- Email: [email protected]
Recognition and treatment of vascular steal secondary to hemodialysis prostheses
Recognition and Treatment of Vascular Steal Secondary to Hemodialysis Prostheses
William J. Maltson, MD, Rapid City, South Dakota
Use of polytetrafluoroethylene prostheses for longterm hemodialysis began in the latter half of the 1970s and progressively has become the most popular form of access [I-3]. Along with the benefits of this prosthesis are its known complications, notably thrombosis, hemorrhage, and infection. Less commonly encountered complications are cardiac failure, false aneurysms, carpal tunnel syndrome, and arterial insufficiency to the distal part. The purpose of this report is to center on arterial insufficiency or steal syndrome, particularly in relation to clinical recognition and treatment. Patients and Methods Between 1978 and 1986,61 patients underwent placement of 76 polytetrafluoroethylene prostheses. Eight prostheses (10.5 percent) clotted within 1 week and were excluded from the study. Three thigh prostheses were also excluded. The resulting 55 patients with 65 functioning arm or forearm prostheses make up the data base of this report. Twenty-one patients, or approximately a third of the group, were diabetic. Forty-four forearm prostheses were constructed, 34 of which were straight 8 mm diameter grafts, with the remainder being looped grafts 8 to 6 mm in diameter. Twenty-one upper arm prostheses were constructed. All upper arm grafts were either 6 mm or tapered prostheses of 5 to 8 or 4 to 7 mm as popularized by Rosental and coworkers [41. Steal syndrome was defined as pain and coolness of the distal part in which an arteriovenous communication is created, accompanied by absence or marked diminution of digital pulse wave as measured by plethysmography and in which the pulse wave augmented with compression of the communication. This concept was first described
From the University of South Dakota, School of Medicine, West River Campus, Rapid City, South Dakota. Requests for reprints should be addressed to William J. Mattson, MD, University of South Dakota School of Medicine, West River Campus, Box 2623, Rapid City, South Dakota 57709. Presented at the 15th Annual Meeting of the Society for Clinical Vascular Surgery, Scottsdale, Arizona, March 25-29, 1967.
196
by Bussell et al [5] in 1971 using pneumatic plethysmography. It was believed that signs and symptoms of vascular steal syndrome developed secondary to placement of a polytetrafluoroethylene hemodialysis prosthesis in four patients (7 percent). Three of the four patients were diabetic. In one patient, steal developed in each of his two forearms where prostheses were placed. Therefore, five prostheses had development of steal syndrome. No corrective treatment was forthcoming for the patient with bilateral steal. The second patient was a 70 year old diabetic woman in whom steal syndrome was verified by photoplethysmography and was treated by ligation of the prosthesis. The last two patients were a 59 year old diabetic man and a 70 year old nondiabetic man with a diagnosis of vascular steal syndrome documented by digital photoplethysmography before and during graft compression. The diagnosis and treatment were essentially identical in both of these patients and are described herein. Statistical analysis utilized chi-square comparison of two independent proportions. A cuffed photoplethysmograph transducer is placed on the contralateral thumb of the limb in question for vascular steal, calibrating the arterial mode of the photoplethysmograph and chart recorder to optimal amplitude. The thumb, index, and middle finger of the affected extremity are then tested. A flat or markedly attenuated signal is the expected finding. Manual compression of the proximal portion of the prosthesis is then performed while observing the plethysmographic waveform. A substantial increase in amplitude constitutes a positive test result and ensures patency of the radial artery and the palmar arcade. Depending on the degree of symptoms, operative correction is then undertaken. Operative banding of the proximal end of the polytetrafluoroethylene prosthesis is begun by again calibrating the plethysmograph using the contralateral digit, preferably the thumb. The transducer is then applied to the thumb of the affected limb and block preparation is carried out over the proximal end of the prosthesis. Under local anesthesia, the proximal prosthesis is mobilized. A band prepared from a polytetrafluoroethylene cardiovaszular patch is then passed around the mobilized prosthe-
The American Journal of Surgery
VascularSteal Secondary to HemodialysisProstheses
Figure 1. A, crossed polytetrafluoroethylene
or shoelace band. B,
of band and appfkatton of iekstat~ clip. C, and D,-&aratton of the band from a polytetrafluoroethylene cardiovascular patch.
