Experience with Dacron Graft Arteriovenous Fistulas for Dialysis Access JAMES F. BURDICK, M.D., WILLIAM SCOTT, M.D., A. BENEDICT COSIMI, M.D.
The technique, hemodynamics, and results of a new, more successful Filamentous Velour Dacron graft dialysis fistula between brachial artery and axillary vein are described. Hemodialysis has been successfully instituted in 15/16 patients. One graft was removed for infection and one for overlying skin erosion. There have been five episodes of thrombosis, all with succegsful restoration to function. After a median nine-month follow-up, nine of 16 grafts have been complication-free and 13 of 16 remain functional or functioned until the time of death. Hemodynamic studies revealed that with high flows (even 1900 cc/min), there was no evidence of cardiac failure or peripheral "steal." One patient with low flow (80% of these grafts should provide long-term dialysis access for this difficult patient population. T HERE IS NOW GENERAL ACCEPTANCE that the
Brescio-Cimino radiocephalic wrist fistula is the best access procedure for chronic dialysis. However, there are many long-term dialysis patients who no longer have vessels suitable for this fistula. These patients require some secondary procedure. Saphenous vein autografts8"2 have been reasonably successful when possible but these have a tendency to eventually develop aneurysmal dilatation, and the extensive dissection required for harvesting increases the morbidity of this procedure. The other alternative with which there has been a large experience utilizes the modified bovine carotid xenograft. Although the bovine grafts have apparently been employed quite successfully for dialysis access in some series,'4'5'20'27 there have been many reports of distressingly high rates of thrombosis, infection and aneurysm formation in these fistulas.3'5'7'12,16,22-24 Dissatisfaction with our bovine graft fistulas because of frequent thromboses and suture-line aneurysms convinced us that a better alternative is needed for patients whose vessels are not suitable for construcSupported in part by the USPHS Training Grant (lT32AI702701) and USCI, a division of C.R. Bard, Inc. Submitted for publication July 5, 1977.
From the Transplantation-Immunology Unit of the General Surgical Services, Massachusetts General Hospital, and the Department of Surgery, Harvard Medical School, Boston, Massachusetts
tion of a wrist fistula. Fistulas of expanded polytetraflouroethylene3 may provide results which are somewhat better than the bovine experience but data on these, as well as on the use of saphenous vein allografts,2' are as yet preliminary. There is experimental and initial clinical evidence that standard internal velour dacron provides an acceptable graft for dialysis access fistulas.18 The Filamentous Velour dacron graft should theoretically be an even better material for dialysis access. Studies with experimental grafts in the dog and other species have shown that the transmural fibrillar structure of this material encourages transmural tissue ingrowth25.26 which should improve resistance to the regular punctures required for dialysis. Brener has had success with this material for dialysis access.6 In view of the problems regularly associated with the previously described secondary access procedures, it was decided to investigate the use of Filamentous Velour dacron grafts for construction of vascular access in a series of dialysis patients who required such a secondary procedure. It was elected to employ a 5 or 6 mm diameter graft in the upper arm and thus a relatively high flow through the fistula was anticipated. Since ischemic symptoms distal to upper arm fistulas have been reported''24 the hemodynamic consequences of these shunts were evaluated in many of the patients. Materials and Methods
Sixteen chronic dialysis patients ranging in age from 30 to 70, in whom a forearm Brescio-Cimino fistula was not possible, were chosen to receive a dacron graft. A history of multiple failed access procedures
0003-4932-78-0300-0262-0075 X J. B. Lippincott Company
262
Vol. 187 . No. 3
DACRON GRAFT EXPERIENCE
was common; three patients were diabetics. Systemic infection was a contraindication and significant arteriosclerosis in the proposed extremity was ruled out preoperatively by physical exam and vascular laboratory evaluation. General anesthesia was usually employed but the procedure has been successfully performed in four patients under local or regional anesthesia. In each instance a Filamentous Velour dacron graft* was implanted end-to-side between the brachial artery just above the antecubital fossa and the axillary vein.4 The graft was 6 mm in diameter in all but two patients who received 5 mm grafts and the length of the graft ranged from 14 to 25 cm. The operative procedure is depicted in Figure 1. A transverse incision is made in the distal axilla starting at the anterior axillary fold. The axillary vein is exposed at the junction of the basilic and brachial veins (Fig. 1 A). The brachial artery is exposed via a "hockey-stick" incision with its transverse component just above the antecubital fold. A longitudinal counterincision equidistant between these is placed laterally and the tunnel curves out to this so that the midportion of the graft lies on the anterior surface of the arm. A relatively deep tunnel is necessary in thin patients to avoid erosion of the graft through the skin. For the same reason the medial side of the counterincision is undermined to form a protective flap at this point. After preclotting the graft with 25 ml ofthe patient's blood and marking it for orientation, the patient is systemically heparinized. Sutures of 5-0 braided polyester on a T-3 needlet are employed. The venous anastomosis is performed prior to placing the graft through the tunnel to facilitate exposure of the entire anastomosis (Figs. 1 B-D). The graft is then drawn through the tunnel (Fig. 1 E), flushed with heparinsaline and occluded in the axilla with a bulldog clamp. The lateral side of the proximal end of the arterial anastomosis is performed with the graft hood tip free (Fig. 1 F), after which the distal hood tip stitch is tied and the anastomosis completed. Each anastomosis is 1.5 cm in length. Placing the needle first through the vessel and then through the graft as shown makes stitching particularly easy and is without risk when the vessel wall is relatively normal. Cephalosporin is employed prophylactically in the perioperative period and the wounds are irrigated with 5% Neomycin. Dialysis via the graft is rarely allowed before an interval of at least two weeks after the operation. All patients underwent preoperative and postoperative evaluation in the vascular laboratory. In many USCI Sauvage Filamentous Velour Prosthesis, USCI, a Division of C.R. Bard, Inc., Billerica, Massachusetts 01821. t Tevdek Sutures, Deknatel, Inc., Queens Village, New York, New York 11429. *
263
F
FIG. 1. Construction of dacron graft fistula between the brachial artery and the axillary vein.
