Axillopopliteal bypass grafting: Indications, late results, and determinants of long-term patency

Transcription

1 Axillopopliteal bypass grafting: Indications, late results, and determinants of long-term patency Enrico Ascer, MD, FACS, Frank J. Veith, MD, FACS, and Sushil Gupta, MD, FACS, New York, N.Y. In the last 12 years we have performed 55 axillopopliteal bypass graftings with 6 nun polytetrafluoroethylene grafts for limb salvage in 50 patients who were at high risk for limb loss. Indications for this procedure were (1) severe atherosclerotic disease of the common, superficial, and deep femoral arteries (33 cases); (2) failed aortofemoral bypass grafting with sufficient fibrosis or disease progression in the deep femoral artery (five cases); (3) insufficient hemodynamic and clinical improvement after axiuofemoral bypass grafting (10 cases); and (4) sepsis in the groin from a previously infected graft (seven cases). The 30-day operative mortality rate was 8%, and the 5-year cumulative patient survival rate was %. Overall 1-, 3-, and 5-year cumulative primary graft patency rates were 58%, 45%, and %, respectively. Comparable limb salvage rates were 83%, 68%, and 58%. Repeat operations increased 5-year patency rates from % to 59% (p < 0.05). Three-year patency rate for grafts placed in the presence of poor angiographic runoff in one vessel was 62% and for good angiographic runoff (two to three vessels) it was 57% (NS). Grafts to the above-knee popliteal artery had a patency rate of 67% at 3 years, whereas for grafts that crossed the knee joint it was 51% (NS). Three-year patency rate for 24 straight axillopopliteal grafts was 42%, and for 31 sequential axillofemoralpopliteal grafts it was 74% (p < 0.05). These results show that axillopopliteal bypass grafting is justified when other standard operations are not possible in patients who are in imminent danger of limb loss, and that every possible effort should be made to use the common or deep femoral artery as part of a sequential axillofemoral-popliteal procedure. (J VASC SURG 1989;10: ) Standard abdominal arterial reconstructions for limb salvage may occasionally be unfeasible or dangerous because of local or systemic risk factors. In an attempt to circumvent these sometimes complex problems, a number of different extraanatomic bypass graft configurations have been devised over the last 3 decades to salvage threatened ischemic lower limbs. During this period much has been learned about the benefits and limitations of the conventional axillofemoral and femorofemoral extraanatomic bypass operations. However, there is limited information regarding the role of the axillopopliteal bypass From the Division of Vascular Surgery, Montefiore Medical Center-Albert Einstein College of Medicine. Supported in part by grants form the Manning Foundation and the New York Institute for Vascular Studies. Presented at the Thirteenth Annual Meeting of the Southern Association for Vascular Surgery, Key West, Fla., Jan , Reprint requests: Enrico Ascer, MD, FACS, Division of Vascular Surgery, Montefiore Medical Center, 11I E. 210th St., New York, NY /14215 graft, and a clear definition of the indications for performing this procedure does not exist. Moreover, there are no reports of late results and determinants of long-term patency for this operation. In 1976 our group first initiated the use of axillopopliteal bypass grafting to salvage limbs in circumstances in which a more standard arterial reconstruction was not feasible? Since then, 55 straight axillopopliteal and sequential axillofemoral-popliteal bypass graftings have been performed in 50 patients. Our initial report of early successful results with this operation was received with reasonable skepticism, since late results for these very long subcutaneous grafts were not available. Despite the fact that others have used this operation with success, 2 the ultimate role of these grafts remained in question until longterm results could be documented. One purpose of the present article is to provide late primary and secondary patency results for axillopopliteal reconstructions. A second purpose is to evaluate whether these graftings will endure as worthwhile limb salvage procedures. In addition, we have assessed the effective- 285