sis, either crossing the tails of the band in a shoelace fashion or bringing them around straight circumferentially (Figure 1A and B). The first pair of tails is tightened using a vascular or right-angle clamp until the desired amplitude of the plethysmographic waveform is obtained. Once this is accomplished, the two abutted ends are secured with a hemostatic clip of appropriate length (a medium-large Liga clip works well). The second set of laces is then tightened in a similar manner to fine-tune the desired amplitude (Figure 1B). The tails of the band are then trimmed just above the hemostatic clip, and the clip is held in place with a figure 8 suture of 5-O cardiovascular polypropylene (Figure 1A and B). The band is prepared from a polytetrafluoroethylene cardiovascular patch by cutting a rectangular piece 1 cm in width and 7 to 10 cm in length (Figure 1C). This band is then cut again lengthwise into four 0.5 cm tails, leaving a center bridge of 0.5 to 0.75 cm (Figure 1D). The wound is closed in the fashion of the operator. Photoplethysmographic tracings can be taken in the recovery room and at bedside the following day. Results
One patient, a 49 year old diabetic man, was an alcoholic and heavy cigarette smoker. Despite admonitions, he continued to smoke and eventually had bilateral forearm amputations. Another patient, a 70 year old diabetic woman, underwent ligation of her prosthesis, with resultant restoration of digital pulse wave and alleviation of symptoms. A 60 year old diabetic man initially did well after banding of his 8 mm looped forearm prosthesis. The prosthesis was banded according to the aforementioned shoelace technique (Figure lA), but the hemostatic clip was not secured with a suture. This patient’s symptoms recurred about 1 week postoperatively. A flat plethysmographic tracing of the
Volume 154, August 1997
c
involved thumb was found, which again augmented with digital compression over the graft. At reoperation, the old band was replaced with a new one in an identical manner, except the hemostatic clip was held in place with a fiie 8 suture of 5-O cardiovascular polypropylene. The last patient, a 70 year old nondiabetic man, underwent banding of a 6 to 8 mm looped forearm prosthesis in a similar fashion but without crossing the tails (Figure 1B). Instead, they were brought circumferentially around the prosthesis and initially secured with both a hemostatic clip and a suture. The patient did well without any recurrent of his symptoms. Both of the latter patients obtained a lasting good result. comments The differential diagnosis of vascular steal syndrome includes the neuropathies of uremia and diabetes, in addition to symptoms attributable to secondary hyperparathyroidism. More recently, carpal tunnel syndrome has been recognized as an uncommon but distinct condition that occurs in the hemodialysis patient [6,4. Clinical presentation consists of symptoms similar to vascular steal syndrome, such as pain, paresthesias, and thenar atrophy of the hand. Diagnosis is verified by electromyographic techniques. The condition is divided into an early and late type, with diabetes being more common in the early form. For the most part, studies on this subject are not clear as to how vascular steal syndrome is ruled out. The suggestion that the two conditions can coexist in the same extremity has been made by DePalma et al [S]. Increasing evidence indicates that the incidence of steal may be dependent upon two subsets of patients: diabetic and nondiabetic. Humphreys et al
199
Mattson
[I] found an extremely low incidence of steal in over 1,000 patients on long-term hemodialysis but also commented that their patient population contained very few diabetics. Haimov et al [9] have described eight instances (1.6 percent) of symptomatic steal in 516 arteriovenous fistulas of various types. Of these eight, four (50 percent) were in diabetics. Unfortunately, the investigators failed to state what number of the total group were diabetic. Of the 55 patients presented in the present report, 21 (38 percent) were diabetic, of which 3 (14.2 percent) had development of clinical steal. Of the remaining 34 patients, steal developed in 1 (3 percent). Chi-square analysis did not show a statistical difference between these two subsets (p >0.05). Looking at the same data with regard to the number of limbs at risk, that is, the chance of steal each time an access is created, the data do reach statistical significance. Forty primary polytetrafluoroethylene prostheses were constructed in nondiabetic limbs, of which only 1 developed a steal. On the other hand, of 25 diabetic limbs in which a polytetrafluoroethylene prosthesis was placed, 4 developed a clinical steal (chi-square 3.949, p = 0.0444). The numbers are small and have not