cases a flow metert was used intraoperatively to measure graft flow and brachial artery flow (with and without graft occlusion). Mean pressure in the brachial artery at the level of the anastomosis was measured intraoperatively with and without graft occlusion using a Tycos manometer guage§ connected by a sterile saline-filled tubing to a needle at the level of the arterial anastomosis. An approximation of the resistance to flow in the distal arm vasculature was calculated from intraoperative measurements.'0 No correction was t Statham Electromagnetic Flow Meter, Model 2201, Gould, Inc. Statham Instruments Division, 2230 Statham Boulevard, Oxnard, California 93030. § Tycos Sphygmomanometer Guage, Taylor Instrument, Consumer Products Division Sybron Corporation, Arden, North Dakota 28704.
264
BURDICK, SCOTT AND COSIMI
Ann. Surg. d March 1978
measurements with the flow probe revealed graft flows ranging from 150 to 1900 ml/min. The mean blood pres-
150-
l4\l
CB
100
GC
50-LM PG
CM**OSF
JB AL e
0
500
1000
1500
2000
GRAFT FLOW mi/min FIG. 2. Relationship between graft flow and distal brachial artery flow measured intraoperatively.
made for venous pressure; since this type of fistula produces comparatively little local venous hypertension'7 and no patient was clinically in congestive failure at the time, this should not markedly alter the results. Results In the 15 of 16 patients whose fistulas have been used, dialysis employing the standard 14 guage cannula has been successful without difficulties in puncturing the graft or in achieving hemostasis after dialysis. Complications have been minimal. One graft was removed for infection and one for skin erosion; both of these patients are stable on dialysis via another access. Of the five grafts which thrombosed, four were returned to good function after simple thrombectomy. The fifth required revision for venous stenosis at the axillary anastomosis similar to previous reports16'23 and is now functioning well. Another complication which has been observed by others was arterial thrombosis distal to the fistula'2 which developed during cardiogenic shock in one of our patients. This fistula was ligated at the time of distal thrombectomy. Two other patients have died from complications of their underlying disease since graft placement, both with functioning fistulas until time of death. After a median follow-up of nine months, nine of 16 grafts have been complication-free and 13 of 16 grafts (8 1%) are either still functioning or functioned until time of death. Some of the results of the hemodynamic measurements are presented in Figures 2 and 3. Intraoperative
sure in the brachial artery at the level of the anastomosis decreased when graft flow was established. The degree to which the mean brachial artery pressure fell upon release of graft occlusion did not appear to correlate with the flow in the graft. In Figure 2 it can be seen that there was perhaps a slight positive correlation between graft flow and distal brachial artery flow and this was true as well between graft flow and postoperative forearm pressure as determined in the vascular laboratory (Fig. 3). There was no apparent pattern to the small changes in calculated resistance in the distal vessels which occurred when graft occlusion was released. The systemic systolic blood pressure of all the patients in these studies was between 100 and 150 mm/Hg and there was no evident correlation between systemic systolic blood pressure and the other measurements depicted. Occlusion of the artery distal to the fistula during hypotension occurred in the patient who had had the lowest intraoperative graft flow (150 ml/min). Transient distal ischemic symptoms were noted in two patients. Intraoperative graft flow in one of these patients was also quite low (260 ml/min); flow was not measured in the other. There was no other evident correlation between the hemodynamic studies and any complication in these grafts. Two patients suffered cardiac arrests months after graft placement; after successful resuscitation the fistula flow and the distal perfusion in the arm remained excellent in both cases.
120-
CB
100Ait 80 60 Nk
40
20
0
500
1000
1500
2000
GRAFT FLOW mi/min FIG. 3. Relationship between intraoperative graft flow and postoperative forearm pressure.
