Predictors of Steal in Hemodialysis Access

ABRAHAM KORN, M.D., HAMID ALIPOUR, M.D., JOSHUA ZANE, M.D., KELSEY GRAY, M.D., TIMOTHY RYAN, M.D.,
AMY KAJI, M.D., PH.D., CHRISTIAN DE VIRGILIO, M.D., NINA BOWENS, M.D.

From the Harbor-UCLA Medical Center, Department of Surgery, Division of Vascular Surgery,
Torrance, California

Pathological steal is an uncommon but devastating complication after the creation of arteriove-
nous access for hemodialysis. In this study, we sought to assess predictors of clinically significant
steal syndrome and to further evaluate the outcome of differing surgical treatment approaches. A
retrospective analysis was performed of 282 consecutive patients undergoing hemodialysis access
at a single center from November 2014 to April 2016. Adequate follow-up to assess for the de-
velopment of steal was obtained in 237 patients. One hundred and fifty-seven (66%) patients were
male, 152 (64%) Hispanic, and 164 (69%) had diabetes. Forty-three (18%) had prior access pro-
cedures. Autologous fistula was created in 218 patients (92%). Pathologic steal occurred in 15
patients (6.7%). On univariate analysis, significant predictors of steal included female sex [P 5
0.03, odds ratio (OR) 5 3.3, CI [1.1–9]), no systemic heparin at operation (P 5 0.02, OR 5 5.0, CI
[1.4–10]), use of angiotensin-converting enzyme inhibitor (P 5 0.003, OR 5 5.6, CI [1.7–18.6]), and
increased vein size (3.1 vs 4.1 mm P 5 0.01). Twelve patients had steal managed with an in-
tervention, but only one patient received distal revascularization. Furthermore, we identify key
predictors of clinically significant steal syndrome while demonstrating that distal re-
vascularization and/or fistula ligation are rarely indicated treatment modalities.

H a cornerstone in the modern

EMODIALYSIS ACCESS IS
management of end-stage renal disease. Vascular
steal occurs when blood flow is diverted from the ar-
The purpose of this study was to identify the in-
cidence and risk factors for steal as well as to provide
a description of the management and outcomes. We
terial tree distal to the fistula into the lower resistance further hypothesized that steal can be successfully
venous system. When clinically significant, steal leads treated in most patients without DRIL or fistula
to inadequate distal perfusion and symptoms of is- ligation.
chemia such as pain, coolness, pulse deficit, and in
severe cases, tissue loss.
Steal can be assessed on a grading scale from 0 to 3. Methods
Grade 0 is asymptomatic. Grade 1 steal is mild and
A retrospective review of all patients who un-
involves coolness and few symptoms. Grade 2 symp-
derwent initial hemodialysis access procedures at the
toms are moderate and include intermittent pain, and
Harbor-UCLA Medical Center between November
grade 3 is severe and manifests with rest pain and
2014 and April 2016 was performed. Institutional
tissue loss. Whereas grades 0 and 1 do not usually
Review Board approval was obtained. Collected de-
require intervention, Grade 2 and Grade 3 often do.
mographics included age, ethnicity, body mass index,
Treatment approaches vary. A commonly described
race, cause of end-stage renal disease, and dialysis
approach to the management of significant steal is
status. Patients were preoperatively evaluated for ar-
distal revascularization with interval ligation (DRIL).
terial disease via formal ultrasound in 151 cases
However, the DRIL procedure is a somewhat complex
concomitant with vein mapping. Arterial evaluation
solution that requires saphenous vein harvest and
consisted of measurement of artery size, presence of
general anesthesia. Plication is a less-utilized but
calcifications, and arterial waveform as well as
simpler alternative and has the advantage of being
completeness of palmer arch. The remaining patients
performed under local anesthesia.1–5
received ultrasound performed by surgical residents
in vascular surgery clinic noting arterial diameter.
Steal was defined as the postoperative development of
symptoms consisting of coolness, signs of ischemia,
Address correspondence and reprint requests to Nina Bowens,
M.D., Division of Vascular Surgery, Department of Surgery, pain, paresthesia or weakness distal to the site of
Harbor-UCLA Medical Center, Torrance, CA. E-mail: nbowens@dhs. hemodialysis access as assessed by the clinicians
lacounty.gov. performing routine care. Patients were routinely seen

