ultrasonography was used to measure the brachial artery flow volume by a VA surgeon just before the procedure. Access flow was measured by duplex DUS using either Prosound Alpha 5 or Prosound Alpha 6 (Hitachi Aloka Medical) equipped with a broadband linear transducer (UST-5413, 411 MHz). Ultrasound was used to determine access depth and presence of adjacent vascular structures and the location of the anastomosis, size of the downstream arteries, and diameter of the banding site. Angiography (Allura Xper FD20; Philips Healthcare, Tokyo, Japan) was performed using a 4F or 5F vascular sheath (Xemex, Tokyo, Japan) for injecting iopamidol (300 mg/ml; Teba, Nagoya, Japan) into the VA and inverting flow in the venous limb with an inflated sphygmomanometer. The banding procedure was performed using local anesthesia with 1% lidocaine. With two parallel 1-cm incisions, the access was dissected subcutaneously using mosquito forceps. For safe dissection, especially below the access, ultrasound was used. After dissecting below and on the surface of the access, 2-0 monofilament polypropylene ligature (Surgipro II; Covidien, Tokyo, Japan) was pulled under and looped around the access. A 0.018-inch Radifocus guide wire (Terumo, Tokyo, Japan) or a 0.014-inch Runthrough guide wire (Terumo) and a Symmetry or Sterling catheter (Boston Scientific, Tokyo, Japan) were used. In patients with steal syndrome, the band size was no larger than the downstream artery. In addition, the lumen diameter was reduced by 60-80%, according to a nomogram (14), in patients with high-flow access.

An angioplasty balloon was then inflated to 15 atm in the area encircled by the suture loop, and six knots were tied (Fig. 1). After banding, the diameter of the banding site was measured by angiography and ultrasound, and Qa was monitored by ultrasound. In patients with steal syndrome, the finger probe of a pulse oximeter was attached to a finger on the ipsilateral side, and the peripheral oxygen saturation (SpO₂) was monitored.

Fig. 1 -

Definition of access outcomes

We calculated patency rates according to the published accepted standard (15). Secondary patency (access survival until abandonment) was defined as the interval from the time of access placement to access abandonment or time of patency measurement, including intervening manipulations (surgical or endovascular interventions) designed to reestablish the thrombosed access.

Statistical analysis

We used t-tests to compare continuous variables and χ² tests to compare categorical variables. Patency after surgery was estimated using the Kaplan-Meier technique. The difference in patency between the two groups was examined using log-rank test. Statistical analysis was performed using SPSS version 14 for Windows (SPSS Inc., Chicago, Illinois, USA). A p-value of less than 0.05 was considered statistically significant. Results are expressed as number of patients (percentage) or as mean ± SD.

Results

Patients

Seven cases of high-flow access and five cases of dialysis-associated steal syndrome were treated using the MILLER procedure. Patient characteristics and comorbid diseases are summarized in Table I. One steal case was complicated with high-flow access. The time between the last VA operation and the banding operation in the steal group was significantly shorter than that in the high-flow group (p = 0.005). Diabetes was more common in the steal group than in the high-flow group (p<0.01). All high-flow patients had the wrist radial-cephalic AVF. In the steal group, three cases had proximal VA (one brachial-cephalic and two brachial-communicating AVFs) and two cases had distal VA (one distal-forearm radial-cephalic AVF and one distal forearm ulnar-cephalic AVG).

Table I -

Scheme of the banding method. An inflated angioplasty balloon was used as a sizing dowel inside the access. A peri-access tunnel was dissected subcutaneously, and a 2-0 monofilament suture was pulled under and looped over the access. The ligature was then tightened around the balloon.

