Is the preferential use of polytetrafluoroethylene grafts for femoropopliteal bypass justified?

Transcription

1 Is the preferential use of polytetrafluoroethylene grafts for femoropopliteal bypass justified? William J. Quifiones-Baldrich, MD, Ronald W. Busuttil, MD, PhD, J. Dennis Baker, MD, Candace L. Vescera, RN, Sam S. Aim, MD, Herbert I. Machleder, MD, and Wesley S. Moore, MD, Los Angeles, Calif. The objective of this review is to analyze the long-term results of femoropopliteal bypass done preferentially with polytetrafluoroethylene (PTFE) grafts in patients who presumably had saphenous vein available. The results are analyzed according to preoperative variables in an attempt to determine those instances in which PTFE grafts may be preferred for the first reconstruction and to identify those patients who benefited from vein preservation. From 1979 to 1985, 146 femoropopliteal bypass operations were performed in 12 patients with 6 mm PTFE grafts used preferentially. The results with follow-up at 5 years are analyzed by actuarial methods. The patency rate at hospital discharge was 1%. The overall primary patency rate at 5 years was 57%. Reconstructions above the knee (11) and below the knee (45) had significantly different 5-year patency rates (63% vs 44%, p <.3). Sixty-two reconstructions done to alleviate disabling claudication had a 5-year primary patency rate of 69% and no amputations. Eighty-one reconstructions were done to treat critical ischemia with a 5-year patency rate of 49% and a 5-year foot salvage rate of 73%. When secondary operations were required to treat graft failures, the 4-year cumultive patency rate of the secondary reconstruction was 18% when performed with a prosthetic graft, in contrast to 7% when performed with the spared saphenous vein. We conclude that femoropopliteal reconstruction with PTFE grafts is a reasonable alternative for older patients with disabling claudication. Patients with critical ischemia will likely benefit from preservation of the vein with initial femoropopliteal reconstruction done with PTFE. Staged infrainguinal revascularization for foot salvage may improve present results. In this regard the sequence PTFE-then-vein carries a higher predicted patency rate than the sequence vein-then-ptfe. (J VASC SURG 1988;8: ) In 1949 Kunlin 1 described the use of autogenous saphenous vein as a bypass conduit to treat occlusive arterial lesions in the femoropopliteal system. Since then, femoropopliteal bypass with saphenous vein has remained the standard reconstruction for lower extremity revascularization. The excellent performance of the saphenous vein in the femoropopliteal position led vascular surgeons to successfully achieve limb salvage with more distal (tibial and pedal) reconstructions, thereby expanding our ability to maintain functional viability of the lower extremity. Inherent in all these reconstructions is a significant failure rate, which ranges from 28% to 5% at 5 years, From the Department of Surgery, University of California, Los Angeles Medical Center. Presented at the Third Annual Meeting of the Western Vascular Society, Monterey, Calif., January 28-31,!988. Reprint requests: William J. Quifiones-Baldrich, MD, Assistant Professor of Surgery, Dept. of Surgery, UCLA Center for the Health Sciences, 1833 LeConte Ave., Los Angeles, CA 924. leading to additional reconstructions. 2-7 The acceptable early results with femoropopliteal reconstructions with polytetrafluoroethylene (PTFE) encouraged us and others to use this graft in preference to the saphenous vein for the initial operation. 8,9 The potential advantages of this approach are a shorter operating time, the need for only two relatively small incisions, and most important, preservation of the saphenous vein for later use, if needed. A lower expected primary patency rate was not seen in our 3- year results, nor in the initial results of a randomized study comparing these two conduits. 8'1 Thus we continued our practice of using PTFE preferentially in the femoropopliteal position. In the meantime, the 6-year results of the randomized vein vs PTFE study became available, clearly showing improved primary patency in femoropopliteal bypasses with vein, although no difference in limb salvage. H This article summarizes our experience with the preferential use of PTFE grafts in the femoropopli- 219