been subjected to more stringent analysis with correction for continuity, but the suggestion remains that the diabetic patient is at increased risk for development of the steal syndrome. In the strict sense, all patients in whom an arteriovenous communication has been constructed have some degree of diversion of blood from the periphery and thereby have a vascular steal. Bussel et al [5] found that all eight patients with a radial artery-tocephalic vein shunt constructed at the wrist had a diminution of ipsilateral digital blood flow. Only one of these eight patients was symptomatic. The diagnosis of clinical steal is made by symptoms of pain and coolness of the distal part (generally the fingers and hand), sometimes accompanied by numbness and trophic changes. The diagnosis is confirmed by lack of or marked diminution of the pulse wave pattern of the digits, particularly the thumb, which augments with compression of the fistulous communication. Reversal of flow in the radial artery per se does not make the diagnosis, nor does radial or ulnar pressure as measured by Doppler analysis. When symptoms warrant, treatment of this condition consists of decreasing or abolishing blood flow through this arteriovenous communication. With regard to the polytetrafluoroethylene prosthesis, Rosental et al [4] have advocated prevention of steal with their tapered graft. From a technical viewpoint, performing a 4 or 5 mm anastomosis from the arterial end of a prosthesis to an arteriosclerotic artery can be difficult, resulting in a high rate of early thrombosis. In the present series reviewed, two patients sustained this complication. In addition, in the majority of patients the tapered
200
graft is not needed. A variety of banding and plicating procedures have been advocated with varying success [2,8,10,11]. Rubio and Farrell [12] have popularized an interposition graft for the treatment of steal syndrome. The drawback of this method is that there is no way of quantitating the degree of narrowing during the time of operation so as to correct the steal and, at the same time, prevent thrombosis. In 1981, Ebeid and Saranchak [.23] reported banding of a polytetrafluoroethylene prosthesis for the treatment of steal syndrome. Their procedure was similar to the method presented herein, with the exception that intraoperative monitoring of the degree of constriction was lacking. The investigators conceded that the procedure calls for constant monitoring in order to obtain an optimal result. A most important point brought out in their study, however, was that the banding should be carried out over a fairly long distance of 1 cm or more, rather than by simple ligature stenosis. This concept relies primarily on Poiseuille’s law, which states that the pressure decrease or resistance within a conduit with fluid of laminar flow is directly proportional to the length of that conduit and inversely proportional to the fourth power of the radius. In plain terms, when the band is relatively wide, greater control can be achieved in obtaining the desired resistance within the prosthesis. A narrow stenosis, as achieved by ligature, produces little decrease in flow until critical stenosis is reached, then produces an exponential increase in resistance. The turbulence created promotes a greater tendency for thrombosis. These hemodynamic concepts have been succinctly reviewed by Barnes [14]. Summary Ischemic symptoms of the distal extremity devel,oped in 4 patients from a group of 55 in whom a primary polytetrafluoroethylene prosthesis was placed for purpose of hemodialysis. Diagnosis of vascular steal syndrome was suggested by clinical symptoms of pain and coolness of the distal part, as well as trophic changes of the hand. The diagnosis was confirmed by placing a cuffed photoplethysmograph transducer on one or more of the digits of the affected limb and recording the waveform before and during manual compression of the arterial end of the prosthesis. The findings showed essentially a flat waveform converting to pulsatile waveform when the proximal graft was compressed. Two patients underwent controlled operative banding of the arterial end of the polytetrafluoroethylene prosthesis using this same method. Apolytetrafluoroethylene cardiovascular patch was cut to a width of 1 cm and then cut again to form two shoelace tails each 0.5 cm in width. The band was placed around the proximal part of the prosthesis and progressively tightened until an adequate digi-
The American Journal of Surgery
Vascular
tal pulse wave returned. Preservation of flow through the prosthesis was maintained so as to allow dialysis and prevent thrombosis. The technical aspects of the procedure have been discussed. Vascular steal syndrome appears to be more common in the diabetic.
6.
7.
8.