Demographic data of the high-flow and steal groups

Group	High-flow		Steal
ACE = angiotensin-converting enzyme inhibitor; ARB = angiotensin receptor blocker; ESA = erythropoiesis-stimulating agent.
Number of patients	7		5
Age (mean)	54.1		63
Gender (male)	6		2
Time from hemodialysis initiation to the banding operation	11.0 ± 4.0	(years)	7.6 ± 8.4	(years)
Time from the last AVF operation to the banding operation	7.1 ± 4.3	(years)	47.4 ± 52.6	(days)
Cause of end-stage renal disease
Diabetes mellitus (%)	0	(0)	4	(80)
Chronic glomerulonephritis (%)	4	(57)	1	(20)
IgA nephritis (%)	1	(14)	0	(0)
Polycystic kidney disease (%)	1	(14)	0	(0)
Hypertension (%)	1	(14)	0	(0)
Comorbidities
Diabetes mellitus (%)	0	(0)	4	(80)
Hypertension (%)	6	(86)	3	(60)
Ischemic heart disease (%)	2	(29)	1	(20)
Cerebrovascular disease (%)	2	(29)	1	(20)
Peripheral vascular disease (%)	2	(29)	2	(40)
Hyperlipidemia (%)	3	(60)	3	(60)
Medications
ACE (%)	2	(29)	0	(0)
ARB (%)	3	(60)	1	(20)
Calcium channel blocker (%)	5	(71)	3	(60)
ESA (%)	6	(86)	5	(100)
Systolic blood pressure (mmHg)	165 ± 25		140 ± 22
Diastolic blood pressure (mmHg)	71 ± 16		63 ± 20

The diameter of the VA, size of the balloon, and Qa both before and after banding on the operation day are summarized in Tables II and III. In the high-flow group (Tab. II), Qa ranged from 1401 to 2623 (mean ± SD 2043 ± 463) ml/min before banding. Balloon size ranged from 3 to 6 mm, with the most common size 3.5 mm (57%). The average diameter around the banding site decreased from 11.1 ± 1.7 to 3.7 ± 0.7 mm. The Qa in all cases of high-flow access decreased from 2043 ± 463 to 1248 ± 388 ml/min (p < 0.001). A second banding using a 1 mm larger balloon was required in one case because the flow was excessively reduced the next day. The mean reduction of the diameter around the banding site was 60.3 ± 15.9%, and the mean reduction in Qa was 39.4 ± 11.8%.

Table II -

Preoperative and postoperative vascular access data in the high-flow group

	Case no.	H1	H2	H3	H4	H5	H6	H7
AVF = arteriovenous fistula; AVG = arteriovenous graft; CV = cephalic vein; RA = radial artery.
Percentage stenosis of diameter in VA = {1 - (Diameter of VA after banding/Diameter of VA before banding)} x 100.
Mode of vascular access		AVF	AVF	AVF	AVF	AVF	AVF	AVF
Artery vein		RA-CV	RA-CV	RA-CV	RA-CV	RA-CV	RA-CV	RA-CV
Diameter of the distal artery	(mm)	5.9	4.1	3.7	5.2	3.4	2.9	3.2
Diameter of the balloon catheter	(mm)	6	3	3.5	3.5	3.5	4.5	3.5
Diameter of VA before banding	(mm)	10.5	4.5	8.1	17.4	20.1	11.1	6.3
Diameter of VA after banding	(mm)	4.8	2.9	3.8	4	3.5	4	2.8
Percent stenosis of diameter in VA	(%)	54.3	35.6	53.1	77.0	82.6	64.0	55.6
Flow volume before banding	(ml/min)	2623	2345	1955	1606	1864	2508	1401
Flow volume after banding	(ml/min)	1520	1160	893	1033	1150	2001	977

In the steal group (Tab. III), the balloon sizes ranged from 2.0 to 4.0 mm, and the mean reduction of the diameter was 67.5 ± 9.9%. SpO₂ values improved in all steal syndrome patients after banding. Symptoms were almost resolved in two steal syndrome patients in whom the AVF anastomosed the brachial artery with a communicating vein. The other steal syndrome patients experienced partial relief with the VA that was anastomosed at a distal location on the forearm and AVF anastomosed the brachial artery with the cephalic vein at the upper arm. The Qa in all cases of steal access decreased from 997 ± 867 to 548 ± 376 ml/min (p = 0.12).

