- Email: [email protected]
Steal syndrome complicating hemodialysis access
PII: S0967-2109(97)00072-0
Cardiovascular Surgery, Vol. 5, No. 6, pp. 648–653, 1997 1997 The International Society for Cardiovascular Surgery Published by Elsevier Science Ltd. Printed in Great Britain 0967–2109/97 $17.00 + 0.00
Steal syndrome complicating hemodialysis access J. D. DeCaprio*, R. J. Valentine*, H. B. Kakish*, R. Awad†, R. T. Hagino* and G. P. Clagett* *Division of Vascular Surgery, Department of Surgery, and †Department of Radiology, Department of Veterans Affairs Medical Center, and The University of Texas Southwestern Medical Center, Dallas, Texas, USA The purposes of this study were to examine potential causes for arterial steal syndrome in patients with hemodialysis shunts and to evaluate the results of treatment. A retrospective analysis was performed on all hemodialysis patients who presented with arterial steal syndrome over a 3-year period. Medical records were reviewed to determine demographic data, time to development of symptoms, operative details, and treatment outcome. Results of upper-extremity arteriography in the last five patients were also documented. Eighteen patients presented with steal syndrome at a mean of 7(4) months after shunt placement. Four patients had prior ipsilateral grafts. Fifteen (83%) patients had forearm shunts in loop configuration between the main brachial artery and a superficial vein, and three had straight upper-arm shunts. Graft sizes were 6 mm in five patients and 4- to 7 step grafts in 13. Graft banding was used to treat steal in 11 shunts, one of which remained patent. All five patients who underwent upper-extremity arteriography had arterial stenoses demonstrated in the inflow circulation. In conclusion, arterial steal is an uncommon complication of hemodialysis shunts. Shunt location and choice of graft do not appear to be major factors: steal can develop regardless of access site, configuration, or use of straight versus step graft. Banding is an ineffective method of treatment. Arteriography may be important to detect unsuspected inflow stenoses in patients with steal syndrome. 1997 The International Society for Cardiovascular Surgery Keywords: steal syndrome, hemodialysis, shunt placement, access site
The onset of arterial steal symptoms portends graft loss in patients with forearm hemodialysis shunts. This complication has been reported in 1.6% to 8% of access patients, who experience symptoms ranging from neurosensory and motor changes in the hand to extensive gangrene [1–4]. Treatment options have generally been aimed at reducing blood flow through the shunt. This has been achieved by localized diameter reduction (banding), graft extension, and graft ligation [5–7]. Although graft banding has become a procedure of choice in these patients, reported longterm graft patency is severely compromised [8].
Correspondence to: Dr. R. J. Valentine, Department of Surgery, U.T. Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, Texas 75235-9157, USA
648
Hemodynamically significant stenoses in donor arteries are well-known causes of steal syndrome in other circulations [9, 10]. While previous authors have commented on the possibility of inflow stenoses in access patients with steal syndrome, few have performed arteriography as part of a rigorous search for flow-limiting lesions. The present authors have recently instituted a policy of routine upperextremity arteriography to evaluate the cause of steal symptoms in patients with hemodialysis access. The purposes of this study were: (i) to determine possible causative factors for steal syndrome; (ii) to evaluate the outcome of treatment; and (iii) to document the findings of routine arteriography in the last five patients with steal symptoms.
CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al.
Patients and methods Between 1993 and 1996, 1020 forearm grafts were placed for hemodialysis access at the authors’ institution. The records of all patients who presented with symptoms of hand ischemia following access placement were reviewed. Patients with objective signs of hemodynamic steal were included in the analysis; this was determined by relief of steal symptoms and/or improvement of photoplethysmographically derived digital pressures during temporary shunt occlusion. Medical records were reviewed for demographic data, degree of finger ischemia, and preoperative assessment of arterial and venous circulations. Graft sizes and locations were documented. Time intervals between placement of shunts and development of symptoms were noted, as were interventional diagnostic tests and treatment regimes. Non-invasive tests obtained following treatment were recorded.