DACRON GRAFT EXPERIENCE
Vol. 187 . No. 3
Discussion
The use of Filamentous Velour dacron grafts as described in this preliminary series has been quite successful, confirming the previous indications that this material6 and internal velour dacron18 provide useful alternatives for dialysis access fistulas. Patency has been excellent and cannulation is convenient. Only one fistula has become infected (this was at the site of skin erosion through a counter-incision) and no aneurysms have developed. In several difficult patients with multiple previous access procedures this procedure proved particularly welcome. The frequency and severity of complications encountered in this preliminary series are similar to the most optimistic reports of bovine grafts14'15'20'27 and the results were much better than the predominant bovine experience.3'5'7'12'16'2224 It is postulated that transmural tissue ingrowth encouraged by the Filamentous Velour graft25'26 produces an optimal potential for the healing of repeated needle punctures. Attention to the technical details of the operation, including the construction of the tunnel and counterincision, should prevent graft erosion through the skin and infection. Allowing tissue fixation to the outside surface of the graft to progress for two weeks prior to its use for dialysis may help to prevent subcutaneous extravasation with resultant infection or induration around the graft which makes later dialysis more difficult. The use of a straight graft in the upper arm position, chosen because of the multiple previous access failures in many of these patients, may also have contributed to the success noted in this series. An additional factor in the comparison with bovine grafts is that dacron grafts are much less expensive. It is possible
265
that expanded polytetraflourethylene grafts3 will also prove to be better than bovine grafts for dialysis fistulas. A disadvantage of the rather soft polytetraflourethylene graft is the possibility of tearing the wall when inserting a dialysis needle3 which does not seem to be a danger with the more durable dacron. The requirement for normal preoperative vascular exam and vascular laboratory findings in the fistula arm may have been responsible for the minimal hemodynamic consequences of these grafts. This requirement does not, however, completely safeguard against transient local "steal" symptoms. The intraoperative flow measurements bore out the fact that many of the fistulas had rather high flows. These are in the range of most previously reported flows for fistulas of this diameter12 although much higher flows may sometimes be observed with similar fistulas in the femoral position.29 It could be postulated that the patients with the highest graft flows might be the ones with the greatest "steal" through the graft and, therefore, the patients most likely to have ischemic symptoms in the extremity distal to the graft. This was clearly not the case in our patients. In fact there was a tendency toward a higher distal forearm flow in the patients with the highest graft flows as seen in Figure 2. The impression that patients with the highest graft flows had the best forearm perfusion distal to the graft junction was confirmed by the forearm pressures measured postoperatively as depicted in Figure 3. The resistance in these grafts is very low and the resistance in the distal vasculature remained fairly constant (Table 1). The factor which determined both graft flow and distal perfusion was apparently resistance in the brachial artery proximal to the graft anastomosis. In spite of the
TABLE 1. Hemodynamic Findings
Calculated Distal Resistance P.R.U.f
Patient M.B. C.B. J.B. G.B. G.C. J.F. S.F. P.G. A.L. C.M. L.M. C.M. J.M. M.S.
Graft Flow ML./min.
Distal Flow* ml./min.
Fall in Mean BPt mmHg
Graft Occluded
Open
1,900 480
115 25
55 43
31 66
39 126
450 1,450 1,200 400 260 1,100 150
60 145 40 40 20 40 40
50 65 25 54 58 44
46 43 56 72 102 114
30 20 78 84 180 138 -
-
Postoperative Forearm BP mmHg
Graft
110 115 50 80 60 60 65 50 80
(flat tracing) 100 120
-90 80
M.W. * In brachial artery distal to the anastomosis with graft flow established. t Measured intraoperatively with the graft occluded then released.
: Peripheral Resistance Unit
=
mm. Hg x sec. ml.
266
BURDICK, SCOTT AND COSIMI
fact that preoperative vascular laboratory studies were normal in each case, the capacity of the brachial artery to provide increased flow through the low-resistance fistula seems to have varied considerably among these patients. When such adaptation was possible due to a low brachial artery resistance proximal to the graft, both the graft flow and distal flow were higher. The lack of dependence of fistula flow or distal flow upon fistula size when large (low-resistance) fistulas are employed has been observed experimentally. 