2 286 Ascer, Veith, and Gupta jollrnal o* VASCULAR SURGERY SO 44~~ 32 IS I,I I, i I I I I ~ I Fig. 1. Cumulative e-table patient survival rates after 55 axillopopliteal bypass operations. Number of patients at risk and standard error are shown at 6-month intervals. ness of an aggressive reoperative approach in cases of graft failure. Finally, we have analyzed several factors that could predict success or failure of this DTpe of operation. MATERIAL AND METHODS Between November 1976 and June 1988, 55 axillopopliteal arterial bypass graftings were performed in 50 patients with 6 mm standard nonringed polytetrafluoroethylene (PTFE) grafts (W.L. Gore & Assoc., Elkton, Md.) at Montefiore Medical Ccm ter-albert Einstein College of Medicine. Thirty men and women with an age range from 49 to 88 years (mean, years) were operated on. Many risk factors were present, with 66% of the patients having hypertension and 60% having diabetes. Seventy-two percent of the patients had had at least one myocardial infarction, and 70% continued to smoke. Fourteen percent of the patients had had a stroke, and 10% were morbidly obese. All patients had limbthreatening ischcmia occurring as severe rest pain in 17 patients (31%), gangrene in 23 (42%), and nonhealing ischemic ulcers in 15 (27%). Indications for axillopopliteal bypass grafting. The indications for performing an axillopopliteal bypass grafting in the present series were as follows: (1) Extensive occlusive disease of the common femoral artery and its superficial and deep branches in combination with local or systemic risk factors, which precluded the use of the aortoiliac segment as an inflow source (33 cases). Many previous aortic operations, severe chronic obstructive pulmonary disease, recent myocardial infarction, and morbid obe- sity were common examples of such risk factors. (2) Failure of a previous aortofemoral bypass graft, progression of disease into and down the deep femoral artery, and a densely scarred groin (five cases), (3) Insufficient hemodynamic and clinical improvement after an axillofemoral bypass grafting (10 cases). (4) Sepsis in the groin from a previously infected graft in patients in whom performance of a standard obturator canal bypass grafting was not possible because of aortoiliac infection, multiple cardiac risk factors, or morbid obesity (seven cases). Types of axillopopliteal bypass graftings. Of the 55 axillopopliteal bypass graftings, 24 were straight axillo-to-popliteal bypasses, and 31 were sequential axillofemoral-to-popliteal bypasses. Crossover axillopopliteal grafts were deemed necessary, in 11 cases in the entire series (%). Thirty-two axillopopliteal bypass graftings were performed as primary procedures (I6 straight axillopopliteal; 16 sequential axillofemoral-to-poplitcal). The remaining 23 cases were distal extensions from an axillofemoral bypass grafting (12 cases), proximal extensions from a femoropopliteal bypass grafting (three cases), and straight axillopopliteal bypass grafting after failed ipsilateral aortofemoral or axillofemoral bypass graftings (eight cases). In four cases the axillopopliteal reconstruction was extended to one of the infrapopliteal vessels: two posterior tibial, one anterior tibial, and one dorsalis pedis artery. Preoperative evaluation. Preoperative antde/brachial pressure indexes and pulse volume recordings consistent with scvere lower limb ischemia

3 Volume 10 Number 3 September 1989 AxillopopliteaI bypass grafting 287 loo I sot I \i ".n 10 8 ~ 5 5 S o-.... ~ % e~ 18 = Primary, Secondary 11 : - - -_ = e % I I I I I I I I I I Fig. 2. Comparison of cumulative life-table primary mad secondary patency rates for 55 axillopopliteal bypass operations. Numbers at risk are shown at &month intervals. The difference in cumulative patency rates at 5 years was statistically significant (p < 0.05). were obtained in all patients. Preoperative arteriographic visualization of the outflow tract was routinely obtained for primary operations and for late graft failures (over 1 month). When standard arteriography was not useful in showing an acceptable runoff, wc have used adjunctive digital subtraction arteriographic techniques (18% of all cases). Over the last few years we have also evaluated the adequacy of the inflow tract by arteriography. This was prompted by our observation that occlusive disease of the subclavian/axillary artery could be an important cause of graft failure for bypass grafts originating from this region. 3 In the present series t'.