syndrome were placed in the forearm, and three (17%) were placed in the upper arm. Three patients had known blood pressure differentials in the upper extremity at the time of graft placement; grafts were placed in the arm with the highest pressure in these cases. The arterial anastomosis was created with the brachial artery in all cases. The venous anastomosis was created with superficial antecubital veins in forearm grafts and with the basilic vein in upper arm grafts. Graft types included straight 6-mm polytetrafluoroethylene (Gore-Tex, Gore Medical Products, Flagstaff, AZ) in five patients and 4- to 7-mm step polytetrafluoroethylene in 13 patients (GoreTex). Hand symptoms included pain and neurosensory deficits in all 18, neuromotor symptoms in two, and tissue loss in three. The mean lowest finger pressure at the time of diagnosis was 32(7) (range 0–84) mmHg. Upper-Extremity arteriography
Results During the study period, 18 (1.8%) hemodialysis patients presented with symptomatic hand steal syndrome at a mean (s.e.m.) of 7(4) (range 0–60) months (median 2 months) after forearm arteriovenous graft placement. Demographics and atherosclerotic risk factors are shown in Table 1. The mean duration of hemodialysis among the 18 patients was 30(9) (range 0–120) months (median 7 months). Nine (50%) patients had previous grafts placed with subsequent thrombosis, and one other patient had a previous primary arteriovenous fistula which occluded. The mean number of prior grafts in the nine patients was two (range one to three). Four (18%) patients had previous grafts in the extremity ipsilateral to the graft which developed steal syndrome. Fifteen (83%) of the grafts which developed steal Table 1 Demographics and atherosclerotic risk factors of the study patients. No. of patients
18
Mean (s.e.m.) age years Race Black White Hispanic No. of males Risk factors Smoking Hypertension Diabetes Dyslipidemia Peripheral vascular disease
61(3) 8 7 3 8
(44) (39) (17) (44)
7 14 14 8 4
(39) (78) (78) (44) (22)
Values in parentheses are percentages (unless stated)
CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
Five patients underwent upper-extremity arteriography to evaluate steal symptoms during the past 12 months. These represented the most recent five patients who presented with steal symptoms following hemodialysis access placement. The mean age of the five patients was 69(3) years. The prevalence of atherosclerotic risk factors and the mean duration of hemodialysis were similar to that noted for the other 13 patients who did not undergo arteriography. One of the five had differential arm blood pressure measurements exceeding 20 mmHg at the time of graft placement, and the graft was placed in the arm with the highest blood pressure, as noted above. Following onset of steal symptoms, all five had lower ipsilateral finger pressures. None of the patients had audible bruits over shoulder or upper-extremity arteries. Arteriography demonstrated arterial stenoses measuring at least 50% diameter loss in the donor arterial circulations of all five patients (Figure 1 and Figure 2). Four of the five patients had stenoses located in the axillosubclavian arteries, one had a mid-brachial artery stenosis, and one had lesions in both locations. The one patient with known blood pressure differentials had bilateral axillary artery stenoses. Treatment Results of treatment are shown in Table 2. Eleven (61%) patients underwent banding procedures. In all cases, graft banding was monitored with intraoperative finger plethysmography. Bands were constructed using large, stainless steel ligating clips (Ligaclip Extra, Ethicon Inc., Summerville, NJ), and grafts were narrowed until finger pressures increased to values above 50 mmHg or until waveforms nor649
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al.
malized, as previously recommended [8]. The mean postoperative finger pressure was 94(11) mmHg in this group. Following banding, 10 patients had resolution of steal symptoms, and one required graft ligation to treat continued hand pain. However, only one of the 10 banded grafts remained patent for longer than 1 week. None of the patients with thrombosed grafts redeveloped steal symptoms. Three patients underwent graft ligations for advanced steal symptoms, including one who had undergone a previous attempt at banding noted above. A fourth patient had resolution of steal symptoms after graft thrombosis and a fifth was lost to follow-up before therapy could be instituted. Two of the five patients with known inflow stenoses underwent percutaneous balloon angioplasty with stent placement (Figure 1 and Figure 2). Both patients had normalization of finger pressures and resolution of symptoms. Of the remaining three patients with inflow stenoses identified angiographically, one had graft ligation without correction of inflow stenosis due to multilevel arterial disease; one failed an attempt to reopen an occluded axillary artery and had resolution of symptoms after graft thrombosis; and one underwent graft banding without correction of inflow lesions. The banded graft thrombosed within 2 days.
Discussion
Figure 1 a Upper extremity arteriogram demonstrating a stenosis at the origin of the left subclavian artery in a patient with steal symptoms. A pressure drop of 27 mmHg was measured across the stenosis. b Completion arteriogram after balloon angioplasty and stent placement to correct the subclavian stenosis. The pressure differential was corrected, and the steal symptoms resolved
650
More than 100,000 patients receive chronic hemodialysis each year in the US, with the majority of these patients expected to live ten years or longer [11]. Graft complications associated with multiple cannulations are common in these patients, especially infection, thrombosis and pseudoaneurysm formation [1, 2, 4, 12]. Although associated vascular disease is prevalent in patients undergoing hemodialysis, circulatory problems associated with arteriovenous shunt formation are rare. The incidence of symptomatic vascular steal syndrome has been reported to occur in 1.6% to 8%, depending on the type of arteriovenous fistula and location of graft [1–4]. The present results confirm that steal syndrome is uncommon in these patients. However, the present findings suggest that steal can occur in shunts regardless of location, configuration, or use of small diameter grafts. The authors cannot determine the influence of graft size on the development of steal per se, since they do not use 8-mm or larger diameter grafts in the creation of shunts and are unaware of any patients with smaller grafts or primary arteriovenous fistulas who developed steal syndrome during the study period. While it has been suggested that use of 4- to 7-mm ‘step’ grafts can decrease the risk of steal syndrome [13], the present study demonstrates that steal complications can occur with these grafts. CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al.