13"9 In addition, the dependence of fistula flow upon inflow rather than fistula size has also been observed for radio-cephalic fistulas at the wrist.I Although an oversimplification of the hemodynamics involved, this seems to explain the finding that the transient ischemic symptoms which occurred were in patients with the lowest, not the highest, fistula flows. It is concluded that there would be no local hemodynamic advantage to employing grafts of a smaller diameter. The Filamentous Velour dacron graft is applicable to most chronic dialysis situations. One patient has subsequently received a successful transplant with continued graft patency. In two patients the grafts have provided convenient sites for administration of antithymocyte globulin. One patient has employed her dacron graft successfully for ten months on home dialysis. Diabetes is not a contraindication to use of this graft. A likelihood of severe systemic vascular instability in any given patient should probably preclude use of this large caliber conduit. However, no patient with an intraoperative fistula flow rate of 400 ml or greater has had local ischemic symptoms, and in none of the 16 patients was there a clinically apparent high output decompensation. No instances of significant arm edema were noted in our series. Patients with thin skin and subcutaneous tissue should have these grafts placed deeply enough to prevent erosion through the skin. The Filamentous Velour dacron graft was only employed in this initial series as a last resort for patients in whom all else had failed. The success with the graft experienced in this small group of patients provides encouragement that it will be a very effective option for more routine dialysis access problems as well as for the very difficult patient. Hemodynamic consequences were minimal and, in particular, large graft flows were very well tolerated. References 1. Anderson, C. B., Etheridge, E. E., Harter, H. R., et al.: Local
Blood Flow Characteristics of Arteriovenous Fistulas in the Forearm for Dialysis. Surg. Gynecol. Obstet., 144:531, 1977. 2. Anderson, B. and Groce, M.: Banding of Arteriovenous Dialysis Fistulas to Correct High-output Cardiac Failure. Surgery, 78: 552, 1975. 3. Baker, L. D., Johnson, J. M. and Goldfarb, D.: Expanded Polytetrafluoroethylene (PTFE) Subcutaneous Arteriovenous Conduit: An Improved Vascular Access for Chronic Hemodialy-
Ann. Surg. * March 1978
sis. In Transactions of The American Society for Artificial Internal Organs, 22:382, 1976. 4. Beven, E. G. and Hertzer, N. R.: Construction of Arteriovenous Fistulas for Hemodialysis. Surg. Clin. North Am., 55:1125, 1975. 5. Biggers, J. A., Remmers, A. R., Glassford, D. M., et al.: Bovine Graft Fistulas in Patients with Vascular Access Problems Receiving Hemodialysis. Surg. Gynecol. Obstet., 140:690, 1975. 6. Brener, B.: Personal Communication, 1977. 7. Burbridge, G. E., Biggers, J. A., Remmers, A. R., et al.: Late Complications and Results of Bovine Xenografts. Trans. Am. Soc. Artif. Intern. Organs, 22:377, 1976. 8. Ehrenfeld, W. K., Gravsz, H. and Wylie, E. J.: Subcutaneous Arteriovenous Fistulas for Hemodialysis. Am. J. Surg., 124: 200, 1972. 9. Fee, H. J., Levisman, J., Doud, R. B., et al.: High-output Congestive Failure from Femoral Arteriovenous Shunts for Vascular Access. Ann. Surg., 183:321, 1976. 10. Guyton, A. C.: Textbook of Medical Physiology, Philadelphia, W. B. Saunders Co., 4th Ed., 221, 1971. 11. Haimov, M., Baez, A., Neff, M., et al.: Complications of Arteriovenous Fistulas for Hemodialysis. Arch. Surg., 110:708, 1975. 12. Haimov, M., Burrows, L., Baez, A., et al.: Alternatives for Vascular Access for Hemodialysis: Experience with Autogenous Saphenous Vein Autografts and Bovine Heterografts. Surgery, 75:447, 1974. 13. Hobson, R. W., Croom, R. D. and Swan, K. G.: Hemodynamics of the Distal Arteriovenous Fistula. J. Surg. Res., 14:483, 1973. 14. Hutchin, P., Jacobs, J. R., Devin, J. B., et al.: Bovine Graft Arteriovenous Fistulas for Maintenance Hemodialysis. Surg. Gynecol. Obstet., 141:255, 1975. 15. Johnson, J. M., Kenoyer, M. R., Johnson, K. E., et al.: The Modified Bovine Heterograft in Vascular Access for Chronic Hemodialysis. Ann. Surg., 183:62, 1976. 16. Katzman, H. E., Schild, A. F. and Vanderwerf, B. A.: Bovine Artegraft Arteriovenous Fistulas for Hemodialysis in OneHundred Patients after "Conventional" Arteriovenous Fistulas Failed. Vasc. Surg., 10:169, 1976. 17. Lavigne, J. E., Brown, C. S., Fewel, J., et al.: Hemodynamics Within a Canine Femoral Arteriovenous Fistula. Surgery, 77: 439, 1975. 18. Levowitz, B. S., Flores, L., Dunn, I., et al.: Prosthetic Arteriovenous Fistula for Vascular Access in Hemodialysis. Am. J. Surg., 132:368, 1976. 19. Lough, F. C., Giordano, J. M. and Hobson, R. W.: Regional Hemodynamics of Large and Small Femoral Arteriovenous Fistulas in Dogs. Surgery, 79:346, 1976. 20. Merickel, J. H., Andersen, R. C., Knutson, R., et al.: Bovine Carotid Artery Shunts in Vascular Access Surgery. Arch. Surg., 109:245, 1974. 21. Piccone, V. A.: Preserved Allograft of saphenous vein for vascular access in hemodialysis, Ann. Surg., 182:727, 1975. 22. Rolley, R. T., Sterioff, S. and Williams, G. M.: Arteriovenous Fistulas for Dialysis Using Modified Bovine Arteries. Surg. Gynecol. Obstet., 142:700, 1976. 23. Rosental, J. J., Spigelman, A., Gaspar, M. R., et al.: Problems with Bovine Heterografts for Hemodialysis. Am. J. Surg., 130: 182, 1975. 24. Sannella, N. A., Mehigan, J. T., Pennell, J. P., et al.: Initial Approach to Blood Access for the Chronic Hemodialysis Patient-The Bovine Heterograft. Trans. Am. Soc. Artif. Intern. Organs, 22:394, 1976. 25. Sauvage, L. R., Berger, K., Beilin, L. B., et al.: Presence of Endothelium in an Axillary-Femoral Graft of Knitted Dacron with an External Velour Surface. Ann. Surg., 182:749, 1975. 26. Sauvage, L. R., Berger, K. E., Wood, S. N., et al.: Interspecies Healing of Porous Arterial Prostheses. Arch. Surg., 109:698, 1974. 27. Sterling, W. A., Taylor, H. L. and Diethelm, A. G.: Vascular Access for Hemodialysis by Bovine Graft Arteriovenous Fistulas. Surg. Gynecol. Obstet., 141:69, 1975.