2 22 Qui~ones-Baldrich et al. )rournal of VASCULAR SURGERY teal position and represents an expansion of our original series with a minimal follow-up of 2 years. PATIENTS AND METHODS A retrospective analysis of all femoropopliteal bypasses done at the University of California, Los Angeles (UCLA) Medical Center between January 1979 and June 1985 identified 349 reconstructions. Of these, 23 were secondary attempts at revascularization referred to our medical center after failure of previous bypass reconstructions. One hundred thirty-eight femoropopliteal bypasses were done with 6 mm PTFE preferentially as the initial attempt at revascularization. An additional eight patients (eight grafts) had this operation after a failed non-bypass procedure (e.g., profundaplasty or dilatation). This group of 146 primary femoropopliteal reconstruction constitutes the basis of this article. The decision to use PTFE was made preoperatively in all patients and no attempt was made to explore the saphenous vein during operation. No patient was excluded who fulfilled these criteria. The first 63 reconstructions (55 patients) were reported in On the basis of these favorable initial results, we continued to use PTFE preferentially as the initial conduit in the femoropopliteal position. All records were reviewed by one of the authors (W. J. Q-B.). Age, sex, presence of hypertension, diabetes, cigarette smoking, coronary artery disease, and/or hyperlipidemia were noted. Previous vascular operations were specifically recorded. All criteria used for this review attempted to conform to standards suggested for reports dealing with lower extremity ischemia. 12 All patients were tested preoperatively with noninvasive vascular procedures and had segmental lower extremity pressures determined by Doppler probe. In diabetic patients with calcified arteries, toe pressures determined by photoplethysmography were obtained. Routine aortography with runoff was performed preoperatively in all patients. All procedures were done with 6 mm PTFE grafts, 5- or 6- polypropylene (Prolene) suture in a running fashion for the anastomosis, and with systemic intravenous heparin during the operation. Prophylactic antibiotics were used routinely. All operations were done by chief residents or vascular fellows under the direct supervision of one of us. Indications for operation included disabling clandication and critical ischemia. Critical ischemia was defined as ischemic rest pain (grade II, chronic limb ischemia) or actual tissue loss caused by ischemia (grade III, chronic limb ischemia). Angiograms were reviewed and the runoff ~y~tcm catcgorlzcd as good if two or three tibia] vessels were in COlatinuity with the popliteal artery. If only one tibial vessel was in continuity with the popliteal artery or if there was an isolated popliteal segment, runoff was considered poor. Details of the operation were reviewed, specifically location of distal anastomosis (above or below knee), injury to the saphenous vein, or other intraoperative complications. Records were examined for postoperative complications. Wound problems such as infection or hematoma, need for a second operation, limb edema, cardiac complications, or early graft thrombosis were specifically recorded. Ankle-arm indexes were obtained within the first 24 hours in all patients. An increase from preoperative status of.15 or greater was considered evidence of graft patency. For the purpose of calculating primary patency, early graft thrombosis (i.e., less than 48 hours) that was successfully treated was not counted as a graft failure. Operative mortality was considered if it occurred within 3 days of operation or during the same hospitalization. Patients were evaluated postoperatively by one of us, regardless of outcome. Ankle-arm Dopplerderived indexes were obtained routinely in all patients during subsequent clinic visits. Graft thrombosis was suspected clinically when a deterioration of the status of the extremity occurred (return of clandication or severe ischemic symptoms), when a significant decrease (greater than.15) in the anklearm pressure ratio occurred, or both. This was confirmed in all cases by physical examination, angiography, and/or operation. Primary graft patency was defined as continued patency without intervention. Thus graft thrombectomy after the first 48 hours was considered a graft failure. Removal of a patent graft because of infection was considered a graft failure; however, amputation with a patent graft was considered a failure in foot salvage but counted as an open graft for patency calculations. Extensions to more distal sites because of progression of the occlusive process were counted as patent grafts if thrombosis never occurred but as failures if thrombectomy was required, regardless of cause. A secondary patency rate was calculated counting as patent all grafts that remained open primarily and as the result of subsequent intervention. For the purpose of this analysis, we calculated the patency of secondary procedures required to treat graft failures. These were divided according to the conduit used for the secondary operation (vein vs prosthetic). Each intervention was treated as a new

3 Volume 8 Number 3 September 1988 Is preferential use of PTFE grafts for femoropopliteal bypass justified? 221 entry in the life table. This was done taking into account that multiple secondary operations leading to prolonged patenet are less desirable than a single operation with similar long-term patenet. If a patient had a second operation with prosthetic material used the first time and vein used in the second operation or vice versa, each patency interval was considered separate. If amputation resulted, it was ascribed to the last intervention for foot salvage calculation. Cumulative foot salvage was recorded for the entire interval for each group, regardless of the number of operations. Patients with critical ischemia were analyzed to determine foot salvage. Amputations limited to the toes and transmetatarsal level were not considered as failures, whereas any amputation at the ankle or above was recorded as limb loss. Because follow-up was continued regardless of graft patency or limb loss, actuarial survival was calculated. Thus outcome extends beyond that of the extremity of interest. All patients were followed up for a minimum of 24 months. Performance of operations that used the spared saphenous vein during follow-up was specifically recorded. All data were analyzed by actuarial methods and presented in the life-table format. Comparisons were done with the nonparametric rank test of Mantel- Cox. When sample size was small, censoring patterns were examined. Calculations were done by the Biostatistics Department at the UCLA Medical Center. RESULTS Between January 1979 and June 1985, 12 consecutive patients had 146 primary femoropopliteal reconstructions at the UCLA Medical Center with preferential use of 6 mm PTFE grafts. All patients potentially had saphenous vein available from either the ipsilateral or contralateral lower extremity. Previous inflow procedures had been done in 51 patients. These were aortoiliae or femoral bypass in 37 patients, axillofemoral bypass in two patients, and balloon dilatation in 12 patients. Three emergency operations in three patients are included. Of the 12 patients, 78 were men and 42 were women. The youngest patient at the time of surgery was 43 years and the oldest was 91 years (mean age 66 years). There were 26 bilateral reconstructions for a total of 146 operations. Indications were disabling claudication in 52 patients (62 grafts, 42%) and critiqal ischemia (foot salvage) in 68 patients (81 grafts, 55%). Three grafts in two patients were placed because of popliteal aneurysms. In one of these, critical ischemia was present because of acute thrombosis of the aneurysm. Thirty-nine patients (48%) with critical ischemia had actual tissue loss with gangrenous toes, or open ulcerations, or both. Diabetes mellitus was present in 35 patients (29%, 42 grafts) and medically controlled hypertension in 7 patients (58%). Sixty-four patients (53%) had documented coronary artery disease. Nine of these patients had coronary artery bypass grafting (CABG) before and seven after their femoropopliteal bypass (CABG vein use, 5.8%). One hundred one reconstructions had the distal anastomosis created above the knee (69%) with the remainder being below the knee (45 grafts, 31%). Good runoff was present in 15 reconstructions (72%), whereas a single runoff vessel or an isolated popliteal segment was present in 41 (28%). Most patients (97 of 12, 8%) were habitual smokers and continued smoking after operation. There were eight minor complications not prolonging hospital stay; five were superficial wound infections defined as wound erythema without drainage and three were small hematomas not requiring drainage. It is our practice to continue antibiotics in those patients with wound erythema for 5 days and to follow up as outpatients. All of these resolved with this regimen. Major complications occurred in 11 patients for an overall morbidity rate of 9.1%. Early thrombosis (less than 48 hours) occurred in five patients; two of these cases were in two patients having simultaneous bilateral femoropopliteal bypass. Four were treated by simple thrombectomy and one required revision and distal extension. All grafts remained patent but the patient who required revision and distal extension needed an amputation despite a patent graft i month after surgery. There was one wound hematoma requiting operative drainage. Groin drainage resulted in a graft infection 2 weeks after implantation in one patient. This patient had Bricker's operation in the right lower quadrant of the abdomen and eventually manifested sepsis of an infected aortic graft, which required removal. Another patient had a groin lymphocele, which responded to aspiration with no furthcr sequelae. There were three myocardial infarctions related to the procedure. One occurred in a patient who had simultaneous bilateral femoropopliteal bypasses and resulted in one death. One other patient died of this complication for an overall mortality rate of 1.6%. Follow-up ranged from 24 to 1 months (mean 46 months). All patients were followed up regardless of the status of the reconstruction. Six patients were lost to follow-up with patent grafts at 8, 16, 26, 4, 43, and 71 months, respectively. Thus for graft patenet, follow-up was complete for 95% of the entire

4 222 Qui~ones-Baldrich et al. Journal of VASCULAR SURGERY o~ lo- ~xx 124 rr 8".- o 6 12_ ~- 4 o 11 " "-... ~ T 8 ""--" ~ Secondary 61 "~ "'--~ iprimary O_, Fig. l. Life-table curve of primary and secondary patency for 146 femoropopliteal bypasses with PTFE graft. Numbers at the intervals represent grafts at risk. 1 -~ 1 44 Foot Salvage o I...,.o/ _> ] l,., ± N Claudicetion a 6 "-- 29 "'-. ~'I... "Critical Ischemia " 4 8 r- ( Fig. 2. Life-table primary patency curve of femoropopliteal bypasses with PTFE graft according to indication for operation and actuarial foot salvage (claudication, n = 62; critical ischemia, n = 81). series. However, for survival analysis, an additional six patients were lost to follow-up after graft failure or amputation. The life-table analysis for primary and secondary patency is shown in Table L The primary patency rate was 68% and 57% at 3 and 5 years, respectively. Secondary patency (that achieved by additional interventions) was 79% and 74% for the same intervals (Fig. 1). Results of femoropopliteal bypass with PTFE grafts according to the indication for operation are presented in Fig. 2. The primary patency rate at 5 years was 69% when the indication for operation was disabling claudication. Patients with critical ischernia had a lower primary patency rate of 49% at 5 years; this difference was statistically significant (p <.3). Foot salvage was excellent, avoiding major amputation in 73% of threatened extremities at 5 years' follow-up. There were 2 major amputations in 81 threatened extremities; 85% of these (17 of 2) were required within the first 36 months after operation. The primary patency rate at 5 years for grafts placed above the knee was 63%. This is significantly different from a primary patency rate of 44% seen

5 Volume 8 Number 3 September 1988 Is preferential use of PTFE grains for femoropopliteal bypass justified? ,r'- q,.) ~_ 4o \I',,. ~o 3\ 52 t5 5 T T Above knee 9 ~... ~[- ---'[_. ~elow knee 3 '; 2 ' 6 '12 ' 2'4 ' 5'6 ' 4'8 ' 6' ' 7'2 Fig. 3. Life-table primary patency curve of femoropopliteal bypasses with PTFE according to location of distal anastomosis. Difference is statistically significant (p <.3) (above-knee, n = 11; below-knee, n = 45) o E (D o 4 13_ 14 5 r TGood "--~----']_ Poor 4 2, ' z'4 ' 3'6 ' 4's ' 6' ' 7'z Fig. 4. Life-table primary patency curve of PTFE femoropopliteal bypasses according to distal runoff. Curves are statistically similar (p >.8) (good runoff, n = 15; poor runoff, n = 41). in grafts with a below-knee distal anastomosis (p <.3; Fig. 3). Long-term results were not significantly different between the groups with good and poor runoff (Fig. 4). The primary patency rate at 5 years was 59% and 53%, respectively (i >.8). This contrasts with our early results where there was a significant difference between these two groups at 3 months. The 5-year primary patency rate of femoropop- liteal bypass with PTFE in diabetic and nondiabetic patients was 46.8% and 61%, respectively (Fig. 5). This trend was suggested in our 3-month results and now is close to being statistically significant with a p value less than.7. Because patients were observed beyond graft thrombosis or amputation, we were able to calculate an actuarial survival for this atherosderotic population. The actuarial survival at 5 years was 65.4%. Ten