Figure 2 a Upper extremity arteriogram demonstrating multiple sequential stenoses in the mid-brachial artery. A pressure differential of 20 mmHg was measured across the stenosis. b Completion arteriogram following balloon angioplasty and stent placement to correct stenotic areas. The patient’s steal symptoms resolved.
CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
651
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al. Table 2 Treatment of patients with steal syndrome, and outcome. Modality
No. of patients
Outcome
Banding
11
9 occluded 1 continued steal symptoms 1 patent, no symptoms Symptoms resolved Symptoms resolved, grafts patent 1 graft thrombosed, symptoms resolved 1 lost to follow-up
Ligation 3 Angioplasty, stent placement 2 None
2
Total
18
Treatment of steal syndrome has focused on reduction of shunt diameter. In this regard, graft banding has become the procedure of choice in many centers. Unfortunately, patency data following graft banding have not been adequately reported. In the single series reporting late results of this technique, Odland et al. [8] acknowledged that their banded graft patency rates of 62.5% at 6 months and 38.5% at 1 year were inferior compared with their unbanded graft results. The present results suggest that banding is an ineffective method of treating steal syndrome, with a 6-month patency rate of 9% among 11 patients who underwent banding according to the method described by Odland et al. [8]. While symptoms resolved in all of the present patients following banding, 10 of 11 had resolution of symptoms due to graft occlusion. This suggests that simple reduction of graft flow is not adequate to treat steal syndrome. Other causes for the steal syndrome should be sought and corrected. One possibility is unsuspected inflow stenosis. The present authors confirmed the presence of inflow stenoses measuring ⱖ 50% diameter loss in their last five patients, all of whom underwent arteriography. It should be stressed that arterial stenoses were not clinically suspected before shunt placement in any of these patients on the basis of ischemic symptoms or blood pressure differentials. Calligaro et al. [14] have confirmed that equal arterial pressures in the upper extremities do not exclude hemodynamically significant stenoses in the axillosubclavian arterial circulation. Crawford and colleagues [15] demonstrated that 27% of patients with symptoms of upper-extremity ischemia due to arterial disease had normal, symmetrical arm blood pressures. Thus, the presence of normal or equal arm blood pressures does not exclude the possibility of occult inflow stenosis. The present authors suspect that the higher flows associated with the arteriovenous shunt circuit may unmask inflow lesions that would otherwise have borderline hemodynamic significance. Since 652
these lesions may not be suspected on the basis of clinical examination alone, it might be argued that upper-extremity arteriography should be performed in all patients before placement of hemodialysis shunts. However, since the prevalence of steal syndrome is low in patients following placement of hemodialysis shunts, the authors do not recommend routine preoperative arteriography. Duplex ultrasonography may offer a non-invasive alternative to arteriography for preoperative evaluation. However, the low prevalence of steal symptoms in the present patient population suggests that there will be relatively few arterial stenoses identified, and certainly too few to justify the cost of routine Duplex scanning. Based on their initial results, the authors would recommend arteriography in all patients who develop steal syndrome after shunt placement. They speculate that the majority of patients in the present series who did not undergo arteriography also had inflow stenoses. There were no other apparent differences in examined variables between these 13 patients and the last five patients who underwent arteriography. Another possible cause of arterial steal syndrome is inadequate arterial flow in the circulation distal to the vascular access. This may occur, for example, in diabetic patients with distal arteriopathy resulting in inadequate collateral flow to the hand or digits. It should be pointed out that 78% of the present study population was diabetic. Therefore, distal disease may have been a prominent feature of the steal syndrome in many of the patients who did not undergo arteriography. However, since four of the five patients with angiographically identified inflow stenoses were diabetic, the authors submit that arteriography should be considered even in patients who are at risk of having distal disease, in the hope of identifying potentially correctable proximal lesions. What should be done with patients who do not have correctable inflow stenoses? Our data corroborate the results of others [8] that banding is tantamount to graft ligation. While interposing narrow graft segments has theoretical appeal, the fact that steal syndrome occurred in patients with 4- to 7-mm step grafts suggests that this technique may not always be curative. Nevertheless, interposition of a narrowed graft segment (e.g. 4 mm diameter) may be worthwhile, with graft ligation reserved for patients with continued symptoms. The risk of tissue loss appears to be significant in symptomatic patients, so the authors do not recommend observation alone. Despite being a large series of hemodialysis patients with steal syndrome, there are limitations in the present study. While it is interesting to speculate on the possible association between steal syndrome and occult inflow stenoses, a cause-and-effect relationship cannot be established since only two CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al.
patients had correction of inflow lesions. However, it is interesting to note that both patients had relief of steal symptoms and continued graft patency. A second limitation is that this study is retrospective in nature; therefore, a number of patients with steal syndrome may have been missed. While it is acknowledged that the steal prevalence data may be underestimated, it is unlikely that symptomatic patients who remained in the authors’ institution’s hemodialysis population were missed due to the diligence of their nephrology services. The present reported steal prevalence is within the range documented by others [1–4]. However, the authors cannot account for the small proportion of patients who are no longer followed at their institution. Another potential limitation of the present study is that there was reliance on the presence of symptoms to detect patients with steal syndrome. Previous reports suggest that there is a high prevalence of symptomless but hemodynamically significant steal in patient with primary arteriovenous fistulas [16]. Since the diameter of the graft nearly always exceeds the diameter of the artery, it is likely that the great majority of patients have measurable, hemodynamic steal in the ipsilateral extremity. The significance of this is unknown, but given the low prevalence of tissue loss in the present symptomless dialysis population, it is suggested that hemodynamic steal without symptoms is probably benign. The authors speculate that some of our institutions dialysis patients with arterial steal may not have been detected due to their being symptomless or minimally symptomatic. It is possible that these symptomless patients are at risk of shunt occlusion or tissue loss. The authors are currently designing a study to evaluate these issues prospectively. In summary, the present study indicates that arterial steal syndrome is an uncommon complication in hemodialysis patients with upper-extremity shunts. Steal symptoms can occur in patients with either 6-mm or 4- to 7-mm step grafts, regardless of shunt location or configuration. Banding maneuvers to decrease the diameter of the shunt do not appear to be effective in treating steal syndrome, suggesting that the cause of symptoms is not related to the size of the graft. Early results with routine arteriography in steal patients suggest that a search for inflow stenosis may be rewarding. The authors recommend routine arteriography in patients who develop steal symptoms after placement of hemodialysis shunts in order to detect potentially correctable lesions.
CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
References 1. Haimov, M., Baez, H., Neff, M. and Slifkin, R., Complications of arteriovenous fistulas for hemodialysis. Archives of Surgery, 1975, 110, 708–712. 2. Zibari, G. B., Rohr, M. S. and Landreneau, M. D. et al., Complications from permanent hemodialysis vascular access. Surgery, 1988, 104, 681–686. 3. Kwun, K. B., Schanzer, H., Finkler, N., Haimov, M. and Burrows, L., Hemodynamic evaluation of angioaccess procedures for hemodialysis. Vascular Surgery, 1979, 13, 170–177. 4. Ballard, J., Blunt, T. J. and Malone, J., Major complications of angioaccess surgery. American Journal of Surgery, 1992, 164, 229–232. 5. Mattson, W., Recognition and treatment of vascular steal secondary to hemodialysis prosthesis. American Journal of Surgery, 1987, 154, 198–201. 6. Rivers, S. P., Scher, L. A. and Veith, F. J., Correction of steal syndrome secondary to hemodialysis access fistulas: a simplified quantitative technique. Surgery, 1992, 112, 593–597. 7. West, J. C., Evans, R. D. and Kelley, S. E. et al., Arterial insufficiency in hemodialysis access procedures: reconstruction by an interposition polytetrafluoro-ethylene graft conduit. American Journal of Surgery, 1987, 153, 300–301. 8. Odland, M. D., Kelley, P. H., Ney, A. L., Andersen, R. C. and Bubrick, M. P., Management of dialysis associated steal syndrome complicating upper extremity arteriovenous fistulas: Use of intraoperative digital photoplethysmography. Surgery, 1991, 110, 664–670. 9. Flanigan, D. P., Pratt, D. G., Goodreau, J. J., Burnham, S. J., Yao, J. S. T. and Bergan, J. J., Hemodynamic and angiographic guidelines in the selection of patients for femorofemoral bypass. Archives of Surgery, 1978, 113, 1257–1262. 10. Fisher, C. M., A new vascular syndrome–the subclavian steal. New England Journal of Medicine, 1961, 265, 912. 11. Bennion, R. S., Socioeconomic implications of vascular access surgery. In Vascular Access: Principles and Practice, ed. S. E. Wilson. Mosby-Year Book Inc., St. Louis, 1996, pp. 284–293. 12. Hill, S. L. and Donato, A. T., Complications of dialysis access; A six-year study. American Journal of Surgery, 1991, 162, 265–267. 13. Rosental, J. J., Bell, D. D., Gaspar, M. R., Mobius, H. J. and Lemire, G. G., Prevention of high flow problems of arteriovenpous grafts: Devlopment of a new tapered graft. American Journal of Surgery, 1980, 140, 231–233. 14. Calligaro, K. D., Ascer, E., Veith, W., Gupta, S. K., Wengerter, K. R., Franco, C. D., Bakal, C. W. and Sprayregen, S., Unsuspected inflow disease in candidates for axillofemoral bypass operations: A prospective study. Journal of Vascular Surgery, 1990, 11, 832–837. 15. Crawford, E. S., De Bakey, M. E., Morris, G. C. and Colley, D. A., Thrombo-obliterative disease of the great vessels arising from the aortic arch. Journal of Thoracic Cardiovascular Surgery, 1962, 43, 38–53. 16. Duncan, H., Ferguson, L. and Faris, I., Incidence of the radial steal syndrome in patients with Brescia fistula for hemodialysis: Its clinical significance. Journal of Vascular Surgery, 1986, 4, 144–147.
Paper accepted 9 June 1997
653
Cardiovascular Surgery, Vol. 5, No. 6, pp. 648–653, 1997 1997 The International Society for Cardiovascular Surgery Published by Elsevier Science Ltd. Printed in Great Britain 0967–2109/97 $17.00 + 0.00
Steal syndrome complicating hemodialysis access J. D. DeCaprio*, R. J. Valentine*, H. B. Kakish*, R. Awad†, R. T. Hagino* and G. P. Clagett* *Division of Vascular Surgery, Department of Surgery, and †Department of Radiology, Department of Veterans Affairs Medical Center, and The University of Texas Southwestern Medical Center, Dallas, Texas, USA The purposes of this study were to examine potential causes for arterial steal syndrome in patients with hemodialysis shunts and to evaluate the results of treatment. A retrospective analysis was performed on all hemodialysis patients who presented with arterial steal syndrome over a 3-year period. Medical records were reviewed to determine demographic data, time to development of symptoms, operative details, and treatment outcome. Results of upper-extremity arteriography in the last five patients were also documented. Eighteen patients presented with steal syndrome at a mean of 7(4) months after shunt placement. Four patients had prior ipsilateral grafts. Fifteen (83%) patients had forearm shunts in loop configuration between the main brachial artery and a superficial vein, and three had straight upper-arm shunts. Graft sizes were 6 mm in five patients and 4- to 7 step grafts in 13. Graft banding was used to treat steal in 11 shunts, one of which remained patent. All five patients who underwent upper-extremity arteriography had arterial stenoses demonstrated in the inflow circulation. In conclusion, arterial steal is an uncommon complication of hemodialysis shunts. Shunt location and choice of graft do not appear to be major factors: steal can develop regardless of access site, configuration, or use of straight versus step graft. Banding is an ineffective method of treatment. Arteriography may be important to detect unsuspected inflow stenoses in patients with steal syndrome. 1997 The International Society for Cardiovascular Surgery Keywords: steal syndrome, hemodialysis, shunt placement, access site
The onset of arterial steal symptoms portends graft loss in patients with forearm hemodialysis shunts. This complication has been reported in 1.6% to 8% of access patients, who experience symptoms ranging from neurosensory and motor changes in the hand to extensive gangrene [1–4]. Treatment options have generally been aimed at reducing blood flow through the shunt. This has been achieved by localized diameter reduction (banding), graft extension, and graft ligation [5–7]. Although graft banding has become a procedure of choice in these patients, reported longterm graft patency is severely compromised [8].
Correspondence to: Dr. R. J. Valentine, Department of Surgery, U.T. Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, Texas 75235-9157, USA
648
Hemodynamically significant stenoses in donor arteries are well-known causes of steal syndrome in other circulations [9, 10]. While previous authors have commented on the possibility of inflow stenoses in access patients with steal syndrome, few have performed arteriography as part of a rigorous search for flow-limiting lesions. The present authors have recently instituted a policy of routine upperextremity arteriography to evaluate the cause of steal symptoms in patients with hemodialysis access. The purposes of this study were: (i) to determine possible causative factors for steal syndrome; (ii) to evaluate the outcome of treatment; and (iii) to document the findings of routine arteriography in the last five patients with steal symptoms.
CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al.
Patients and methods Between 1993 and 1996, 1020 forearm grafts were placed for hemodialysis access at the authors’ institution. The records of all patients who presented with symptoms of hand ischemia following access placement were reviewed. Patients with objective signs of hemodynamic steal were included in the analysis; this was determined by relief of steal symptoms and/or improvement of photoplethysmographically derived digital pressures during temporary shunt occlusion. Medical records were reviewed for demographic data, degree of finger ischemia, and preoperative assessment of arterial and venous circulations. Graft sizes and locations were documented. Time intervals between placement of shunts and development of symptoms were noted, as were interventional diagnostic tests and treatment regimes. Non-invasive tests obtained following treatment were recorded.
syndrome were placed in the forearm, and three (17%) were placed in the upper arm. Three patients had known blood pressure differentials in the upper extremity at the time of graft placement; grafts were placed in the arm with the highest pressure in these cases. The arterial anastomosis was created with the brachial artery in all cases. The venous anastomosis was created with superficial antecubital veins in forearm grafts and with the basilic vein in upper arm grafts. Graft types included straight 6-mm polytetrafluoroethylene (Gore-Tex, Gore Medical Products, Flagstaff, AZ) in five patients and 4- to 7-mm step polytetrafluoroethylene in 13 patients (GoreTex). Hand symptoms included pain and neurosensory deficits in all 18, neuromotor symptoms in two, and tissue loss in three. The mean lowest finger pressure at the time of diagnosis was 32(7) (range 0–84) mmHg. Upper-Extremity arteriography
Results During the study period, 18 (1.8%) hemodialysis patients presented with symptomatic hand steal syndrome at a mean (s.e.m.) of 7(4) (range 0–60) months (median 2 months) after forearm arteriovenous graft placement. Demographics and atherosclerotic risk factors are shown in Table 1. The mean duration of hemodialysis among the 18 patients was 30(9) (range 0–120) months (median 7 months). Nine (50%) patients had previous grafts placed with subsequent thrombosis, and one other patient had a previous primary arteriovenous fistula which occluded. The mean number of prior grafts in the nine patients was two (range one to three). Four (18%) patients had previous grafts in the extremity ipsilateral to the graft which developed steal syndrome. Fifteen (83%) of the grafts which developed steal Table 1 Demographics and atherosclerotic risk factors of the study patients. No. of patients
18
Mean (s.e.m.) age years Race Black White Hispanic No. of males Risk factors Smoking Hypertension Diabetes Dyslipidemia Peripheral vascular disease
61(3) 8 7 3 8
(44) (39) (17) (44)
7 14 14 8 4
(39) (78) (78) (44) (22)
Values in parentheses are percentages (unless stated)
CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
Five patients underwent upper-extremity arteriography to evaluate steal symptoms during the past 12 months. These represented the most recent five patients who presented with steal symptoms following hemodialysis access placement. The mean age of the five patients was 69(3) years. The prevalence of atherosclerotic risk factors and the mean duration of hemodialysis were similar to that noted for the other 13 patients who did not undergo arteriography. One of the five had differential arm blood pressure measurements exceeding 20 mmHg at the time of graft placement, and the graft was placed in the arm with the highest blood pressure, as noted above. Following onset of steal symptoms, all five had lower ipsilateral finger pressures. None of the patients had audible bruits over shoulder or upper-extremity arteries. Arteriography demonstrated arterial stenoses measuring at least 50% diameter loss in the donor arterial circulations of all five patients (Figure 1 and Figure 2). Four of the five patients had stenoses located in the axillosubclavian arteries, one had a mid-brachial artery stenosis, and one had lesions in both locations. The one patient with known blood pressure differentials had bilateral axillary artery stenoses. Treatment Results of treatment are shown in Table 2. Eleven (61%) patients underwent banding procedures. In all cases, graft banding was monitored with intraoperative finger plethysmography. Bands were constructed using large, stainless steel ligating clips (Ligaclip Extra, Ethicon Inc., Summerville, NJ), and grafts were narrowed until finger pressures increased to values above 50 mmHg or until waveforms nor649
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al.
malized, as previously recommended [8]. The mean postoperative finger pressure was 94(11) mmHg in this group. Following banding, 10 patients had resolution of steal symptoms, and one required graft ligation to treat continued hand pain. However, only one of the 10 banded grafts remained patent for longer than 1 week. None of the patients with thrombosed grafts redeveloped steal symptoms. Three patients underwent graft ligations for advanced steal symptoms, including one who had undergone a previous attempt at banding noted above. A fourth patient had resolution of steal symptoms after graft thrombosis and a fifth was lost to follow-up before therapy could be instituted. Two of the five patients with known inflow stenoses underwent percutaneous balloon angioplasty with stent placement (Figure 1 and Figure 2). Both patients had normalization of finger pressures and resolution of symptoms. Of the remaining three patients with inflow stenoses identified angiographically, one had graft ligation without correction of inflow stenosis due to multilevel arterial disease; one failed an attempt to reopen an occluded axillary artery and had resolution of symptoms after graft thrombosis; and one underwent graft banding without correction of inflow lesions. The banded graft thrombosed within 2 days.
Discussion
Figure 1 a Upper extremity arteriogram demonstrating a stenosis at the origin of the left subclavian artery in a patient with steal symptoms. A pressure drop of 27 mmHg was measured across the stenosis. b Completion arteriogram after balloon angioplasty and stent placement to correct the subclavian stenosis. The pressure differential was corrected, and the steal symptoms resolved
650
More than 100,000 patients receive chronic hemodialysis each year in the US, with the majority of these patients expected to live ten years or longer [11]. Graft complications associated with multiple cannulations are common in these patients, especially infection, thrombosis and pseudoaneurysm formation [1, 2, 4, 12]. Although associated vascular disease is prevalent in patients undergoing hemodialysis, circulatory problems associated with arteriovenous shunt formation are rare. The incidence of symptomatic vascular steal syndrome has been reported to occur in 1.6% to 8%, depending on the type of arteriovenous fistula and location of graft [1–4]. The present results confirm that steal syndrome is uncommon in these patients. However, the present findings suggest that steal can occur in shunts regardless of location, configuration, or use of small diameter grafts. The authors cannot determine the influence of graft size on the development of steal per se, since they do not use 8-mm or larger diameter grafts in the creation of shunts and are unaware of any patients with smaller grafts or primary arteriovenous fistulas who developed steal syndrome during the study period. While it has been suggested that use of 4- to 7-mm ‘step’ grafts can decrease the risk of steal syndrome [13], the present study demonstrates that steal complications can occur with these grafts. CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al.
Figure 2 a Upper extremity arteriogram demonstrating multiple sequential stenoses in the mid-brachial artery. A pressure differential of 20 mmHg was measured across the stenosis. b Completion arteriogram following balloon angioplasty and stent placement to correct stenotic areas. The patient’s steal symptoms resolved.
CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
651
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al. Table 2 Treatment of patients with steal syndrome, and outcome. Modality
No. of patients
Outcome
Banding
11
9 occluded 1 continued steal symptoms 1 patent, no symptoms Symptoms resolved Symptoms resolved, grafts patent 1 graft thrombosed, symptoms resolved 1 lost to follow-up
Ligation 3 Angioplasty, stent placement 2 None
2
Total
18
Treatment of steal syndrome has focused on reduction of shunt diameter. In this regard, graft banding has become the procedure of choice in many centers. Unfortunately, patency data following graft banding have not been adequately reported. In the single series reporting late results of this technique, Odland et al. [8] acknowledged that their banded graft patency rates of 62.5% at 6 months and 38.5% at 1 year were inferior compared with their unbanded graft results. The present results suggest that banding is an ineffective method of treating steal syndrome, with a 6-month patency rate of 9% among 11 patients who underwent banding according to the method described by Odland et al. [8]. While symptoms resolved in all of the present patients following banding, 10 of 11 had resolution of symptoms due to graft occlusion. This suggests that simple reduction of graft flow is not adequate to treat steal syndrome. Other causes for the steal syndrome should be sought and corrected. One possibility is unsuspected inflow stenosis. The present authors confirmed the presence of inflow stenoses measuring ⱖ 50% diameter loss in their last five patients, all of whom underwent arteriography. It should be stressed that arterial stenoses were not clinically suspected before shunt placement in any of these patients on the basis of ischemic symptoms or blood pressure differentials. Calligaro et al. [14] have confirmed that equal arterial pressures in the upper extremities do not exclude hemodynamically significant stenoses in the axillosubclavian arterial circulation. Crawford and colleagues [15] demonstrated that 27% of patients with symptoms of upper-extremity ischemia due to arterial disease had normal, symmetrical arm blood pressures. Thus, the presence of normal or equal arm blood pressures does not exclude the possibility of occult inflow stenosis. The present authors suspect that the higher flows associated with the arteriovenous shunt circuit may unmask inflow lesions that would otherwise have borderline hemodynamic significance. Since 652
these lesions may not be suspected on the basis of clinical examination alone, it might be argued that upper-extremity arteriography should be performed in all patients before placement of hemodialysis shunts. However, since the prevalence of steal syndrome is low in patients following placement of hemodialysis shunts, the authors do not recommend routine preoperative arteriography. Duplex ultrasonography may offer a non-invasive alternative to arteriography for preoperative evaluation. However, the low prevalence of steal symptoms in the present patient population suggests that there will be relatively few arterial stenoses identified, and certainly too few to justify the cost of routine Duplex scanning. Based on their initial results, the authors would recommend arteriography in all patients who develop steal syndrome after shunt placement. They speculate that the majority of patients in the present series who did not undergo arteriography also had inflow stenoses. There were no other apparent differences in examined variables between these 13 patients and the last five patients who underwent arteriography. Another possible cause of arterial steal syndrome is inadequate arterial flow in the circulation distal to the vascular access. This may occur, for example, in diabetic patients with distal arteriopathy resulting in inadequate collateral flow to the hand or digits. It should be pointed out that 78% of the present study population was diabetic. Therefore, distal disease may have been a prominent feature of the steal syndrome in many of the patients who did not undergo arteriography. However, since four of the five patients with angiographically identified inflow stenoses were diabetic, the authors submit that arteriography should be considered even in patients who are at risk of having distal disease, in the hope of identifying potentially correctable proximal lesions. What should be done with patients who do not have correctable inflow stenoses? Our data corroborate the results of others [8] that banding is tantamount to graft ligation. While interposing narrow graft segments has theoretical appeal, the fact that steal syndrome occurred in patients with 4- to 7-mm step grafts suggests that this technique may not always be curative. Nevertheless, interposition of a narrowed graft segment (e.g. 4 mm diameter) may be worthwhile, with graft ligation reserved for patients with continued symptoms. The risk of tissue loss appears to be significant in symptomatic patients, so the authors do not recommend observation alone. Despite being a large series of hemodialysis patients with steal syndrome, there are limitations in the present study. While it is interesting to speculate on the possible association between steal syndrome and occult inflow stenoses, a cause-and-effect relationship cannot be established since only two CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
Steal syndrome complicating hemodialysis access: J. D. DeCaprio et al.
patients had correction of inflow lesions. However, it is interesting to note that both patients had relief of steal symptoms and continued graft patency. A second limitation is that this study is retrospective in nature; therefore, a number of patients with steal syndrome may have been missed. While it is acknowledged that the steal prevalence data may be underestimated, it is unlikely that symptomatic patients who remained in the authors’ institution’s hemodialysis population were missed due to the diligence of their nephrology services. The present reported steal prevalence is within the range documented by others [1–4]. However, the authors cannot account for the small proportion of patients who are no longer followed at their institution. Another potential limitation of the present study is that there was reliance on the presence of symptoms to detect patients with steal syndrome. Previous reports suggest that there is a high prevalence of symptomless but hemodynamically significant steal in patient with primary arteriovenous fistulas [16]. Since the diameter of the graft nearly always exceeds the diameter of the artery, it is likely that the great majority of patients have measurable, hemodynamic steal in the ipsilateral extremity. The significance of this is unknown, but given the low prevalence of tissue loss in the present symptomless dialysis population, it is suggested that hemodynamic steal without symptoms is probably benign. The authors speculate that some of our institutions dialysis patients with arterial steal may not have been detected due to their being symptomless or minimally symptomatic. It is possible that these symptomless patients are at risk of shunt occlusion or tissue loss. The authors are currently designing a study to evaluate these issues prospectively. In summary, the present study indicates that arterial steal syndrome is an uncommon complication in hemodialysis patients with upper-extremity shunts. Steal symptoms can occur in patients with either 6-mm or 4- to 7-mm step grafts, regardless of shunt location or configuration. Banding maneuvers to decrease the diameter of the shunt do not appear to be effective in treating steal syndrome, suggesting that the cause of symptoms is not related to the size of the graft. Early results with routine arteriography in steal patients suggest that a search for inflow stenosis may be rewarding. The authors recommend routine arteriography in patients who develop steal symptoms after placement of hemodialysis shunts in order to detect potentially correctable lesions.
CARDIOVASCULAR SURGERY
DECEMBER 1997 VOL 5 NO 6
References 1. Haimov, M., Baez, H., Neff, M. and Slifkin, R., Complications of arteriovenous fistulas for hemodialysis. Archives of Surgery, 1975, 110, 708–712. 2. Zibari, G. B., Rohr, M. S. and Landreneau, M. D. et al., Complications from permanent hemodialysis vascular access. Surgery, 1988, 104, 681–686. 3. Kwun, K. B., Schanzer, H., Finkler, N., Haimov, M. and Burrows, L., Hemodynamic evaluation of angioaccess procedures for hemodialysis. Vascular Surgery, 1979, 13, 170–177. 4. Ballard, J., Blunt, T. J. and Malone, J., Major complications of angioaccess surgery. American Journal of Surgery, 1992, 164, 229–232. 5. Mattson, W., Recognition and treatment of vascular steal secondary to hemodialysis prosthesis. American Journal of Surgery, 1987, 154, 198–201. 6. Rivers, S. P., Scher, L. A. and Veith, F. J., Correction of steal syndrome secondary to hemodialysis access fistulas: a simplified quantitative technique. Surgery, 1992, 112, 593–597. 7. West, J. C., Evans, R. D. and Kelley, S. E. et al., Arterial insufficiency in hemodialysis access procedures: reconstruction by an interposition polytetrafluoro-ethylene graft conduit. American Journal of Surgery, 1987, 153, 300–301. 8. Odland, M. D., Kelley, P. H., Ney, A. L., Andersen, R. C. and Bubrick, M. P., Management of dialysis associated steal syndrome complicating upper extremity arteriovenous fistulas: Use of intraoperative digital photoplethysmography. Surgery, 1991, 110, 664–670. 9. Flanigan, D. P., Pratt, D. G., Goodreau, J. J., Burnham, S. J., Yao, J. S. T. and Bergan, J. J., Hemodynamic and angiographic guidelines in the selection of patients for femorofemoral bypass. Archives of Surgery, 1978, 113, 1257–1262. 10. Fisher, C. M., A new vascular syndrome–the subclavian steal. New England Journal of Medicine, 1961, 265, 912. 11. Bennion, R. S., Socioeconomic implications of vascular access surgery. In Vascular Access: Principles and Practice, ed. S. E. Wilson. Mosby-Year Book Inc., St. Louis, 1996, pp. 284–293. 12. Hill, S. L. and Donato, A. T., Complications of dialysis access; A six-year study. American Journal of Surgery, 1991, 162, 265–267. 13. Rosental, J. J., Bell, D. D., Gaspar, M. R., Mobius, H. J. and Lemire, G. G., Prevention of high flow problems of arteriovenpous grafts: Devlopment of a new tapered graft. American Journal of Surgery, 1980, 140, 231–233. 14. Calligaro, K. D., Ascer, E., Veith, W., Gupta, S. K., Wengerter, K. R., Franco, C. D., Bakal, C. W. and Sprayregen, S., Unsuspected inflow disease in candidates for axillofemoral bypass operations: A prospective study. Journal of Vascular Surgery, 1990, 11, 832–837. 15. Crawford, E. S., De Bakey, M. E., Morris, G. C. and Colley, D. A., Thrombo-obliterative disease of the great vessels arising from the aortic arch. Journal of Thoracic Cardiovascular Surgery, 1962, 43, 38–53. 16. Duncan, H., Ferguson, L. and Faris, I., Incidence of the radial steal syndrome in patients with Brescia fistula for hemodialysis: Its clinical significance. Journal of Vascular Surgery, 1986, 4, 144–147.
Paper accepted 9 June 1997
653