The technique, hemodynamics, and results of a new, more successful Filamentous Velour Dacron graft dialysis fistula between brachial artery and axillary vein are described. Hemodialysis has been successfully instituted in 15/16 patients. One graft was removed for infection and one for overlying skin erosion. There have been five episodes of thrombosis, all with succegsful restoration to function. After a median nine-month follow-up, nine of 16 grafts have been complication-free and 13 of 16 remain functional or functioned until the time of death. Hemodynamic studies revealed that with high flows (even 1900 cc/min), there was no evidence of cardiac failure or peripheral "steal." One patient with low flow (80% of these grafts should provide long-term dialysis access for this difficult patient population. T HERE IS NOW GENERAL ACCEPTANCE that the
Brescio-Cimino radiocephalic wrist fistula is the best access procedure for chronic dialysis. However, there are many long-term dialysis patients who no longer have vessels suitable for this fistula. These patients require some secondary procedure. Saphenous vein autografts8"2 have been reasonably successful when possible but these have a tendency to eventually develop aneurysmal dilatation, and the extensive dissection required for harvesting increases the morbidity of this procedure. The other alternative with which there has been a large experience utilizes the modified bovine carotid xenograft. Although the bovine grafts have apparently been employed quite successfully for dialysis access in some series,'4'5'20'27 there have been many reports of distressingly high rates of thrombosis, infection and aneurysm formation in these fistulas.3'5'7'12,16,22-24 Dissatisfaction with our bovine graft fistulas because of frequent thromboses and suture-line aneurysms convinced us that a better alternative is needed for patients whose vessels are not suitable for construcSupported in part by the USPHS Training Grant (lT32AI702701) and USCI, a division of C.R. Bard, Inc. Submitted for publication July 5, 1977.
From the Transplantation-Immunology Unit of the General Surgical Services, Massachusetts General Hospital, and the Department of Surgery, Harvard Medical School, Boston, Massachusetts
tion of a wrist fistula. Fistulas of expanded polytetraflouroethylene3 may provide results which are somewhat better than the bovine experience but data on these, as well as on the use of saphenous vein allografts,2' are as yet preliminary. There is experimental and initial clinical evidence that standard internal velour dacron provides an acceptable graft for dialysis access fistulas.18 The Filamentous Velour dacron graft should theoretically be an even better material for dialysis access. Studies with experimental grafts in the dog and other species have shown that the transmural fibrillar structure of this material encourages transmural tissue ingrowth25.26 which should improve resistance to the regular punctures required for dialysis. Brener has had success with this material for dialysis access.6 In view of the problems regularly associated with the previously described secondary access procedures, it was decided to investigate the use of Filamentous Velour dacron grafts for construction of vascular access in a series of dialysis patients who required such a secondary procedure. It was elected to employ a 5 or 6 mm diameter graft in the upper arm and thus a relatively high flow through the fistula was anticipated. Since ischemic symptoms distal to upper arm fistulas have been reported''24 the hemodynamic consequences of these shunts were evaluated in many of the patients. Materials and Methods
Sixteen chronic dialysis patients ranging in age from 30 to 70, in whom a forearm Brescio-Cimino fistula was not possible, were chosen to receive a dacron graft. A history of multiple failed access procedures
0003-4932-78-0300-0262-0075 X J. B. Lippincott Company
262
Vol. 187 . No. 3
DACRON GRAFT EXPERIENCE
was common; three patients were diabetics. Systemic infection was a contraindication and significant arteriosclerosis in the proposed extremity was ruled out preoperatively by physical exam and vascular laboratory evaluation. General anesthesia was usually employed but the procedure has been successfully performed in four patients under local or regional anesthesia. In each instance a Filamentous Velour dacron graft* was implanted end-to-side between the brachial artery just above the antecubital fossa and the axillary vein.4 The graft was 6 mm in diameter in all but two patients who received 5 mm grafts and the length of the graft ranged from 14 to 25 cm. The operative procedure is depicted in Figure 1. A transverse incision is made in the distal axilla starting at the anterior axillary fold. The axillary vein is exposed at the junction of the basilic and brachial veins (Fig. 1 A). The brachial artery is exposed via a "hockey-stick" incision with its transverse component just above the antecubital fold. A longitudinal counterincision equidistant between these is placed laterally and the tunnel curves out to this so that the midportion of the graft lies on the anterior surface of the arm. A relatively deep tunnel is necessary in thin patients to avoid erosion of the graft through the skin. For the same reason the medial side of the counterincision is undermined to form a protective flap at this point. After preclotting the graft with 25 ml ofthe patient's blood and marking it for orientation, the patient is systemically heparinized. Sutures of 5-0 braided polyester on a T-3 needlet are employed. The venous anastomosis is performed prior to placing the graft through the tunnel to facilitate exposure of the entire anastomosis (Figs. 1 B-D). The graft is then drawn through the tunnel (Fig. 1 E), flushed with heparinsaline and occluded in the axilla with a bulldog clamp. The lateral side of the proximal end of the arterial anastomosis is performed with the graft hood tip free (Fig. 1 F), after which the distal hood tip stitch is tied and the anastomosis completed. Each anastomosis is 1.5 cm in length. Placing the needle first through the vessel and then through the graft as shown makes stitching particularly easy and is without risk when the vessel wall is relatively normal. Cephalosporin is employed prophylactically in the perioperative period and the wounds are irrigated with 5% Neomycin. Dialysis via the graft is rarely allowed before an interval of at least two weeks after the operation. All patients underwent preoperative and postoperative evaluation in the vascular laboratory. In many USCI Sauvage Filamentous Velour Prosthesis, USCI, a Division of C.R. Bard, Inc., Billerica, Massachusetts 01821. t Tevdek Sutures, Deknatel, Inc., Queens Village, New York, New York 11429. *
263
F
FIG. 1. Construction of dacron graft fistula between the brachial artery and the axillary vein.