1099
1100 THE AMERICAN SURGEON October 2017 Vol. 83

in clinic by a vascular surgeon 1 to 2 weeks and 4 to
6 weeks after surgery. Statistical analysis of potential
risk factors for steal was performed by univariate
logistic regression with SAS software. Significance
of the univariate variable was preselected with a two-
tailed P value of 0.05.
Results
There were 282 consecutive patients who underwent
hemodialysis access procedures during this study pe-
riod. Adequate follow-up, defined as at least one in-
teraction with vascular surgery service sufficient to
assess and diagnose steal after one month following
surgery, was available in 237 cases to assess for the
development of steal. One hundred and sixty (66.4%)
patients were male. One hundred and fifty-four
(63.9%) patients were Hispanic. Hypertension was
present in 215 patients (89.2%), diabetes was present
in 164 (68%). Current smoking or history of smoking
was present in 95 patients (39.4%). Patients had pre-
viously undergone dialysis access procedures in 45
(18.7%) cases. Median follow-up time was 12 months.
The procedures were performed by eight surgeons in
total.
Clinically significant steal occurred in 15 patients
(6.3%). The demographic features of the patients are
presented in Table 1. On univariate analysis, signifi-
cant predictors of steal included female sex (P 4 0.03,
OR 4 3.3, CI [1.1–9]), no systemic heparin at operation

TABLE 1. Patient Demographics, Medical History, and Medication History in Steal

No steal (n4222) Steal (n415) OR [95% CI] P value
Male 151 (68%) 6 (40%) 0.3 [0.1-0.9] 0.03*
Hispanic 141 (63.5%) 11 (73.3%) 1.6 [0.5-5.1] 0.4
Median age (y) 54.9 56.2 0.5
Median BMI 27.7 27.9 0.9
Race 0.4
White 155 (69.8%) 13 (86.7%)
African-American 45 (20.3%) 2 (13.3%)
Asian 17 (7.7%) 0 (0%)
Native Hawaiian/Pacific islands 5 (2.3%) 0 (0%)
Cause of ESRD 0.9
Hypertension and Diabetes 126/200 (63%) 8 (53.3%)
Hypertension 37/200 (18.5%) 4 (26.7%)
Diabetes 13/200 (6.5%) 1 (6.7%)
Primary kidney disease 20/200 (10%) 2 (13.3%)
Malignancy 3/200 (1.5%) 0 (0%)
Congenital 1/200 (0.5%) 0 (0%)
Diabetes 153 (78.9%) 10 (66.7%) 0.9
Hypertension 198 (89.2%) 13 (86.7%) 0.8 [0.2-3.7] 0.8
Dialysis at time of surgery 141 (63.5) 6 (40%) 0.4 [0.1-1.1] 0.07
Prior CABG or PCI 13 (5.9%) 2 (13.3%) 2.5 [0.5-12.1] 0.2
Current smoking 27 (12.2%) 2 (13.3%) 0.8
Previous access procedure 41 (18.5%) 2 (13.3%) 0.9
Statin 114/217 (52.5%) 7/14 (50%) 0.9 [0.3-2.7] 0.9
ACE Inhibitor 66/215 (30.7%) 10/14 (71.4%) 5.6 [1.5-18.6] 0.003*
Aspirin 91/216 (42.1%) 4/14 (28.6%) 0.4

TABLE 2. Operative Characteristics of Steal Patients

Features of Operation
No Steal (n4222) Steal (n415) P value
AV Graft 17 (8%) 2 (13%) 0.3
Regional anesthesia 35 (16%) 3 (20%) 0.9
Systemic Heparin 182/214 (82%) 8/14 (57%) 0.02*
Artery diameter (mm) 4 n4140 4.1 n410 0.7
Vein diameter (mm) 3.3 n4131 4.2 n411 0.01*
Autologous Fistula Configuration (n4218) 0.4
Radiocephalic 37 (18.2%) 0 (0%)
Brachiocephalic 100 (49.3%) 10 (66.7%)
Brachiobasilic 54 (26.6%) 3 (20%)
Brachiobrachial 9 (4.4%) 0 (0%)
Brachioantecubital 5 (2.5%) 0 (0%)
No. 10 PREDICTORS OF STEAL ? Korn et al. 1101

(P 4 0.02, OR 4 5, CI [1.4–10]), and the use of in only one patient. Ligation was not utilized as
angiotensin-converting enzyme inhibitor (ACE) in- a treatment for steal in this study group.
hibitor (P 4 0.003, OR 4 5.6, CI [1.7–18.6]). Op-
erative factors associated with steal, are shown in
Discussion
Table 2. Larger vein size was associated with an in-
creased risk of steal (3.1 vs 4.1 mm, P 4 0.01). Clinically symptomatic steal is a serious complica-
Nonsignificant factors for the development of steal tion of hemodialysis access creation due to the asso-
included diabetes, age, body mass index, location of ciated morbidity including ischemic symptoms and the
fistula, artery size on preoperative ultrasound, intra- frequent need for revision surgery. Knowledge of both
operative arterial size, and tobacco use. Results of patient and operative characteristics associated with
a multivariate analysis of intraoperative heparin, vein the development of clinically significant steal can aid
diameter, and patient sex are listed in Table 3. On mul- in its prevention and management. Risk factors for the
tivariate analysis, intraoperative heparin does not main- development of steal in previous studies included
tain significance as the P value is 0.1. Female gender brachial artery access, peripheral arterial disease,
demonstrated a strong trend with a P value of 0.054. diabetes, and female sex.3, 6, 7 In this study, steal was
Of the 15 patients who developed steal, three had noted in 6.3 per cent of patients. Several factors were
Grade 1 steal and were observed. Four had Grade 2 found to be associated with an increased incidence of
steal and eight had Grade 3 steal. All three Grade 1 steal including female sex, use of an ACE inhibitor,
steal patients were observed and had improvement or and larger vein size. The most important was female
no progression. Intervention type is noted in Table 4. gender, a finding that has been previously reported.5, 8
The 12 remaining patients with steal received One proposed explanation identifies the smaller cali-
interventions. The most common intervention was ber arteries in women as inadequate to carry sufficient
plication, which was performed in five cases. Other blood to supply both the fistula and distal tissues.5, 9 A
interventions included proximalization of inflow, proximal fistula configuration has also been associated
tributary ligation, transposition with proximalization with the development of steal. Higher arterial flow in
of inflow, distal revascularization, and venous angio- the proximal feeding artery may lead to higher blood
plasty. Of the 12 patients who received interventions, flow in the fistula therefore diverting more blood away
nine (75%) improved after the initial intervention. from the arterial system distal to the fistula anasto-
Three patients (25%) did not improve and so un- mosis.3, 9 Accordingly, in our study, no distal radio-
derwent subsequent interventions that were additional cephalic fistulas developed steal, although this did not
plication in two cases and plication with subclavian achieve significance.
artery stent and angioplasty in one case. After the The use of heparin during hemodialysis access is
subsequent interventions, two (67%) of the three had debatable. Eighty percent of patients received heparin
improvement of symptoms and one (33%) did not. The in the present study. It is not obvious why intra-
patient who did not improve underwent subsequent operative systemic heparin may have resulted in
resection of stenotic portion of vein with plication and a lower risk of steal. It may reflect a selection bias.
had continued symptoms. Notably, DRIL was utilized Given the retrospective nature of our study, the
indication for when to use heparin was not clearly
elucidated. Furthermore, the low incidence of steal
TABLE 3. Multivariate Analysis of Risk Factors for Steal precluded the use of multivariate analysis to elucidate
OR CI P-Value interdependence of variables.
Female 3.85 0.98-16.7 0.05 A novel finding of this study is the association of
Systemic Heparin 0.3 0.1-1.4 0.1 ACE-inhibitor use with an increased risk for the de-
Vein Diameter 1.2 1-1.5 0.07 velopment of steal syndrome. Inhibition of peripheral
vascular response to the renin-aldosterone pathway
that promotes vascular remodeling and has previously
TABLE 4. Management of Steal
been associated with improved primary patency in
Management of Steal N 415 fistulas.10, 11
Plication 5 Our results confirm that the DRIL procedure and
Proximalization of inflow 2 access ligation are rarely necessary to treat steal. At
Transposition with proximalization 1 our institution, plication has been used as the primary
Plication with superficialization 1
Distal revascularization 1 treatment modality. It is a highly effective treatment
Tributary ligation 1 option for clinically significant steal. This procedure
Angioplasty 1 can be performed under local anesthesia, can achieve
No Intervention 3 precise cessation of flow reversal when used with
1102 THE AMERICAN SURGEON
intraoperative ultrasound, and can even be repeated if
steal symptoms persist.12 In addition, some of the pa-
tients with mild steal were able to be managed without
any interventions with good outcomes. Previous stud-
ies have reported a much higher rate of ligation.5, 8
Avoidance of ligation allows for continued use of the
fistula, whereas ligation requires additional access
procedures that are also at risk of development of steal.
As experience increases with alternative approaches,
current practice is evolving to reduce the number of
ligations performed for management of steal syn-
drome. The prevention and management of steal syn-
drome in patients requiring hemodialysis access is
multifaceted and requires an understanding of both
modifiable and nonmodifiable risk factors.
REFERENCES
1. Leon C, Asif A. Arteriovenous access and hand pain: the
distal hypoperfusion ischemic syndrome. Clin J Am Soc Nephrol
2007;2:175–83.
2. Zamani P, Kaufman J, Kinlay S. Ischemic steal syndrome
following arm arteriovenous fistula for hemodialysis. Vasc Med
2009;14:371–6.
3. Rocha A, Silva F, Queir ós J, et al. Predictors of steal
syndrome in hemodialysis patients. Hemodial Int 2012;16:
539–44.
October 2017 Vol. 83
4. Crononwett JL. Rutherford’s Vascular Surgery, 8th Ed,
Elsevier Health Sciences, 2014.
5. Gupta N, Yuo TH, Konig G, et al. Treatment strategies of
arterial steal after arteriovenous access. J Vasc Surg 2011;54:
162–7.
6. Kudlaty EA, Kendrick DE, Allemang MT, et al. Upper
extremity steal syndrome is associated with atherosclerotic
burden and access configuration. Ann Vasc Surg 2016;35:
82–7.
7. Morsy AH, Kulbaski M, Chen C, et al. Incidence and char-
acteristics of patients with hand ischemia after a hemodialysis
access procedure. J Surg Res 1998;74:8–10.
8. DeCaprio JD, Valentine RJ, Kakish HB, et al. Steal syndrome
complicating hemodialysis access. Cardiovasc Surg 1997;5:
648–53.
9. Pounds LL, Teodorescu VJ. Chronic kidney disease and
dialysis access in women. J Vasc Surg 2013;57:49S–53S.e1.
10. Pruthi D, McCurley A, Aronovitz M, et al. Aldosterone
promotes vascular remodeling by direct effects on smooth muscle
cell mineralocorticoid receptors. Arterioscler Thromb Vasc Biol
2014;34:355–64.
11. Jackson RS, Sidawy AN, Amdur RL, et al. Angiotensin
receptor blockers and antiplatelet agents are associated with im-
proved primary patency after arteriovenous hemodialysis access
placement. J Vasc Surg 2011;54:1706–12.
12. Yaghoubian A, de Virgilio C. Plication as primary treatment
of steal syndrome in arteriovenous fistulas. Ann Vasc Surg 2009;
23:103–7.
Reproduced with permission of copyright owner. Further reproduction
prohibited without permission.