6 224 Qui~ones-Baldrich et al. Journal of VASCULAR SURGERY 4\',..79 -J 84 ~ ' " t'- n 6 ' '-" "'---Z J Nondi(]betic ± ~L Diabetic 4 I 2 i i I i o, 2'4 4'8 6b ' - ' 72 Fig. 5. Life-table curve of primary patency of TFE femoropopliteal grafts in diabetic and nondiabetic patients. Difference approaches statistical significance (p <.7) (nondiabetics, n -- 14; diabetics, n = 42). o-.t g - u 4- o_ 2, I I I I I %, 6 2 I T T.., ff Vein % "2-. 1'2 ' 2'4 ' '.J_ "1 Prosthetic 5 2 Fig. 6. Life-table curve of primary patency of secondary reconstructions according to type of graft used. Difference is statistically significant (p <.2) (PTFE, n = 57; vein, n = 1). patients were lost to follow-up during the period of observation (24 to 1 months). Twenty-six patients (21%) have been followed up for 5 years or more. Analysis of failures: prosthetic graft vs vein. Forty-nine primary graft thromboses in 146 grafts occurred during follow-up. These were manifested as critical ischemia in 3 extremities, of which six were primarily amputated and 24 had secondary operations for foot salvage. Nineteen patients with graft failure had recurrent claudication; four of these pa- tients did not find symptoms disabling and no further intervention was indicated. Thus a total of 39 patients had second operations after primary graft thrombosis. The indications for operation were critical ischemia in 24 limbs and recurrent disabling claudication in 15 limbs. This group of 39 limbs was analyzed according to the method of secondary reconstruction depending whether vein or prosthetic material was used for the secondary procedures. Ten limbs had second operations with the spared autogenous saphenous vein, as either a new bypass or extension to a distal tibial vessel in a sequential manner. Ten operations were done in 1 patients. In 29 patients, thrombectomy of the thrombosed graft with or without patch (vein or prosthetic) or extension to a distal vessel or a new prosthetic graft was used. Fifty-seven operations in 29 patients were performed in this group. Results are shown in Fig. 6. Actuarial patency at 4 years was better in patients having secondary reconstructions with the spared saphenous vein. The 4-year patency rate for secondary operations with prosthetic material was 18% in comparison to 7% with saphenous vein. Despite the small numbers, this reaches statistical significance (2 <.2). More importandy, of 24 limbs with critical ischemia, 18 had secondary operations with prosthetic material, yielding a 3-year actuarial foot salvage of 29%. In comparison, six limbs in which the spared autogenous saphenous vein was used yielded an actuarial foot salvage of 67% (Fig. 7). However, this difference is not statistically significant.

7 Volume 8 Number 3 September 1988 Is preferential use of PTFE grafts for femoropopliteal bypass justified? 225 6' \ % I T T T T Vein o O o o U_ x 2 I.... Prosthetic 'I Fig. 7. Life-table curve for secondary reconstructions according to graft used for foot salvage. Difference is not statistically significant probably because of small numbers (PTFE, n = 18; vein, n = 6). Table I. Life-table analysis of primary patency No. of grafts No. withdrawn patent became of: Interva2 patency Cumulative Standard Interval at risk at No. of rate patency error (too) start failed grafts Duration Follow-up Death (%) (%) (%) i , Of 57 secondary operations with prosthetic material, nine consisted of thrombectomy alone and 48 involved additional reconstructions. In the six patients in whom only thrombectomy was done as the second operation, no specific cause was found for the first thrombosis. All six patients were given warfarin postoperatively. Three grafts were patent at 32, 34, and 44 months after the second operation; one graft failed 62 months after reoperation and the remaining two grafts failed at 4 and 5 months, respectively. The latter occurred after anticoagulation was discontinued. These grafts were successfully retrieved with repeat thrombectomy. In contrast, of 48 second operations in the remaining 23 patients in whom additional reconstructions with PTFE were done (new bypass, extension, or patch), only eight grafts remained patent at 7, 9, 9, 12, 13, 18, 37, and 4 months, respectively. In three of these eight grafts, multiple revisions were required to achieve prolonged patency.

8 226 Qui~ones-Baldrich et al. Journal of VASCULAR SURGERY DISCUSSION The controversy regarding the preferred graft (vein vs PTFE) in infrainguinal revascularization prompted investigators to launch a multiinstitutional prospective randomized clinical study comparing these alternatives) This admirable effort was completed and the &year results were published in n Several important conclusions were drawn from their study. First, the 2-year primary patency of vein and PTFE bypasses was similar; however, at 5 years vein bypasses had superior primary patency rates. Second, there was a marked difference in favor of vein in infrapopliteal (tibial) reconstructions in early and late results, Third, no difference in limb salvage was seen at 4 years between the two groups. Fourth, there was a significant difference among patients within the PTFE group, with those randomized to PTFE having better primary patency compared with those who had no saphenous vein available and thus had to be entered into this group (obligatory PTFE). The superiority of the saphenous vein in regard to patency was clearly established, and important guidelines regarding clinical approach could be abstracted. Inherent in all infrainguinal reconstructions is a long-term failure rate. In the randomized study, this was 32% at 5 years for vein grafts, n Clinical series that used autogenous saphenous vein for femoropopliteal bypasses with follow-up to at least 5 years have failure rates that range from 28% to 48%. 2,4-7 Thus it is estimated that on the average, one third of patients with autogenous saphenous vein femoropopliteal bypasses will have graft failure by 5 years. More recent series with in situ or reversed autogenous saphenous vein femoropopliteal bypass have not shown an improvement over these figures, reporting primary patency rates of approximately 75% at 3 yearsy 3'14 The role of secondary operations in the management of patients with failed femoropopliteal grafts is well established. Of the various alternatives, distal extension or a new more distal bypass is frequently required. Excluding failures caused by proximal progression of the occlusive process, Brewster et al)s noted that of 14 secondary reconstructions, 68% required a more distal graft. The remainder were early failures treated by thrombectomy alone. On the basis of our experience, a new, more distal graft has the best long-term patency when done with autogenous saphenous vein. Our results with the preferential use of PTFE in the femoropopliteal position suggest a viable approach to lower limb revascularization. In patients with critical ischemia, improved limb salvage may be achieved by staged procedures. In this regard, an initial PTFE femoropopliteal graft may be the only procedure required in 5% of limbs. If a secondary procedure were needed (5% of patients), a reconstruction with vein would have an expected 4-year patency rate of 7%. Had all secondary reconstructions in this series been done with saphenous vein, an additional eight limbs could have been saved. This would have yielded an 85% limb salvage rate compared with the actual 73%. Patients initially given a saphenous vein femoropopliteal bypass may be at a disadvantage when their reconstruction fails. Present techniques allow vascular surgeons to treat advanced ischemia with infrapopliteal bypasses and to date the saphenous vein remains the best conduit for distal reconstructions. Thus when dealing with critical ischemia, the primary patency of the initial reconstruction may not be so important as the combined interval patency of the alternatives. One could refine this judgment further, proceeding with vein bypass initially in those patients unlikely to require secondary intervention. Unfortunately, such guidelines remain unclear. The concept of staged infrainguinal revascularization to treat critical ischemia has been evaluated by others. Sterpetti et al.16 reported 9 PTFE aboveknee femoropopliteal bypasses followed for 5 to 7 years. The actuarial primary patency was 58%. This group was compared with 17 above-knee and 77 below-knee femoropopliteal bypasses with autogenous saphenous vein. The primary patency rate for this group with autogenous reconstruction was 63% and 65% at 5 years, respectively. Of their graft failures, 2 limbs required a secohdary bypass. The 3- year cumulative patency rate for these secondary attempts was 58% and 16% for vein and PTFE, respectively. These results are remarkably similar to ours. They concluded that preservation of the saphenous vein for future use in lower extremity revascularization was important. However, foot salvage by this approach was not calculated. Similar differences in patency rates of secondary procedures for infrainguinal revascularization were noted by Brewster et al)s Nineteen secondary grafts with autogenous vein had a 63% 5-year patency rate compared with an 18% 5-year patency rate for 52 cases done with prosthetic grafts. The authors commented on the advisability of obtaining vein from other areas for secondary reconstruction and admitted that this was often not feasible, thereby accounting for the large number of secondary prosthetic reconstructions in their series.