cases a flow metert was used intraoperatively to measure graft flow and brachial artery flow (with and without graft occlusion). Mean pressure in the brachial artery at the level of the anastomosis was measured intraoperatively with and without graft occlusion using a Tycos manometer guage§ connected by a sterile saline-filled tubing to a needle at the level of the arterial anastomosis. An approximation of the resistance to flow in the distal arm vasculature was calculated from intraoperative measurements.'0 No correction was t Statham Electromagnetic Flow Meter, Model 2201, Gould, Inc. Statham Instruments Division, 2230 Statham Boulevard, Oxnard, California 93030. § Tycos Sphygmomanometer Guage, Taylor Instrument, Consumer Products Division Sybron Corporation, Arden, North Dakota 28704.
264
BURDICK, SCOTT AND COSIMI
Ann. Surg. d March 1978
measurements with the flow probe revealed graft flows ranging from 150 to 1900 ml/min. The mean blood pres-
150-
l4\l
CB
100
GC
50-LM PG
CM**OSF
JB AL e
0
500
1000
1500
2000
GRAFT FLOW mi/min FIG. 2. Relationship between graft flow and distal brachial artery flow measured intraoperatively.
made for venous pressure; since this type of fistula produces comparatively little local venous hypertension'7 and no patient was clinically in congestive failure at the time, this should not markedly alter the results. Results In the 15 of 16 patients whose fistulas have been used, dialysis employing the standard 14 guage cannula has been successful without difficulties in puncturing the graft or in achieving hemostasis after dialysis. Complications have been minimal. One graft was removed for infection and one for skin erosion; both of these patients are stable on dialysis via another access. Of the five grafts which thrombosed, four were returned to good function after simple thrombectomy. The fifth required revision for venous stenosis at the axillary anastomosis similar to previous reports16'23 and is now functioning well. Another complication which has been observed by others was arterial thrombosis distal to the fistula'2 which developed during cardiogenic shock in one of our patients. This fistula was ligated at the time of distal thrombectomy. Two other patients have died from complications of their underlying disease since graft placement, both with functioning fistulas until time of death. After a median follow-up of nine months, nine of 16 grafts have been complication-free and 13 of 16 grafts (8 1%) are either still functioning or functioned until time of death. Some of the results of the hemodynamic measurements are presented in Figures 2 and 3. Intraoperative
sure in the brachial artery at the level of the anastomosis decreased when graft flow was established. The degree to which the mean brachial artery pressure fell upon release of graft occlusion did not appear to correlate with the flow in the graft. In Figure 2 it can be seen that there was perhaps a slight positive correlation between graft flow and distal brachial artery flow and this was true as well between graft flow and postoperative forearm pressure as determined in the vascular laboratory (Fig. 3). There was no apparent pattern to the small changes in calculated resistance in the distal vessels which occurred when graft occlusion was released. The systemic systolic blood pressure of all the patients in these studies was between 100 and 150 mm/Hg and there was no evident correlation between systemic systolic blood pressure and the other measurements depicted. Occlusion of the artery distal to the fistula during hypotension occurred in the patient who had had the lowest intraoperative graft flow (150 ml/min). Transient distal ischemic symptoms were noted in two patients. Intraoperative graft flow in one of these patients was also quite low (260 ml/min); flow was not measured in the other. There was no other evident correlation between the hemodynamic studies and any complication in these grafts. Two patients suffered cardiac arrests months after graft placement; after successful resuscitation the fistula flow and the distal perfusion in the arm remained excellent in both cases.
120-
CB
100Ait 80 60 Nk
40
20
0
500
1000
1500
2000
GRAFT FLOW mi/min FIG. 3. Relationship between intraoperative graft flow and postoperative forearm pressure.