References 9. 1. Humphries AL Jr, Nesbit RR Jr, Caruana RJ, Hutchins RS, Heimburger RA, Wray CH. Thirty-six recommendations for vascular access operations: lessons learned from our first thousand operations. Am Surg 1981; 47: 145-51. 2. Tellis VA, Kohlberg WE, Bhat DJ, et al. Expanded polytetrafluoroethylene graft fistulas for chronic hemodialysis. Ann Surg 1979; 189: 101-5. 3. Hurt AV, Batello-Cruz M, Skipper BJ, Teaf SR, Sterling WA. Bovine carotid artery heterografts versus polytetrafluoroethylene grafts. Am J Surg 1983; 146: 644-7. 4. Rosental JJ, Bell RD, Gaspar MR, Lemire GG. Prevention of high flow problems of arteriovenous grafts. Am J Surg 1980; 140: 231-3. 5. Bussell JA, Abbott JA, Lim RC. A radial steal syndrome with
Volume 154, August lgS7
10.
11. 12. 13.
14.
Steal Secondary to Hemodialysis
Prostheses
arteriovenous fistula for hemodialysis. Ann Intern Med 1971; 75: 387-94. Spertini P, Wauters JP, Poulenas I. Carpal tunnel syndrome: a frequent, invalidating, long-term complication of chronic hemodialysis. Clin Nephrol 1984; 21: 98-101. Zamora JL, Rose JE, Rosario V, Noon GP. Hemodialysisassociated carpal tunnel syndrome. A clinical review. Nephron 1985; 41: 70-4. DePalma JR. Vannix R, Bahuth J. Abukurah A. Steal syndrome, ischemia, congestive failure, and peripheral neuropathy. Proc Clin Dial Transplant Forum 1973; 3: 31 l-9. Haimov M, Baez A, Neff M, Slifkin Ft. Complications of arteriovenous fistulas for hemodialysis. Arch Surg 1975; 110: 708-12. Anderson CB, Grace MA. Banding of arteriovenous dialysis fistulas to correct high-output cardiac failure. Surgery 1975; 78: 552-4. Fee HJ, Golding AL. Lower extremity ischemia after femoral arteriovenous bovine shunts. Ann Surg 1976; 183: 42-5. Rubio PA, Farrell EM Atlas of angioaccess surgery. Chicago: Year Book Medical Publishers, 1983: 238-45. Ebeib A, Saranchak HJ. Banding of a PTFE hemodialysis fistula in the treatment of steal syndrome. Clin Exp Dial Apheresis 1981; 5(3): 251-7. Barnes JW. Hemodynamlcs for the vascular surgeon. Arch Surg 1980; 115: 216-23.
201
William J. Maltson, MD, Rapid City, South Dakota
Use of polytetrafluoroethylene prostheses for longterm hemodialysis began in the latter half of the 1970s and progressively has become the most popular form of access [I-3]. Along with the benefits of this prosthesis are its known complications, notably thrombosis, hemorrhage, and infection. Less commonly encountered complications are cardiac failure, false aneurysms, carpal tunnel syndrome, and arterial insufficiency to the distal part. The purpose of this report is to center on arterial insufficiency or steal syndrome, particularly in relation to clinical recognition and treatment. Patients and Methods Between 1978 and 1986,61 patients underwent placement of 76 polytetrafluoroethylene prostheses. Eight prostheses (10.5 percent) clotted within 1 week and were excluded from the study. Three thigh prostheses were also excluded. The resulting 55 patients with 65 functioning arm or forearm prostheses make up the data base of this report. Twenty-one patients, or approximately a third of the group, were diabetic. Forty-four forearm prostheses were constructed, 34 of which were straight 8 mm diameter grafts, with the remainder being looped grafts 8 to 6 mm in diameter. Twenty-one upper arm prostheses were constructed. All upper arm grafts were either 6 mm or tapered prostheses of 5 to 8 or 4 to 7 mm as popularized by Rosental and coworkers [41. Steal syndrome was defined as pain and coolness of the distal part in which an arteriovenous communication is created, accompanied by absence or marked diminution of digital pulse wave as measured by plethysmography and in which the pulse wave augmented with compression of the communication. This concept was first described
From the University of South Dakota, School of Medicine, West River Campus, Rapid City, South Dakota. Requests for reprints should be addressed to William J. Mattson, MD, University of South Dakota School of Medicine, West River Campus, Box 2623, Rapid City, South Dakota 57709. Presented at the 15th Annual Meeting of the Society for Clinical Vascular Surgery, Scottsdale, Arizona, March 25-29, 1967.