~e last 18 patients had preoperative angiographic assessment of the aortic arch and its branches. Three of these cases (16%) were found to have significant arterial stenosis (over 50%) proximal to the proposed inflow site. All three cases had successful crossover axillopopliteal bypasses originating from an undiseased contralateral inflow tract. A pulmonary artery balloon-tipped catheter inserted to monitor left ventricular function has been helpful in the perioperative management of these patients who are at high risk. Techniques. The technical principles involved in the construction of axillopopliteal bypass grafts are similar to those previously described for axillofemoral and femoropopliteal grafts. However, some technical aspects of axillopopliteal grafts are worth reemphasizing. 4 Although this procedure can be entirely performed with local anesthestic, most patients at high risk will tolerate light general anesthesia, particularly when given by a skilled anesthesiologist. All our patients had the latter anesthetic, since it provides a comfortable environment for the surgeon to perform careful dissections, tunneling, and meticulous anastomoses. In addition, it provides optimal control of the patient's blood pressure and respiratory function. The choice of outflow site(s) and the route of the tunnel(s) for graft placement were dependent on whether or not infection or scarring were present and on the location and quality of patent arterial segments. In cases of uncomplicated straight axillopopliteal bypasses, the tunnel was constructed from the axillary artery to the ipsilateral groin along the mid axillary line and then to the poplitcal artery via a subsartorial route. When the ipsilateral groin was densely scarred or infected, the tunnel was created at a distance from this scar and infection, even being placed lateral to the anterosuperior iliac spine, and then was carried infcriorly along the lateral aspect of the upper thigh. From there it was gradually curved toward the medial aspect of the lower thigh. From this point the tunnel followed the anatomic route of the popliteal artery. However, if the standard medial approach to the popliteal artery was not feasible, the above-knee or below-l~ee popliteal segments were safely exposed through a lateral approach (eight cases)3 Another helpful strategy involved a direct approach to the distal two thirds of the deep femoral artery. 6 When the common femoral artery and its proximal branches were severely diseased or the groin had been subjected to multiple repeat operations, we

4 288 Ascer, Veith, and Gupta Journal of VASCULAR SURGERY g ~ I, I i I,i -- I I i i Fig. 3. Cumulative life-table limb salvage rates after 55 axillopopliteal bypass operations. Numbers at risk and standard error are shown at 6-month intervals. used an incision through virginal tissue in the middle of the thigh along the medial or lateral border of the sartorius muscle to expose the distal portion of the deep femoral artery. This approach was used as part of a sequential axillopopliteal bypass grafting in six cases. When the ischernic limb was on the side opposite the axillary inflow site, the graft should not be passed across the chest or the middle of the abdomen, as this may complicate unanticipated abdominal or chest operations. Rather, the tunnel should be constructed from the axillary artery to the ipsilateral groin and then continued to the contralateral groin in a gradually curving subcutaneous, suprapubic direction (11 cases). Another technical detail relates to the axillary artery-to-graft anastomosis. Performing a long 1.8 to 2 cm oblique anastomosis will allow for an intentional slight redundancy of the proximal portion of the graft; thus motion and kinking of the donor artery caused by shoulder motion will be eliminated. Also, the lack of undue tension at this anastomosis may decrease the incidence of false aneurysm or disruption. 4 We preferentially perform the distal anastomoses first, since the recipient arteries can usually be unclamped for the remainder of the procedure. This offers two advantages: (1) it permits earlier reestablishment of the collateral circulation in the ischemic limb, and (2) it decreases the total blood loss as platelet aggregates plug the anastomotic suture holes be- fore this anastomosis is subjected to full arterial pressure. Reoperative approach. Our plan of treatment for repeat operations on axillopopliteal PTFE graft failure has been similar to the ones described for thrombosed axillofemoral and femoropopliteal bypasses. 