DACRON GRAFT EXPERIENCE
Vol. 187 . No. 3
Discussion
The use of Filamentous Velour dacron grafts as described in this preliminary series has been quite successful, confirming the previous indications that this material6 and internal velour dacron18 provide useful alternatives for dialysis access fistulas. Patency has been excellent and cannulation is convenient. Only one fistula has become infected (this was at the site of skin erosion through a counter-incision) and no aneurysms have developed. In several difficult patients with multiple previous access procedures this procedure proved particularly welcome. The frequency and severity of complications encountered in this preliminary series are similar to the most optimistic reports of bovine grafts14'15'20'27 and the results were much better than the predominant bovine experience.3'5'7'12'16'2224 It is postulated that transmural tissue ingrowth encouraged by the Filamentous Velour graft25'26 produces an optimal potential for the healing of repeated needle punctures. Attention to the technical details of the operation, including the construction of the tunnel and counterincision, should prevent graft erosion through the skin and infection. Allowing tissue fixation to the outside surface of the graft to progress for two weeks prior to its use for dialysis may help to prevent subcutaneous extravasation with resultant infection or induration around the graft which makes later dialysis more difficult. The use of a straight graft in the upper arm position, chosen because of the multiple previous access failures in many of these patients, may also have contributed to the success noted in this series. An additional factor in the comparison with bovine grafts is that dacron grafts are much less expensive. It is possible
265
that expanded polytetraflourethylene grafts3 will also prove to be better than bovine grafts for dialysis fistulas. A disadvantage of the rather soft polytetraflourethylene graft is the possibility of tearing the wall when inserting a dialysis needle3 which does not seem to be a danger with the more durable dacron. The requirement for normal preoperative vascular exam and vascular laboratory findings in the fistula arm may have been responsible for the minimal hemodynamic consequences of these grafts. This requirement does not, however, completely safeguard against transient local "steal" symptoms. The intraoperative flow measurements bore out the fact that many of the fistulas had rather high flows. These are in the range of most previously reported flows for fistulas of this diameter12 although much higher flows may sometimes be observed with similar fistulas in the femoral position.29 It could be postulated that the patients with the highest graft flows might be the ones with the greatest "steal" through the graft and, therefore, the patients most likely to have ischemic symptoms in the extremity distal to the graft. This was clearly not the case in our patients. In fact there was a tendency toward a higher distal forearm flow in the patients with the highest graft flows as seen in Figure 2. The impression that patients with the highest graft flows had the best forearm perfusion distal to the graft junction was confirmed by the forearm pressures measured postoperatively as depicted in Figure 3. The resistance in these grafts is very low and the resistance in the distal vasculature remained fairly constant (Table 1). The factor which determined both graft flow and distal perfusion was apparently resistance in the brachial artery proximal to the graft anastomosis. In spite of the
TABLE 1. Hemodynamic Findings
Calculated Distal Resistance P.R.U.f
Patient M.B. C.B. J.B. G.B. G.C. J.F. S.F. P.G. A.L. C.M. L.M. C.M. J.M. M.S.
Graft Flow ML./min.
Distal Flow* ml./min.
Fall in Mean BPt mmHg
Graft Occluded
Open
1,900 480
115 25
55 43
31 66
39 126
450 1,450 1,200 400 260 1,100 150
60 145 40 40 20 40 40
50 65 25 54 58 44
46 43 56 72 102 114
30 20 78 84 180 138 -
-
Postoperative Forearm BP mmHg
Graft
110 115 50 80 60 60 65 50 80
(flat tracing) 100 120
-90 80
M.W. * In brachial artery distal to the anastomosis with graft flow established. t Measured intraoperatively with the graft occluded then released.
: Peripheral Resistance Unit
=
mm. Hg x sec. ml.
266
BURDICK, SCOTT AND COSIMI
fact that preoperative vascular laboratory studies were normal in each case, the capacity of the brachial artery to provide increased flow through the low-resistance fistula seems to have varied considerably among these patients. When such adaptation was possible due to a low brachial artery resistance proximal to the graft, both the graft flow and distal flow were higher. The lack of dependence of fistula flow or distal flow upon fistula size when large (low-resistance) fistulas are employed has been observed experimentally. 13"9 In addition, the dependence of fistula flow upon inflow rather than fistula size has also been observed for radio-cephalic fistulas at the wrist.I Although an oversimplification of the hemodynamics involved, this seems to explain the finding that the transient ischemic symptoms which occurred were in patients with the lowest, not the highest, fistula flows. It is concluded that there would be no local hemodynamic advantage to employing grafts of a smaller diameter. The Filamentous Velour dacron graft is applicable to most chronic dialysis situations. One patient has subsequently received a successful transplant with continued graft patency. In two patients the grafts have provided convenient sites for administration of antithymocyte globulin. One patient has employed her dacron graft successfully for ten months on home dialysis. Diabetes is not a contraindication to use of this graft. A likelihood of severe systemic vascular instability in any given patient should probably preclude use of this large caliber conduit. However, no patient with an intraoperative fistula flow rate of 400 ml or greater has had local ischemic symptoms, and in none of the 16 patients was there a clinically apparent high output decompensation. No instances of significant arm edema were noted in our series. Patients with thin skin and subcutaneous tissue should have these grafts placed deeply enough to prevent erosion through the skin. The Filamentous Velour dacron graft was only employed in this initial series as a last resort for patients in whom all else had failed. The success with the graft experienced in this small group of patients provides encouragement that it will be a very effective option for more routine dialysis access problems as well as for the very difficult patient. Hemodynamic consequences were minimal and, in particular, large graft flows were very well tolerated. References 1. Anderson, C. B., Etheridge, E. E., Harter, H. R., et al.: Local
Blood Flow Characteristics of Arteriovenous Fistulas in the Forearm for Dialysis. Surg. Gynecol. Obstet., 144:531, 1977. 2. Anderson, B. and Groce, M.: Banding of Arteriovenous Dialysis Fistulas to Correct High-output Cardiac Failure. Surgery, 78: 552, 1975. 3. Baker, L. D., Johnson, J. M. and Goldfarb, D.: Expanded Polytetrafluoroethylene (PTFE) Subcutaneous Arteriovenous Conduit: An Improved Vascular Access for Chronic Hemodialy-
Ann. Surg. * March 1978
sis. In Transactions of The American Society for Artificial Internal Organs, 22:382, 1976. 4. Beven, E. G. and Hertzer, N. R.: Construction of Arteriovenous Fistulas for Hemodialysis. Surg. Clin. North Am., 55:1125, 1975. 5. Biggers, J. A., Remmers, A. R., Glassford, D. M., et al.: Bovine Graft Fistulas in Patients with Vascular Access Problems Receiving Hemodialysis. Surg. Gynecol. Obstet., 140:690, 1975. 6. Brener, B.: Personal Communication, 1977. 7. Burbridge, G. E., Biggers, J. A., Remmers, A. R., et al.: Late Complications and Results of Bovine Xenografts. Trans. Am. Soc. Artif. Intern. Organs, 22:377, 1976. 8. Ehrenfeld, W. K., Gravsz, H. and Wylie, E. J.: Subcutaneous Arteriovenous Fistulas for Hemodialysis. Am. J. Surg., 124: 200, 1972. 9. Fee, H. J., Levisman, J., Doud, R. B., et al.: High-output Congestive Failure from Femoral Arteriovenous Shunts for Vascular Access. Ann. Surg., 183:321, 1976. 10. Guyton, A. C.: Textbook of Medical Physiology, Philadelphia, W. B. Saunders Co., 4th Ed., 221, 1971. 11. Haimov, M., Baez, A., Neff, M., et al.: Complications of Arteriovenous Fistulas for Hemodialysis. Arch. Surg., 110:708, 1975. 12. Haimov, M., Burrows, L., Baez, A., et al.: Alternatives for Vascular Access for Hemodialysis: Experience with Autogenous Saphenous Vein Autografts and Bovine Heterografts. Surgery, 75:447, 1974. 13. Hobson, R. W., Croom, R. D. and Swan, K. G.: Hemodynamics of the Distal Arteriovenous Fistula. J. Surg. Res., 14:483, 1973. 14. Hutchin, P., Jacobs, J. R., Devin, J. B., et al.: Bovine Graft Arteriovenous Fistulas for Maintenance Hemodialysis. Surg. Gynecol. Obstet., 141:255, 1975. 15. Johnson, J. M., Kenoyer, M. R., Johnson, K. E., et al.: The Modified Bovine Heterograft in Vascular Access for Chronic Hemodialysis. Ann. Surg., 183:62, 1976. 16. Katzman, H. E., Schild, A. F. and Vanderwerf, B. A.: Bovine Artegraft Arteriovenous Fistulas for Hemodialysis in OneHundred Patients after "Conventional" Arteriovenous Fistulas Failed. Vasc. Surg., 10:169, 1976. 17. Lavigne, J. E., Brown, C. S., Fewel, J., et al.: Hemodynamics Within a Canine Femoral Arteriovenous Fistula. Surgery, 77: 439, 1975. 18. Levowitz, B. S., Flores, L., Dunn, I., et al.: Prosthetic Arteriovenous Fistula for Vascular Access in Hemodialysis. Am. J. Surg., 132:368, 1976. 19. Lough, F. C., Giordano, J. M. and Hobson, R. W.: Regional Hemodynamics of Large and Small Femoral Arteriovenous Fistulas in Dogs. Surgery, 79:346, 1976. 20. Merickel, J. H., Andersen, R. C., Knutson, R., et al.: Bovine Carotid Artery Shunts in Vascular Access Surgery. Arch. Surg., 109:245, 1974. 21. Piccone, V. A.: Preserved Allograft of saphenous vein for vascular access in hemodialysis, Ann. Surg., 182:727, 1975. 22. Rolley, R. T., Sterioff, S. and Williams, G. M.: Arteriovenous Fistulas for Dialysis Using Modified Bovine Arteries. Surg. Gynecol. Obstet., 142:700, 1976. 23. Rosental, J. J., Spigelman, A., Gaspar, M. R., et al.: Problems with Bovine Heterografts for Hemodialysis. Am. J. Surg., 130: 182, 1975. 24. Sannella, N. A., Mehigan, J. T., Pennell, J. P., et al.: Initial Approach to Blood Access for the Chronic Hemodialysis Patient-The Bovine Heterograft. Trans. Am. Soc. Artif. Intern. Organs, 22:394, 1976. 25. Sauvage, L. R., Berger, K., Beilin, L. B., et al.: Presence of Endothelium in an Axillary-Femoral Graft of Knitted Dacron with an External Velour Surface. Ann. Surg., 182:749, 1975. 26. Sauvage, L. R., Berger, K. E., Wood, S. N., et al.: Interspecies Healing of Porous Arterial Prostheses. Arch. Surg., 109:698, 1974. 27. Sterling, W. A., Taylor, H. L. and Diethelm, A. G.: Vascular Access for Hemodialysis by Bovine Graft Arteriovenous Fistulas. Surg. Gynecol. Obstet., 141:69, 1975.