196
by Bussell et al [5] in 1971 using pneumatic plethysmography. It was believed that signs and symptoms of vascular steal syndrome developed secondary to placement of a polytetrafluoroethylene hemodialysis prosthesis in four patients (7 percent). Three of the four patients were diabetic. In one patient, steal developed in each of his two forearms where prostheses were placed. Therefore, five prostheses had development of steal syndrome. No corrective treatment was forthcoming for the patient with bilateral steal. The second patient was a 70 year old diabetic woman in whom steal syndrome was verified by photoplethysmography and was treated by ligation of the prosthesis. The last two patients were a 59 year old diabetic man and a 70 year old nondiabetic man with a diagnosis of vascular steal syndrome documented by digital photoplethysmography before and during graft compression. The diagnosis and treatment were essentially identical in both of these patients and are described herein. Statistical analysis utilized chi-square comparison of two independent proportions. A cuffed photoplethysmograph transducer is placed on the contralateral thumb of the limb in question for vascular steal, calibrating the arterial mode of the photoplethysmograph and chart recorder to optimal amplitude. The thumb, index, and middle finger of the affected extremity are then tested. A flat or markedly attenuated signal is the expected finding. Manual compression of the proximal portion of the prosthesis is then performed while observing the plethysmographic waveform. A substantial increase in amplitude constitutes a positive test result and ensures patency of the radial artery and the palmar arcade. Depending on the degree of symptoms, operative correction is then undertaken. Operative banding of the proximal end of the polytetrafluoroethylene prosthesis is begun by again calibrating the plethysmograph using the contralateral digit, preferably the thumb. The transducer is then applied to the thumb of the affected limb and block preparation is carried out over the proximal end of the prosthesis. Under local anesthesia, the proximal prosthesis is mobilized. A band prepared from a polytetrafluoroethylene cardiovaszular patch is then passed around the mobilized prosthe-
The American Journal of Surgery
VascularSteal Secondary to HemodialysisProstheses
Figure 1. A, crossed polytetrafluoroethylene
or shoelace band. B,
of band and appfkatton of iekstat~ clip. C, and D,-&aratton of the band from a polytetrafluoroethylene cardiovascular patch.
sis, either crossing the tails of the band in a shoelace fashion or bringing them around straight circumferentially (Figure 1A and B). The first pair of tails is tightened using a vascular or right-angle clamp until the desired amplitude of the plethysmographic waveform is obtained. Once this is accomplished, the two abutted ends are secured with a hemostatic clip of appropriate length (a medium-large Liga clip works well). The second set of laces is then tightened in a similar manner to fine-tune the desired amplitude (Figure 1B). The tails of the band are then trimmed just above the hemostatic clip, and the clip is held in place with a figure 8 suture of 5-O cardiovascular polypropylene (Figure 1A and B). The band is prepared from a polytetrafluoroethylene cardiovascular patch by cutting a rectangular piece 1 cm in width and 7 to 10 cm in length (Figure 1C). This band is then cut again lengthwise into four 0.5 cm tails, leaving a center bridge of 0.5 to 0.75 cm (Figure 1D). The wound is closed in the fashion of the operator. Photoplethysmographic tracings can be taken in the recovery room and at bedside the following day. Results
One patient, a 49 year old diabetic man, was an alcoholic and heavy cigarette smoker. Despite admonitions, he continued to smoke and eventually had bilateral forearm amputations. Another patient, a 70 year old diabetic woman, underwent ligation of her prosthesis, with resultant restoration of digital pulse wave and alleviation of symptoms. A 60 year old diabetic man initially did well after banding of his 8 mm looped forearm prosthesis. The prosthesis was banded according to the aforementioned shoelace technique (Figure lA), but the hemostatic clip was not secured with a suture. This patient’s symptoms recurred about 1 week postoperatively. A flat plethysmographic tracing of the
Volume 154, August 1997
c