Table III -

Preoperative and postoperative vascular access data in the steal group

	Case no.	S1	S2	S3	S4	S5
AVF = arteriovenous fistula; AVG = arteriovenous graft; ComV = communicating vein; CV = cephalic vein; RA = radial artery.
Percentage stenosis of diameter in VA = {1 - (Diameter of VA after banding/Diameter of VA before banding)} x 100.
Mode of vascular access		AVF	AVF	AVF	AVG	AVF
Artery vein		BA-ComV	RA-CV	BA-ComV	UA-CV	BA-CV
Diameter of the distal artery	(mm)	3.3	1.6	3.8	2.5	3
Diameter of the balloon catheter	(mm)	4	2	3.5	3	3
Diameter of VA before banding	(mm)	7	5.3	10.6	7.6	6.4
Diameter of VA after banding	(mm)	2.6	1.5	2.9	2.4	1.4
Percent stenosis of diameter in VA	(%)	62.9	71.7	72.6	68.4	78.1
Flow volume before banding	(ml/min)	560	354	2453	484	1133
Flow volume after banding	(ml/min)	325	247	1129	312	726

The Qa of the high-flow cases was significantly higher in the steal group not only before banding (2043 ± 463 vs. 997 ± 867 ml/min, p<0.05) but also after banding (1248 ± 388 vs. 548 ± 376 ml/min, p<0.05).

Outcome

The secondary patency rates of the high-flow and steal groups at 6 months were 83.3% and 50% and at 1 year were 83.3% and 25%, respectively. In the steal group, three patients needed VA ligation because of residual symptoms after the banding procedure (on 30, 90, and 126 days after banding, respectively). Among these ligated patients, diabetes and peripheral vascular disease were complications in three and two patients, respectively. In the radial-cephalic case, the radial artery was complicated by a high degree of arterial sclerosis with an extremely small diameter. The other distal forearm case was radial-basilic AVG. One patient in the high-flow group needed AVF ligation because of infection (298 days after banding). Another patient in the high-flow group required AVF reconstruction because of AVF failure (1112 days after banding).

Discussion

AVF creation induced a reduction in peripheral resistance and resulted in increased CO. High-flow access resulted in high-output heart failure, which was improved by either ligation or access flow reduction. One difficult issue with banding is how to improve the symptoms without a thrombosis. Several banding procedures are reported to reduce the flow of high-flow access (4, 5), but it is difficult to control the degree of Qa after banding. The MILLER banding procedure modulates a precise band size using an intraluminal balloon as a sizing dowel (13). Murray et al (14) pointed out that a little change in flow was seen with induction of at least 50% stenosis. To achieve a flow reduction to 1 l/min in a 1-cm-diameter fistula, 75% stenosis would be required. On the basis of a sizing nomogram (14), MILLER banding results in a 60-80% reduction in lumen diameter (13). In our high-flow group, the mean reduction of the diameter around the banding site was 60.3 ± 15.9%. Moreover, it is easy to correct the degree of the banding with this procedure. If the Qa is not reduced, additional banding using a smaller balloon is performed. On the contrary, if the Qa is too low, the band can be stretched by a larger-diameter angioplasty balloon (13). If stretching the banding was not effective, we rewound the banding and re-banded with new banding using a larger balloon. With other banding methods, readjustment to a suitable Qa is more complicated. With MILLER banding, correcting the degree of banding is simple and easy.

We added some modifications to the original MILLER banding procedure. Using intraoperative flow monitoring by DUS, we can adjust the degree of banding to ensure Qa after banding. Because we used the flow volume of the brachial artery on the ipsilateral upper limb to the VA in routine VA surveillance (16), we also checked ipsilateral brachial artery flow intraoperatively.