7'8 This included dissection of the distal anastomosis, longitudinal incision in the hood of the graft directly over the anastomosis, and proximal graft thromboectomy. If no cause of failure could be identiffed, the procedure was limited to simple thrombectomy (nine cases). However, if progression of distal disease was found to be the cause of graft thrombosis, then an extension graft was constructed (six cases). Last, if intimal hyperplasia was stenosing the outflow, the graft incision was extended into the recipient artery, and a patch of PTFE was used to widen the lumen (two cases). Patient follow-up and statistical analysis. Complete follow-up information was obtained in 95% of our 50 patients. Patients were examined by a senior vascular attending surgeon every 4 to 6 months. Patency of the graft was assessed by palpation of graft pulses or distal pulses or both and noninvasive parameters. Recurrence of symptoms, deterioration of pulse volume recordings, and more recently decreased flow velocities by duplex scanning were indications for arteriographic studies. Primary and secondary cumulative graft pateney rates, limb salvage rates, and patient survival rates were calculated by the life-table method. Appropriate

5 Volume i0 Number 3 September 1989 Axillopopliteal bypass grafling 289 I00 ~ \ I ". 18 _ F' ~\\ \ \ '~ ~ ~ 51% j " ' " ' Straight o Sequentiai ' ' '8 ' 310 ' Fig. 4. Cumulative life-table secondary patency rates for 31 sequential axillofemoral-popliteal bypasses and 24 straight axillopopliteal bypasses. Numbers at risk are shown at 6-month intervals. The difference in cumulative patency rates at 3 years was statistically significant (p < 0.05). = = Ak - Pop "---o Bk - POp i I 1 I I I Fig. 5. Cumulative life-table secondary patency rates for 28 axillopopliteal bypasses to the above-knee popliteal artery and 27 to the below-knee popliteal artery. Numbers at risk are shown at 6-month intervals. The difference in cumulative patency rates at 3 years was not statistically significant (p < 0.25). statistical comparisons of all cumulative life-table rates were performed by the log rank test. Statistical significance was defined as p < RESULTS Patient survival. The 30-day operative mortality rate for the entire series was 8%. All fottr deaths resulted from myocardial infarctions. The severity of the underlying systemic disease in this group of patients at high risk was reflected in the cumulative survival rate of % at 5 years (Fig. 1). No late deaths were related to the axillopopliteal bypass grafting. Overall graft patency rates. As shown in Fig. 2, the overall primary cumulative graft patency rates at 1, 3, and 5 years were 59%, %, and %, respectively. The secondary cumulative patency rates at comparable intervals were 75%, 59%, and 59%, respectively. Comparison between these life-table curves was statistically significant by the log rank test (p < 0.05). Overall limb salvage rates. Overall cumulative limb salvage rates were 83%, 68%, and 58% at 1, 3, and 5 years, respectively (Fig. 3). Straight versus sequential axillopopliteal graft patency. The a-year graft patency results for the 31 sequential axillofemoral-popllteal bypasses and for the 24 straight axillopopliteal bypasses arc depicted in Fig. 4. As shown, the results obtained with sequential bypasses were superior to those obtained with straight ones (p < 0.05). Patency rates and level of distal anastomosis. To verify whether the passage of the graft over an additional joint was an important prognostic factor for patency, we compared the patency results of 28 bypass graftings to the above-knee Popliteal artery with those of 27 reconstructions to the below-knee Popliteal artery. The graft patency rates at 1 year were superior for the above-knee grafts (p < 0.005). However, after 2 and 3 years of follow-up the patency differences did not reach statistical significance (Fig. 5). Patency rates and runoff. Twenty-eight axillo- Popliteal bypass grafts were primarily inserted into popliteal arteries with poor angiographic runoff (one vessel or isolated), whereas the remaining 27 bypass graftings were performed to popliteal arteries with good angiographic runoff (two or three vessels). As shown in Fig. 6, the a-year patency results were similar, 62% and 57%, respectively. Complications. The overall graft infection rate was 3.6%. Both infected axillopopliteal grafts were successfially treated with local operative debridements, systemic antibiotics, and frequent dressings changes. One additional patient had a false aneurysm in the groin that was treated by resection of the distal anastomosis and interposition of the graft to the deep femoral artery. The patient's postoperative course remained uneventful until his death 15 months later. All local wound cultures and systemic blood cultures were negative for bacteria.