9 Volume 8 Number 3 September 1988 Is preferential use of PTFE grafts for femoropopliteal bypass justified? 227 Evaluating the validity of the concept of staged infrainguinal revascularization for extending foot salvage, Rosen et al.9 compared 9 prosthetic and 78 in situ grafts done because of critical ischemia. The cumulative secondary limb salvage rate at 36 months was 73% for prosthetic graft alone, 78% for in situ saphenous vein, and 87% for staged reconstructions. Thus improved foot salvage was seen when the saphenous vein was used as a backup for primary prosthetic failures. The approach to patients with claudication may need to be different because prolonged initial patency is most important. Life expectancy must also be considered. On the basis of our experience, 69% of patients operated on to relieve claudication will enjoy primary patency of their reconstruction for 5 years or more. Thus for the older patient with limited life expectancy, this is a reasonable approach. The risk of eventual amputation is minimal (zero in our series) and the saphenous vein would be available for later use. However, younger patients would do best with the graft that has the best primary patency, that is, autogenous saphenous vein. Preservation of the saphenous vein for future use in CABG in our experience is a weak argument. Only 5.8% of our patients benefited from this. This is similar to the experience of Sterpetti et al.16 who reported a CABG use rate of.4%. With increased utilization of the internal mammary artery for grafting and percutaneous techniques to treat coronary artery disease, preservation of the saphenous vein solely for this reason does not seem justified. Several interesting observation s were made in our review of 5-year results. The 5-year cumulative patency rate was similar between grafts with good and poor runoff. In our 3-month review there was a significant difference between these two groups, s No particular trend in the 5-year life-table curve (Fig. 4) is evident to explain this difference. It is possible that runoff criteria used in this study to segregate the two groups may not be sufficiently sensitive to demonstrate the expected difference in long-term results between these two groups. Patients with above-knee femoropopliteal bypass with PTFE grafts had significantly better patency at 5 years than patients with a below-knee distal anastomosis (63% vs 44%). This differs from the results of the randomized study where below-knee vein and PTFE grafts did better than above-knee reconstructions, n Our results support the preferential use of the above-knee position when possible. Management of primary prosthetic graft failures has allowed us to identify two heterogeneous groups of patients. One group represents patients with failed grafts in whom no cause for failure is found at second operation. These patients seem to respond well to thrombectomy and anticoagulation with warfarin. The second group has graft thrombosis resulting from anastomotic hyperplasia, progression of distal disease, or both. In these patients, a new bypass with the spared saphenous vein seems to yield the best results. Actuarial survival in our population was better than expected (65% at 5 years) despite the high incidence of significant risk factors. This compares favorably with survival estimates of 5% at 5 years reported by Szilagyi et al. 6 i years ago. Improved survival in modern series of patients with atherosclerosis is being reported by others. 16 This underscores the importance of a planned approach to improve present long-term results of lower extremity vascular reconstructions. The primary use of PTFE in femoropopliteal reconstructions seems justified when the goal is limb salvage. Graft failures are best managed by utilization of the spared saphenous vein. This approach has the potential to improve overall foot salvage. An aboveknee distal anastomosis seems preferable, with a better predicted patency than below-knee reconstructions. The primary use of PTFE for femoropopliteal bypass to treat disabling claudication is a reasonable option in older patients. Younger persons with claudication will likely benefit from an initial femoropopliteal reconstruction with autogenous saphenous vein with a higher predicted primary patency. We acknowledge the excellent assistance of Ms. Glenda L. Fischer. REFERENCES 1. Ktmlin J. Le traitement de l'arterite obliterante par la greffe veineuse. Arch Mal Coeur 1949;42: DcWecse JA, Rob CG. Autogenous venous grafts ten years later. Surgery 1977;82: ' 3. Watelet J, Chaeysson E, Pods D, Menard JF, Papion H, Saour N, Testart J. In situ versus reversed saphenous vein for femoropopliteal bypass: a prospective randomized study of 1 cases. Ann Vase Surg 1986;1: Cranley JJ, Hafner CD. Revascularization of the femoropopliteal arteries using saphenous vein, polytetrafluoroethylene, and umbilical vein grafts: 5 and 6 year results. Arch Surg 1982;117: Darling RC, Linton RR. Durability of femoropopliteal reconstructions. Am J Surg 1972;123: Szilagyi DE, Hageman JH, Smith RF, Elliott JP, Brown F, Diem P. Autogenous vein grafting "in femoropopliteal atherosclerosis: the limits of its effectiveness. Surgery 1979; 36: Reichle FA, Rankin KP, Tyson RR. Long-term results of

10 228 Qui~ones-Baldrich st al. Journal of VASCULAR SURGERY 474 arterial reconstructions for severely ischemic limbs: a 14- year follow-up. Surgery 1979;85: Quifiones-Baldrich WJ, Ma~in-Paredero V, Baker JD, Busuttil RW, Machleder HI, Moore W8. Polytetrafluoroethyiene grafts as the first choice arterial substitute in femoropopliteal revascularization. Arch Surg 1984;119: Rosen RC, Johnson WC, Bush HL, Cho SI, O'Hara ET, Nabseth DC. Staged infrainguinal revascailarizafion: initial prosthetic above-knee bypass followed by a distal vein bypass for recurrent ischemia: a valid concept for extending limb salvage? Am J Surg 1986;152: Bergan 5J, Veith FJ, Bernhard VM, et al. Randomization of autogenous vein and polytetrafluoroethylene grafts in femoral distal reconstruction. Surgery 1982;92: Veith FJ, Gupta S, Ascer E, et al. Six-year prospective mulficenter randomized comparison of autologous saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J VASC SURG 1986;3: Rutherford RB, Flanigan DP, Gupta SK, et al Suggested standards for reports dealing with lower extremity ischemia. J VASC SURG 1986;4: Simone ST, Dubner B, Sail AR, et al. Comparative review of early and intermediate patency rates of polytetrafluoro-- ethylene and autogenous saphenous vein grafts for lower extremity ischemia. Surgery 1981;9: Taylor LM, Phinney ES, Porter JM. Present status of reversed vein bypass for lower extremity revascularization. J VAsc SURG 1986;3: Brewster DC, LaSalle AJ, Robinson JG, Strayhorn EC, Darling C. Femoropopliteal graft failures: clinical consequences and success of secondary reconstructions. Arch Surg 1983; 118: Sterpetti AV, Schultz RD, Feldhaus RJ, Peetz DJ Jr. Sevenyear experience with polytetrafluoroethylene as above-knee femoropopliteal bypass graft. Is it worthwhile to preserve the autologous saphenous vein? J VAse SURG 1985;2:97-12.