High CO is defined as a resting CO in adults greater than 8 l/min or cardiac index of greater than 3.9/min/m² (17). End-stage renal disease patients without an arteriovenous (AV)access have an average CO of 4.6 ± 0.9 l/min (18). Patients are reported to have symptomatic high-output heart failure in the setting of high AV access blood flow (3-4 l/min) and an increased CO of 7-10 l/min (3, 19, 20). High-flow access had been reported to result in high-output heart failure. Basile et al (6) reported that Qa values of at least 2.0 l/min and Qa/CO more than 20% predicted high-output cardiac failure. Moreover, the presence of an AVF negatively influences left ventricular (LV) performance and cardiac geometry in heart disease patients because its closure is associated with long-term improvement of echocardiographic functional and structural findings (21). Closure of an AVF is associated with a significant improvement in LV ejection fraction (LVEF), a significant decrease in LV mass and LV mass index, and a more favorable shift of cardiac geometry toward normality. Although only two patients (cases 2 and 4) underwent an echocardiogram before and after banding, EF was preserved in these two patients before banding. CO decreased after banding in both cases (from 7.6 to 5.9 in case 2 and from 7.5 to 5.6 in case 4). Although LVH and reduced LVEF are recognized as prognostic predictors in dialysis patients, most end-stage renal disease patients have LVH and preserved LVEF (EF ≥50%) (22, 23). Indeed, our high-flow group had preserved LVEF before banding. In patients with steal syndrome, the band size was no larger than the downstream artery. Moreover, flow reduction increased venous outflow resistance through a VA, thereby improving distal artery flow, the reason banding is used for steal syndrome and heart insufficiency (12, 13). In the steal group, balloon sizes ranged from 2.0 to 4.0 mm, and the mean reduction of the diameter was 67.5 ± 9.9%. Another modification we added to the original MILLER banding procedure in patients with steal syndrome is the use of a pulse oximeter attached to a finger of the ipsilateral upper limb to monitor the SpO₂ of the peripheral lesion, thus ensuring primary clinical success of banding during the operation.

Although Miller et al (13) obtained good patency in steal syndrome patients, our results in the steal group were not satisfactory. In our steal group, three patients needed VA ligation because of residual symptoms after banding. Two of these cases had distal forearm anastomosis, one requiring bypass of the radial artery to the basilic vein with a short prosthetic graft near the wrist joint and the other complicated by a high degree of arterial sclerosis. The Qa of these two cases was not very high (354 and 484 ml/min, respectively). As mentioned in the study by Miller et al (13), distal revascularization and interval ligation and proximalization of the arterial inflow may be the optimal treatments in cases of low-flow steal (<600 ml/min).

Our study had some limitations. First, the sample size was small, especially in the steal group. Second, only a few patients underwent echocardiography both before and after banding. More cases will need to be analyzed to clarify the effect of banding on cardiac function. We aim to resolve these issues in our future analyses.

Conclusion

The MILLER banding procedure with intraoperative access flow monitoring is effective to treat both high-flow VA and steal syndrome cases.

• 1. Foley RN. Clinical epidemiology of cardiac disease in dialysis patients: left ventricular hypertrophy, ischemic heart disease, and cardiac failure. Semin Dial 2003; 16: 111-117
• 2. Engelberts I.,Tordoir JH.,Boon ES.,Schreij G. High-output cardiac failure due to excessive shunting in a hemodialysis access fistula: an easily overlooked diagnosis. Am J Nephrol 1995; 15: 323-326
• 3. Stern AB.,Klemmer PJ. High-output heart failure secondary to arteriovenous fistula. Hemodial Int 2011; 15: 104-107
• 4. Zanow J.,Petzold K.,Petzold M.,Krueger U.,Scholz H. Flow reduction in high-flow arteriovenous access using intraoperative flow monitoring. J Vasc Surg 2006; 44: 1273-1278
• 5. Schneider CG.,Gawad KA.,Strate T.,Pfalzer B.,Izbicki JR. T-banding: a technique for flow reduction of a hyperfunctioning arteriovenous fistula. J Vasc Surg 2006; 43: 402-405
• 6. Basile C.,Lomonte C.,Vernaglione L.,Casucci F.,Antonelli M.,Losurdo N. The relationship between the flow of arteriovenous fistula and cardiac output in haemodialysis patients. Nephrol Dial Transplant 2008; 23: 282-287
• 7. KDOQI clinical practice guidelines and clinical practice recommendations for 2006 updates: hemodialysis adequacy, peritoneal dialysis adequacy and vascular access. Am J Kidney Dis 2006; 48: S1-S322
• 8. Tordoir J.,Canaud B.,Haage P. EBPG on vascular access. Nephrol Dial Transplant 2007; 22: ii88-ii117
• 9. 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: 8-10
• 10. Lazarides MK.,Staramos DN.,Panagopoulos GN.,Tzilalis VD.,Eleftheriou GJ.,Dayantas JN. Indications for surgical treatment of angioaccess-induced arterial “steal”. J Am Coll Surg 1998; 187: 422-426
• 11. Tordoir JH.,Dammers R.,van der Sande FM. Upper extremity ischemia and hemodialysis vascular access. Eur J Vasc Endovasc Surg 2004; 27: 1-5
• 12. Goel N.,Miller GA.,Jotwani MC.,Licht J.,Schur I.,Arnold WP. Minimally invasive limited ligation endoluminal-assisted revision (MILLER) for treatment of dialysis access-associated steal syndrome. Kidney Int 2006; 70: 765-770
• 13. Miller GA.,Goel N.,Friedman A. The MILLER banding procedure is an effective method for treating dialysis-associated steal syndrome. Kidney Int 2010; 77: 359-366
• 14. Murray BM.,Rajczak S.,Herman A.,Leary D. Effect of surgical banding of a high-flow fistula on access flow and cardiac output: intraoperative and long-term measurements. Am J Kidney Dis 2004; 44: 1090-1096
• 15. Sidawy AN.,Gray R.,Besarab A. Recommended standards for reports dealing with arteriovenous hemodialysis accesses. J Vasc Surg 2002; 35: 603-610
• 16. Shintaku S.,Kawanishi H.,Moriishi M.,Bansyodani M.,Tsuchiya S. Distal ulnar-basilic fistula as the first hemodialysis access. J Vasc Access 2014; 15: 83-87
• 17. Anand IS.,Florea VG. High output cardiac failure. Curr Treat Options Cardiovasc Med 2001; 3: 151-159
• 18. Coffey JP.,Woywodt A.,Hill JC. SPECT MIBI imaging for cardiac output and index in end stage renal disease. Hemodial Int 2011; 15: 320-325
• 19. Isoda S.,Kajiwara H.,Kondo J.,Matsumoto A. Banding a hemodialysis arteriovenous fistula to decrease blood flow and resolve high output cardiac failure: report of a case. Surg Today 1994; 24: 734-736
• 20. Engelberts I.,Tordoir JH.,Boon ES.,Schreij G. High-output cardiac failure due to excessive shunting in a hemodialysis access fistula: an easily overlooked diagnosis. Am J Nephrol 1995; 15: 323-326
• 21. Movilli E.,Viola BF.,Brunori G. Long-term effects of arteriovenous fistula closure on echocardiographic functional and structural findings in hemodialysis patients: a prospective study. Am J Kidney Dis 2010; 55: 682-689
• 22. Sharma R.,Gaze DC.,Pellerin D. Cardiac structural and functional abnormalities in end stage renal disease patients with elevated cardiac troponin T. Heart 2006; 92: 804-809
• 23. Wang AY.,Lam CW.,Wang M. Diagnostic potential of serum biomarkers for left ventricular abnormalities in chronic peritoneal dialysis patients. Nephrol Dial Transplant 2009; 24: 1962-1969