6 290 Ascer, Veith, and Gupta ~ournal of VASCULAR SURGERY Cu " ~%'K 1~ = -- Good Runoff o. - -t Poor Runoff I I I 7 4 j 17-'-'%_ % 57% Fig. 6. Cumulative life-table secondary patency rates for 27 axillopopliteal bypass graftings performed in the presence of good runoff and 28 with poor runoff. Numbers at risk are shown at &month intervals. The difference in cumulative patency rates at 3 years was not statistically significant (p < 0.9). Twelve major amputations were performed in this series. Nine of these amputations were above the knee, and three were below the knee. DISCUSSION Our early experience suggested that limbthreatening ischemia could be completely reversed in patients in whom the only other alternative was a major amputation, by performing long prosthetic bypass graftings from the axillary to the poptitcal artery. ~ This conclusion was supported by an earlier case report of Smith et al.9 Our initial use of this operation was based on the acceptable longtcrm results achieved with axillofemoral and axillocross-femoral reconstructions couplcd with the fact that many of our patients at high risk could not tolerate the increased cardiac work load required for amputation after major amputation. The safety and case with which an axillopopliteal bypass cma be constructed renders it an attractivc palliative proccdure for patients with severe disease; scarring or infection of the iliac arteries and the common, decp, and superficial femoral arteries; and advanced heart or lung aiimcnts. Thc original conccrn that axillopopliteal bypass grafting would not pro. vide durable patency and limb salvage in a high proportion of cases has been partly alleviated by the present results, which show % primary graft patcncy and 58% limb salvage at 5 years. The relatively frequent failure rate and the need for repeat opera- tions are not surprising in view of the advanced stage of the disease process in patients subjected to this operation. The observation that sequential grafts had better patency rates than straight bypass grafts (Fig. 4) is relevant in this regard for two reasons. First, the sequential graft provides a means for improving primary patency in some cases without groin infection. Second, if primary axillopopliteal grafting performed for groin infection fails and the groin wound heals, at repeat operation it can be converted to a sequential axillo / distal deep femoral/poptiteal bypass graft. 6 This was done in four of our recent cases. It may be that the construction of an additional outflow source contributed to the improved results by generating increased flow rates through the longest portion of the bypass graft. We believe that in the presence of active groin infection, axillopopliteal bypass grafting is a superior alternative to obturator bypass grafting. The latter operation may not be feasible in the presence of aortoiliac sepsis, and the axitlopopliteal bypass is technicauy easier and less stressful to the patient because it is carried out via an extracavitary approach and is largely subcutaneous. An added advantage of axillopopliteal bypass grafts is that they afford the flexibility for creating many types of tunnel configurations to circumvent areas of infection or dense scarring. In this regard lateral nmnels and some unusual approaches to infrainguinal arteries have been particularly helpful. These include the lateral approach to the popliteal artery and the direct approach to the distal deep femoral artery. 5,6 It is of interest that the 3-year and 5-year primary patency results for axillopopliteal bypass grafts were similar to those we previously reported for axillopopliteal bypass grafts at comparable intervals (47% and %, respectively), s However, the secondary 5- year patent y rates for a,xillopopliteal grafts were not as favorable as those for axillofemoral grafts (55% vs 75%, respectively). This may be due to our initial reluctance to operate again on these very long grafts. However, repeat operations proved to be effective and increased patency rates sigiaificantly (Fig. 2). These findings justify an aggressive approach toward repeat operations for axillopopiiteal graft failure associated with recurrence of symptoms. A crucial technical factor related to the success of these repeat operations is approaching the thrombosed graft at its distal anastomotic end rather than in the middle portion. This ensures complete removal of the distal dot while sparing the recipient artery from potential damage by blind passage of balloon catheters. <9 We cannot explain why axillopopliteal bypass