involved thumb was found, which again augmented with digital compression over the graft. At reoperation, the old band was replaced with a new one in an identical manner, except the hemostatic clip was held in place with a fiie 8 suture of 5-O cardiovascular polypropylene. The last patient, a 70 year old nondiabetic man, underwent banding of a 6 to 8 mm looped forearm prosthesis in a similar fashion but without crossing the tails (Figure 1B). Instead, they were brought circumferentially around the prosthesis and initially secured with both a hemostatic clip and a suture. The patient did well without any recurrent of his symptoms. Both of the latter patients obtained a lasting good result. comments The differential diagnosis of vascular steal syndrome includes the neuropathies of uremia and diabetes, in addition to symptoms attributable to secondary hyperparathyroidism. More recently, carpal tunnel syndrome has been recognized as an uncommon but distinct condition that occurs in the hemodialysis patient [6,4. Clinical presentation consists of symptoms similar to vascular steal syndrome, such as pain, paresthesias, and thenar atrophy of the hand. Diagnosis is verified by electromyographic techniques. The condition is divided into an early and late type, with diabetes being more common in the early form. For the most part, studies on this subject are not clear as to how vascular steal syndrome is ruled out. The suggestion that the two conditions can coexist in the same extremity has been made by DePalma et al [S]. Increasing evidence indicates that the incidence of steal may be dependent upon two subsets of patients: diabetic and nondiabetic. Humphreys et al
199
Mattson
[I] found an extremely low incidence of steal in over 1,000 patients on long-term hemodialysis but also commented that their patient population contained very few diabetics. Haimov et al [9] have described eight instances (1.6 percent) of symptomatic steal in 516 arteriovenous fistulas of various types. Of these eight, four (50 percent) were in diabetics. Unfortunately, the investigators failed to state what number of the total group were diabetic. Of the 55 patients presented in the present report, 21 (38 percent) were diabetic, of which 3 (14.2 percent) had development of clinical steal. Of the remaining 34 patients, steal developed in 1 (3 percent). Chi-square analysis did not show a statistical difference between these two subsets (p >0.05). Looking at the same data with regard to the number of limbs at risk, that is, the chance of steal each time an access is created, the data do reach statistical significance. Forty primary polytetrafluoroethylene prostheses were constructed in nondiabetic limbs, of which only 1 developed a steal. On the other hand, of 25 diabetic limbs in which a polytetrafluoroethylene prosthesis was placed, 4 developed a clinical steal (chi-square 3.949, p = 0.0444). The numbers are small and have not been subjected to more stringent analysis with correction for continuity, but the suggestion remains that the diabetic patient is at increased risk for development of the steal syndrome. In the strict sense, all patients in whom an arteriovenous communication has been constructed have some degree of diversion of blood from the periphery and thereby have a vascular steal. Bussel et al [5] found that all eight patients with a radial artery-tocephalic vein shunt constructed at the wrist had a diminution of ipsilateral digital blood flow. Only one of these eight patients was symptomatic. The diagnosis of clinical steal is made by symptoms of pain and coolness of the distal part (generally the fingers and hand), sometimes accompanied by numbness and trophic changes. The diagnosis is confirmed by lack of or marked diminution of the pulse wave pattern of the digits, particularly the thumb, which augments with compression of the fistulous communication. Reversal of flow in the radial artery per se does not make the diagnosis, nor does radial or ulnar pressure as measured by Doppler analysis. When symptoms warrant, treatment of this condition consists of decreasing or abolishing blood flow through this arteriovenous communication. With regard to the polytetrafluoroethylene prosthesis, Rosental et al [4] have advocated prevention of steal with their tapered graft. From a technical viewpoint, performing a 4 or 5 mm anastomosis from the arterial end of a prosthesis to an arteriosclerotic artery can be difficult, resulting in a high rate of early thrombosis. In the present series reviewed, two patients sustained this complication. In addition, in the majority of patients the tapered
200
graft is not needed. A variety of banding and plicating procedures have been advocated with varying success [2,8,10,11]. Rubio and Farrell [12] have popularized an interposition graft for the treatment of steal syndrome. The drawback of this method is that there is no way of quantitating the degree of narrowing during the time of operation so as to correct the steal and, at the same time, prevent thrombosis. In 1981, Ebeid and Saranchak [.23] reported banding of a polytetrafluoroethylene prosthesis for the treatment of steal syndrome. Their procedure was similar to the method presented herein, with the exception that intraoperative monitoring of the degree of constriction was lacking. The investigators conceded that the procedure calls for constant monitoring in order to obtain an optimal result. A most important point brought out in their study, however, was that the banding should be carried out over a fairly long distance of 1 cm or more, rather than by simple ligature stenosis. This concept relies primarily on Poiseuille’s law, which states that the pressure decrease or resistance within a conduit with fluid of laminar flow is directly proportional to the length of that conduit and inversely proportional to the fourth power of the radius. In plain terms, when the band is relatively wide, greater control can be achieved in obtaining the desired resistance within the prosthesis. A narrow stenosis, as achieved by ligature, produces little decrease in flow until critical stenosis is reached, then produces an exponential increase in resistance. The turbulence created promotes a greater tendency for thrombosis. These hemodynamic concepts have been succinctly reviewed by Barnes [14]. Summary Ischemic symptoms of the distal extremity devel,oped in 4 patients from a group of 55 in whom a primary polytetrafluoroethylene prosthesis was placed for purpose of hemodialysis. Diagnosis of vascular steal syndrome was suggested by clinical symptoms of pain and coolness of the distal part, as well as trophic changes of the hand. The diagnosis was confirmed by placing a cuffed photoplethysmograph transducer on one or more of the digits of the affected limb and recording the waveform before and during manual compression of the arterial end of the prosthesis. The findings showed essentially a flat waveform converting to pulsatile waveform when the proximal graft was compressed. Two patients underwent controlled operative banding of the arterial end of the polytetrafluoroethylene prosthesis using this same method. Apolytetrafluoroethylene cardiovascular patch was cut to a width of 1 cm and then cut again to form two shoelace tails each 0.5 cm in width. The band was placed around the proximal part of the prosthesis and progressively tightened until an adequate digi-
The American Journal of Surgery
Vascular
tal pulse wave returned. Preservation of flow through the prosthesis was maintained so as to allow dialysis and prevent thrombosis. The technical aspects of the procedure have been discussed. Vascular steal syndrome appears to be more common in the diabetic.
6.
7.
8.
References 9. 1. Humphries AL Jr, Nesbit RR Jr, Caruana RJ, Hutchins RS, Heimburger RA, Wray CH. Thirty-six recommendations for vascular access operations: lessons learned from our first thousand operations. Am Surg 1981; 47: 145-51. 2. Tellis VA, Kohlberg WE, Bhat DJ, et al. Expanded polytetrafluoroethylene graft fistulas for chronic hemodialysis. Ann Surg 1979; 189: 101-5. 3. Hurt AV, Batello-Cruz M, Skipper BJ, Teaf SR, Sterling WA. Bovine carotid artery heterografts versus polytetrafluoroethylene grafts. Am J Surg 1983; 146: 644-7. 4. Rosental JJ, Bell RD, Gaspar MR, Lemire GG. Prevention of high flow problems of arteriovenous grafts. Am J Surg 1980; 140: 231-3. 5. Bussell JA, Abbott JA, Lim RC. A radial steal syndrome with
Volume 154, August lgS7
10.
11. 12. 13.
14.
Steal Secondary to Hemodialysis
Prostheses
arteriovenous fistula for hemodialysis. Ann Intern Med 1971; 75: 387-94. Spertini P, Wauters JP, Poulenas I. Carpal tunnel syndrome: a frequent, invalidating, long-term complication of chronic hemodialysis. Clin Nephrol 1984; 21: 98-101. Zamora JL, Rose JE, Rosario V, Noon GP. Hemodialysisassociated carpal tunnel syndrome. A clinical review. Nephron 1985; 41: 70-4. DePalma JR. Vannix R, Bahuth J. Abukurah A. Steal syndrome, ischemia, congestive failure, and peripheral neuropathy. Proc Clin Dial Transplant Forum 1973; 3: 31 l-9. Haimov M, Baez A, Neff M, Slifkin Ft. Complications of arteriovenous fistulas for hemodialysis. Arch Surg 1975; 110: 708-12. Anderson CB, Grace MA. Banding of arteriovenous dialysis fistulas to correct high-output cardiac failure. Surgery 1975; 78: 552-4. Fee HJ, Golding AL. Lower extremity ischemia after femoral arteriovenous bovine shunts. Ann Surg 1976; 183: 42-5. Rubio PA, Farrell EM Atlas of angioaccess surgery. Chicago: Year Book Medical Publishers, 1983: 238-45. Ebeib A, Saranchak HJ. Banding of a PTFE hemodialysis fistula in the treatment of steal syndrome. Clin Exp Dial Apheresis 1981; 5(3): 251-7. Barnes JW. Hemodynamlcs for the vascular surgeon. Arch Surg 1980; 115: 216-23.
201