7 Volume 10 Number 3 September 1989 Axillopopliteal bypass graftir~g 291 grafts placed into popliteal arteries with better angiographic runoff did not have better patency rates than those inserted into arteries with poor runoff. However, arteriographic runoff criteria have also failed to predict patency of femoropopliteal bypass grafts. 1 In addition, we have previously shown that the outflow from the popliteal artery is rarely sufficiently poor to cause graft failure, n Although analyses of graft patency for this type of bypass operation are important and should not be overlooked, an even more important consideration is limb salvage. In our series the 5-year cumulative limb-salvage rate was an acceptable 58%. Moreover, 77% of the 18 patients who died in the late followup period had functional limbs at the time of their death. We recognize that the ultimate level of amputation should also be discussed here in. Several factors may lead to the deterioration of distal circulation after failure ofinfrainguinal bypass operations. These include intraoperative technical mishaps such as ligation of important collateral vessels, propagation of the clot into the recipient artery, progression of atherosclerotic disease, and other less common factors. However, in the present series we believe that the reason why we performed more supragenicular amputations (nine out of 12 amputations) was because these patients had had a severely diseased or occluded ipsilateral iliofemoral arterial system before the performance of the axillopopliteal bypass grafting. The collateral circulation was inadequate to achieve healing at the infragenicular level when the graft was closed. The satisfactory long-term results presented in this report clearly reaffirm our previous impression, and that of others, that axillopopliteal bypass grafting is an effective and durable limb salvage operation for the patient at high risk in whom other more standard arterial reconstructive operations are not feasible and in whom the only other option is a major amputation. 1,2,~2 Thus although use of axillopopliteal bypass graft- ing should be restricted to these relatively unusual circumstances, this extended extraanatomic bypass operation deserves a permanent place in the armamentarium of vascular surgeons. REFERENCES 1. Veith FJ, Moss CM, Daly V, Fell SC, Haimovici H. New approaches to limb salvage by extended extra-anatomic bypasses and prosthetic reconstructions to foot arteries. Surgery 1978;84: Connolly JE, Kwaan JHM, Brownell D~ McCart PM, Levine E. Newer developments of extra-anatomic bypass. Surg Gynecol Obstet 1984;158: Ascer E, Veith FJ, Gupta SK, et al. Comparison of axitlounifemoral and axillofemoral bypass operations. Surgery 1985;97: Ascer E, Veith FJ. Extraanatomic bypasses. In: Haimovici H, ed. Haimovici's vascular surgery: principles and techniques. 3rd ed.norwalk: Appleton & Lange, 1989; Veith FJ, Ascer E, Gupta SK, Wengerter KR. Lateral approach to the popliteal artery. J VASC SURG 1987;6: Nunez A, Veith FJ, Collier P, Ascer E, White Flores S, Gupta SK. Direct approaches to the distal portions of the deep femoral artery for limb salvage bypasses. I VASC SURG 1988; 8:576-8I. 7. Veith FJ, Gupta SK, Daly V. Management of early and late thrombosis of expanded polytetrafluoroethylene (PTFE) femoropopliteal bypass grafts: favorable prognosis with appropriate reoperation. Surgery I9;87: Ascer E, Collier P, Gupta SK, Veith FJ. Reoperation for polytetrafluoroethylene bypass failure: the importance of distal outflow site and operative technqiue in determining outcome. J VAse SURG 1987;5: Smith RB, Perdue GD, Hyatt HC, Ansley ID. Management of the infected aortofemoral prosthesis including use of an axillopopliteal bypass. Am Surg 1977;73: Veith FJ, Gupta SK, Daly V. Femoropopliteal bypass to the isolated popliteal segment: is polytetrafluoroethylene graft acceptable? Surgery 1981;89: Ascer E, Veith FJ, Morin L, et al. Components of outflow resistance and their correlation with graft patency in lower extremity arterial reconstructions. J VASC SURG 1984;1: i2. Ascer E, Veith FJ, Gupta SK, et al. Six-year experience with expanded polytetrafluoroethylene arterial grafts for limb salvage. J Cardiovasc Surg (Torino) i985;26: