- Email: [email protected]
Femoropopliteal reconstruction with knitted, nonvelour Dacron versus expanded polytetrafluoroethylene
Femoropopliteal reconstruction with knitted, nonvelour Dacron versus expanded polytetrafluoroethylene W'dliam C. Pevec, MD, R. Clement Darling, MD, Gilbert J. L'Italien, BS, and William M. Abbott, MD, Boston, Mass. No ideal prosthetic conduit exists for femoropopliteal reconstruction. Expanded polytetrafluoroethylene (PTFE) has gained much greater popularity than the less expensive Dacron, despite disappointing long-term patency for each material. Few direct comparisons of results with the two materials have been performed. The concurrent experience of a single surgeon with femoropopliteal reconstructions of PTFE (n = 85) or knitted, nonvelour Dacron (KNVD, n = 38) was reviewed. Risk factors were comparable in the two groups. Overall patency was significantly better for KNVD than for PTFE (p = 0.0096 by log rank;p = 0.039 by Cox proportional hazards; 48% vs 27% at 5 years). Since in this series the overall performance of KNVD was significantly better than PTFE, a reassessment of the prosthetic materials used in femoropopliteal reconstruction is necessary. These results justify a prospective comparison of PTFE and KNVD. (J VASC SURG 1992;16:60-5.)
Expanded polytetrafluoroethylene (PTFE) has become the most popular conduit for femoropopliteal reconstruction when autogenous vein is not used. The use of PTFE has exceeded other alternatives, including allograft vein, heterograft artery, human umbilical vein, and Dacron. This preference for PTFE has developed for several reasons. Its smooth, hydrophobic surface intuitively appears "thromboresistent." It handles well and does not require preclotting, making it attractive to the surgeon. Articles published after its clinical release in 1976 reported excellent early patency rates. ~-4 Dacron was one of the early materials used in these operations. As the use of PTFE for femoropopliteal reconstruction increased, the use of Dacron decreased. However, the preference for PTFE over Dacron is not based on solid scientific evidence. A recent review of the literature found only one large series directly comparing the two prosthetic materials in femoropopliteal reconstructions; that retrospective series showed similar 5- and 10-year patencies for PTFE and Dacron.S Since Dacron is considerably less expensive than PTFE, if PTFE offers no clear From Massachusetts General Hospital, Department of Surgery, Division of Vascular Surgery, Boston. Reprint requests: William Abbott, MD, Division of Vascular Surgery, Massachusetts General Hospital, ACC 458, Boston, MA 02114. 24/1/35441
60
advantage over Dacron, then perhaps the preferential use of PTFE needs to be reassessed. The clinical impression of one of the authors was that a specific configuration of Dacron, a nonvelour, high-porosity knitted graft (the LoPor graft; Golaski Labs, Philadelphia, Pa.; porosity = 2000 to 3000 ml water/cm2/min), had a particularly good patency in the femoropopliteal position. A review was therefore undertaken of the experience of this single surgeon, comparing the results of femoropopliteal reconstructions performed with this material with those performed with PTFE (Gore-Tex; W. L. Gore and Associates, Inc., Elkton, Md.). PATIENTS AND METHODS
A retrospective review was performed of one surgeon's concurrent experience with the two graft materials. All patients undergoing femoropopliteal reconstructions with knitted, nonvelour Dacron (KNVD) or PTFE by one of the authors (R.C.D.) between the dates of June 1,1974, and Jan. 31,1986, were included. The KNVD grafts were implanted from June 1974 through January 1986; the PTFE grafts were implanted from May 1977 through December 1985. During this same time period, autogenous saphenous vein was the conduit of choice. Other conduits, including human umbilical vein, cryopreserved saphenous vein, and other configurations of Dacron, were also used, but they are
Volume 16 Number 1 July 1992
Femoropopliteal reconstruction with Dacron versus PTFE
61
Table I. Demographic and risk factors PTTE
KNVD (n =
n
Age (yr) Male Cardiac disease Carotid disease Hypertension Diabetes Smoking Indication Claudication Rest pain/tissue loss Acute ischemia Adjunctive Redo Size = 6 mm Proximal reconstruction Runoff -< 1 Below-knee
%
n
64.5 +- 10.4 23/38 23/36 13/32 22/36 15/38 29/33
61 64 41 61 39 88
11/37 20/37 6/37 5/38 11/38 33/37 22/38 16/31 17/37
30 54 16 13 29 89 58 52 46
not included in this analysis. The selection of one particular graft material over another did not follow any specific criteria. Rather, the use of multiple different materials reflected the surgeon's dissatisfaction with the performance of each of the nonautogenous materials. Of note, the authors have never had any commercial relationship with Golaski Labs or W. L. Gore and Associates, Inc. Information was obtained through review of the patient's chart, noninvasive vascular laboratory reports, angiogram reports, operative note, and operative diagram. Excluded were grafts to the tibial vessels, short grafts from the distal superficial femoral artery to the popliteal artery for popliteal aneurysms, composite grafts, and grafts from the femoral artery to the contralateral popliteal artery without an ipsilateral groin anastomosis. Thirty-eight KNVD and 85 PTFE grafts met the inclusion criteria. Demographic and risk factors assessed were age; sex; indication; principal versus adjunctive procedure; primary versus secondary operation; presence of proximal arterial reconstruction; inflow vessel; outflow vessel; number of patent runoff vessels; graft size; presence of documented cardiac disease, carotid disease, hypertension, and diabetes mellitus; and history of smoking (a patient who smoked at any time, past or present, was considered a smoker). All terms were defined according to the guidelines suggested by the Ad Hoc Committee on Reporting Standards of the Society for Vascular Surgery and the North American Chapter of the International Society for Cardiovascnlar Surgery. 6
(n = ss)
38)
%
p
63/85 39/72 19/67 37/73 35/82 65/76
74 54 28 51 43 86
0.48 0.13 0.34 0.22 0.31 0.74 0.74 0.30
18/84 48/84 18/84 16/85 15/85 70/81 43/85 31/66 48/84
21 57 21 19 18 86 51 47 57
65.9 - 9.7
0.44 0.16 0.46 i 0.67 ~ 0.26
Patency was assessed by documentation by the operating surgeon (R.C.D.) of a palpable pulse distal to the graft and/or maintenance of the postoperative improvement in pulse volume recordings and Doppler pressures. In some instances patency was documented by angiography or direct intraoperative assessment. In four cases the assessment of another physician was accepted as documentation ofpatency. These patency criteria do not strictly adhere to the guidelines of the Ad Hoc Committee. Strict adherence is difficult in a retrospective series that predates the Committee's report. Every effort was made to follow these guidelines as closely as possible, and in most cases the criteria of the Committee were fulfilled. A graft was considered patent only to the date of the last examination that documented patency. After this date, the graft was considered lost to follow-up. In most instances the date of graft occlusion could be unequivocally determined, became the patient developed acute symptoms that led to prompt evaluation. Less commonly, the graft occluded without symptoms. When an asymptomatic occlusion was diagnosed, the graft was considered to have occluded halfway through the interval between examinations, as recommended by the Ad Hoc C o m m i t t e e . 6 Primary and secondary patency, as defined by the Ad Hoc Committee, are reported. Grafts undergoing thrombectomy or revision within 24 hours of the original operation, with subsequent sustained patency, were considered to be primarily patent. Grafts that occluded more than 24 hours after implantation, in which patency was subsequently reestablished, were considered to be secondarily
62
Journal of VASCULAR SURGERY
Pevec et al.
Table II. Overall patency Interval (too)
Grafts at risk (no.)
Failedgrafts (no.)
Withdrawn grafts (no.)
Interval patency (%)
Cumulative patency (%)
SE (%)
PTFE
0-1 1-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 48-54 54-60
85 72 52 38 30 24 21 18 17 15 12
10 13 13 5 4 2 1 0 0 2 1
3 7 1 3 2 1 2 1 2 1 0
88 82 75 87 87 92 95 100 100 87 92
88 71 54 46 39 36 34 34 34 29 27
3.5 5.1 5.8 5.8 5.8 5.8 5.8 5.8 5.8 5.9 5.9
KNVD
0-1 1-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 48-54 54-60
38 33 27 24 20 19 17 16 14 13 12
5 4 2 3 1 0 1 1 1 0 0
0 2 1 1 0 2 0 1 0 1 0
87 88 93 88 95 100 94 94 93 100 100
87 76 70 61 58 58 55 51 48 48 48
5.5 7.0 7.6 8.2 8.3 8.3 8.5 8.6 8.8 8.8 8.8
patent. Limb salvage, defined as the proportion of grafted limbs maintained without below- or aboveknee amputation, is also reported. Chi-square analysis was used to compare the demographic and risk factors between the two groups of patients. Student's t test was used to compare the ages of the two groups. Percent graft patency and limb salvage rate were determined by the Kaplan-Meier (product limit) method. Patency and limb salvage curves were compared by the log rank test. In addition, Cox Proportional Hazards regression was used to evaluate the influence of graft type (KNVD or PTFE) on primary patency while adjusting simultaneously for the effects of all listed risk factors, according to the method described by Kalbfleisch and Prentice. 7 All statistical computations were performed by use of SAS software (SAS Institute, Cary, N.C.). RESULTS No statistically significant differences occurred in demographic or risk factors between the KNVD and PTFE groups (Table I). One death occurred shortly after operation in the KNVD group (operative mortality rate, 2.8%), and three occurred in the PTFE group (operative mortality rate, 3.8%). Two deaths were due to myocardial infarction, one was due to a perforated gallbladder, and one was due to bowel infarction. Four
complications occurred in the KNVD group (morbidity, 10.5%), and 15 complications occurred in the PTFE group (morbidity, 17.6%). Acute thrombectomy and/or graft revision (less than 24 hours after the original operation), with subsequent sustained patency, were performed in one KNVD (2.6%) and six PTFE (7.1%) grafts. One pseudoaneurysm developed in each group. A KNVD graft developed a pseudoaneurysm at the popliteal anastomosis, became of disruption of the Dacron prosthetic, 8 years and 6 months after implantation. One PTFE graft developed a pseudoaneurysm at the site of the proximal anastomosis to a previously placed aortofemoral prosthesis, 5 years and 10 months after implantation of the femoropopliteal graft. The patency curves for the two graft materials were compared by use of the log rank test. Overall primary patency of the KNVD grafts was superior to the primary patency of the PTFE grafts (p = 0.0096, Table LI, Fig. 1). Five-year primary patency of the Dacron was 48% versus 27% for the PTFE. The log rank test also showed that overall secondary patency was superior for KNVD over PTFE (p = 0.011, Fig. 2). Reporting of secondary patency resulted in slightly improved early graft patency rates, but at 5 years secondary patency was equal to primary patency. Because this was a retrospective study, with a similar but not identical distribution of risk factors in
Volume
16
Number
1
Femoropopliteal reconstruction with Dacron versus PTFE 63
July 1992
i-.
100
0 0
8O
>. o to
~
I1.
u E
,'I
60
!ol 3o
1
T
,__
,.
I
I
4o
[
13
1 t
--
12 p = 0.0096
15
11
2o
"c
13_
0
I
0
I
12
I
24
I
36
I
48
60
Months
Fig. 1. Overall primary patency. Solid line, KNVD, n = 38; broken line, PTFE, n = 85.
" r® 0
'>.
o
r-
I00
80 I I
60
--
o
a_
--
1
24
14
-- - - I
40
20
12
12
>. L
0
-o tO 0 (D O3
p = 0.011
I---
20 0
0
I
I
I
I
I
12
24
36
48
60
Months
Fig. 2. Overall secondary patency. Solid line, KNVD, n = 38; broken line, PTFE, n = 85.
the K N V D and PTFE groups, a Cox Proportional Hazards regression was applied. This indicated a significant association between graft type and primary patency, favoring K N V D over PTFE (p = 0.039), after simultaneous adjustment for the effects of smoking, indication, diabetes, runoff, prior proximal reconstruction, principal versus adjunctive procedure, primary versus secondary bypass, and above- versus below-knee outflow. N o significant difference was observed in limb salvage rates for limbs with grafts of K N V D or PTFE (to = 0.079 by log rank, Fig. 3). The 5-year limb salvage rate for K N V D was 79%, and for PTFE it was 61%. (The number o f limbs in Fig. 3 is smaller than the number of grafts in Figs. 1 and 2, because three limbs had more than one graft implanted.)
DISCUSSION
In this series of femoropopliteal reconstructions, the overall patency was better for grafts o f K N V D than for grafts o f P T F E . Chi-square analysis showed the two populations to be similar, suggesting that log rank comparison of the patency curves was appropriate for this retrospective series. To further control for population differences, a Cox Proportional Hazards regression was also applied. This statistical test is designed to simultaneously correct for the disproportionate distribution ofmuhiple risk factors in two populations. The Cox regression confirmed the significant advantage in patency for K N V D over PTFE. The rate of limb salvage was not significantly different for the two graft materials. This is not
64
Journalof VASCULAR SURGERY
Pevec et al.
-,(o
100
T 8O
19
T
- - , --
--
~
(D CL
--
--I 37--
--
T 14
7 I
24
I ----~----~
60 03 0 >
T 16
p = 0.079 19
4O
C~ (f) ..0
2O
E
. - -
--J
0 12
24
36
48
60
Monfhs
Fig. 3. Limb salvage. Solid line, KNVD, n = 37; broken line, PTFE, n = 83. surprising, because limb salvage is mostly dependent on the adequacy of collateral circulation, which is difficult to quantitate and stratify. Comparison with other series of femoropopliteal reconstructions can be misleading, since population characteristics are quite variable, as are results. But such comparisons may be useful to gain a proper perspective regarding the results in this report. The best designed study of prosthetic femoropopliteal reconstructions to date is the prospective, randomized, multicenter study reported by Veith et al.s in 1986. The overall 5-year primary patency for PTFE grafts was 38%, better than the rate of 27% in this series. However, Veith et al. found that in patients without an adequate saphenous vein, the situation that pertained to most of the patients in the present series, the 4-year patency of PTFE was 29%, very similar to the results reported here. Other comparable retrospective series report 5-year primary patency ofPTFE femoropopliteal grafts ranging from 11% to 5 7 % . 9-12
No prospective evaluation of Dacron femoropopliteal grafts was found on review of the literature. A comparable retrospective series of knitted, external velour Dacron grafts had a 5-year secondary patency of 50%, and another series of externally supported Dacron grafts had a 4-year patency of 76%. 13'~ The results of the former study are comparable to, and the latter study are better than, the 48% 5-year patency in the present series. The only previous study directly comparing Dacron and PTFE as femoropopliteal grafts was reported by Rosenthal et al.s in 1990. The type of Dacron used was not specified. The series included only patients receiving above-knee grafts for inter-
mittent claudication. The patency of Dacron and PTFE was similar at 5 years (57% versus 65%) and at 10 years (32% versus 31%). In contrast, the current study showed KNVD to be superior to PTFE in all patients, including those with above-knee grafts and those bypassed for claudication (log rank comparison of primary patency for these subgroups showed KNVD to be superior to PTFE, with p < 0.05, but the number of individuals in each subgroup is small). The KNVD graft was designed to fidfill several theoretic requirements. It was believed that high graft porosity was desirable, because porosity would allow ingrowth of arterioles and other host tissues, resulting in complete graft incorporation and greater durability. However, high porosity makes effective preclotting very difficult and often results in leakage from the graft. Other manufacturers thus decreased porosity by adding velour to the graft surface. However, the process of adding velour increased the amount of foreign material, and was thought to increase the permeability of the Dacron fibers to water, resulting in the eventual decomposition of the graft. The Golaski Labs developed a process to manufacture a graft of slightly lower porosity without velour, theoretically allowing both effective predotting and tissue ingrowth, without creating a susceptibility to graft decomposition. Whether any of these theoretic considerations is truly operative in the good patency and durability described in this report is purely speculative. Although the data in this report suggest that KNVD is superior to IrFFE as a graft material for femoropopliteal reconstruction, this is a small, retrospective study. As such, it is subject to potential
Volume 16 Number 1
july 1992
biases and inaccuracies.Is An attempt was made to minimize the biases by stratifying the patients according to the appropriate risk factors and by applying the Cox Proportional Hazards regression to adjust for the influence of risk factors that may be disproportionately distributed between the two groups. An attempt was made to minimize inaccuracies by adherence to the recommendations of the Ad Hoc Committee on Reporting Standards to the greatest extent possible. Appropriate and rigorous statistical methods were applied. The apparent advantage in patency of KNVD over PTFE in femoropopliteal reconstruction is great enough to justify a prospective, randomized trial comparing these two materials. REFERENCES 1. Haimov H, Giron F, Jacobson JH. The expanded polytetrafluoroethylene graft Arch Surg 1979;114:673-7. 2. Johnson WC. Preliminary experience with expanded polytetrafluoroethylene grafts. Surgery 1979;85:123-8. 3. Raithel D, Groid H. Small artery reconstruction with a new vascular prosthesis. World J Surg 1980;4:223-30. 4. Pradhan DJ, Juanteguy JM, Michelson E, Abosch O. Results of polytetrafluoroethylene grafts in the femoropopliteal region: two-year evaluation of 146 bypass procedures. Am Surg 1981;47:355-8. 5. Rosenthal D, Evans D, McKinsey J, et al. Prosthetic above-knee femoropopliteal bypass for intermittent claudication. J Cardiovasc Surg (Torino) 1990;31:462-8. 6. Rutherford RB, Flanigan DP, Gupta SK, et al. Suggested
Femoropopliteal reconstruction with Dacron versus PTFE
7. 8.
9,
10. 11. 12. 13. 14.
15.
65
standards for reports dealing with lower extremity ischemia. J VAsc SURG 1986;4:80-94. KalbfleischJD, Prentice RL. The statistical analysis of failure time data. New York: John Wiley and Sons, 1980. Veith FJ, Gupta SK, Ascer E, et al. Six-year prospective multicenter randomized comparison of autologons saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J VAsc SuR~ 1986;3:104-14. Harris JI', O'Brien CJ, Stephen MS, Sheil AGR, May J. Should polytetrafluoroethylene grafts be used in preference to saphenous vein for femoropopliteal arterial bypass? Aust N Z J Surg 1985;55:579-83. Quinones-Baldrich WJ, Busuttil RW, Baker JD, et al. Is the preferential use of polytetrafluoroethylene grafts for femoropopliteal bypass justified? J VASC SuRe 1988;8:219-28. Budd JS, Brennan J, Beard JD, Warren H, Burton PR, Bell PRF. Infralnguinal bypass surgery: factors determining late graft patency. Br J Surg 1990;77:1382-7. McAuley CE, Steed DL, Webster MW. Seven-yearfollow-up of expanded polytetrafluoroethylene (PTFE) femoropopliteal bypass grafts. Ann Surg 1984;199:57-60. Mosley JG, Marson A. A 5 year follow-up of Dacron femoropopliteal bypass grafts. Br J Surg 1986;73:24-7. Matsubara J, Nagasue M, Tsuchishima S, Nakatani B, Shimizu T. Clinical results of femoropopliteal bypass using externally supported (EXS) Dacron grafts: with a comparison of above- and below-knee anastomosis. J Cardiovasc Surg (Torino) 1990;31:731-4. Goldman M, Kenchington G. Prospective and retrospective data compared in patients with PTFE femoropopliteal bypass grafts. Ann R Coil Surg Engl 1989;71:243-4.
Submitted Sept. 10, 1991; accepted Nov. 29, 1991.
Expanded polytetrafluoroethylene (PTFE) has become the most popular conduit for femoropopliteal reconstruction when autogenous vein is not used. The use of PTFE has exceeded other alternatives, including allograft vein, heterograft artery, human umbilical vein, and Dacron. This preference for PTFE has developed for several reasons. Its smooth, hydrophobic surface intuitively appears "thromboresistent." It handles well and does not require preclotting, making it attractive to the surgeon. Articles published after its clinical release in 1976 reported excellent early patency rates. ~-4 Dacron was one of the early materials used in these operations. As the use of PTFE for femoropopliteal reconstruction increased, the use of Dacron decreased. However, the preference for PTFE over Dacron is not based on solid scientific evidence. A recent review of the literature found only one large series directly comparing the two prosthetic materials in femoropopliteal reconstructions; that retrospective series showed similar 5- and 10-year patencies for PTFE and Dacron.S Since Dacron is considerably less expensive than PTFE, if PTFE offers no clear From Massachusetts General Hospital, Department of Surgery, Division of Vascular Surgery, Boston. Reprint requests: William Abbott, MD, Division of Vascular Surgery, Massachusetts General Hospital, ACC 458, Boston, MA 02114. 24/1/35441
60
advantage over Dacron, then perhaps the preferential use of PTFE needs to be reassessed. The clinical impression of one of the authors was that a specific configuration of Dacron, a nonvelour, high-porosity knitted graft (the LoPor graft; Golaski Labs, Philadelphia, Pa.; porosity = 2000 to 3000 ml water/cm2/min), had a particularly good patency in the femoropopliteal position. A review was therefore undertaken of the experience of this single surgeon, comparing the results of femoropopliteal reconstructions performed with this material with those performed with PTFE (Gore-Tex; W. L. Gore and Associates, Inc., Elkton, Md.). PATIENTS AND METHODS
A retrospective review was performed of one surgeon's concurrent experience with the two graft materials. All patients undergoing femoropopliteal reconstructions with knitted, nonvelour Dacron (KNVD) or PTFE by one of the authors (R.C.D.) between the dates of June 1,1974, and Jan. 31,1986, were included. The KNVD grafts were implanted from June 1974 through January 1986; the PTFE grafts were implanted from May 1977 through December 1985. During this same time period, autogenous saphenous vein was the conduit of choice. Other conduits, including human umbilical vein, cryopreserved saphenous vein, and other configurations of Dacron, were also used, but they are
Volume 16 Number 1 July 1992
Femoropopliteal reconstruction with Dacron versus PTFE
61
Table I. Demographic and risk factors PTTE
KNVD (n =
n
Age (yr) Male Cardiac disease Carotid disease Hypertension Diabetes Smoking Indication Claudication Rest pain/tissue loss Acute ischemia Adjunctive Redo Size = 6 mm Proximal reconstruction Runoff -< 1 Below-knee
%
n
64.5 +- 10.4 23/38 23/36 13/32 22/36 15/38 29/33
61 64 41 61 39 88
11/37 20/37 6/37 5/38 11/38 33/37 22/38 16/31 17/37
30 54 16 13 29 89 58 52 46
not included in this analysis. The selection of one particular graft material over another did not follow any specific criteria. Rather, the use of multiple different materials reflected the surgeon's dissatisfaction with the performance of each of the nonautogenous materials. Of note, the authors have never had any commercial relationship with Golaski Labs or W. L. Gore and Associates, Inc. Information was obtained through review of the patient's chart, noninvasive vascular laboratory reports, angiogram reports, operative note, and operative diagram. Excluded were grafts to the tibial vessels, short grafts from the distal superficial femoral artery to the popliteal artery for popliteal aneurysms, composite grafts, and grafts from the femoral artery to the contralateral popliteal artery without an ipsilateral groin anastomosis. Thirty-eight KNVD and 85 PTFE grafts met the inclusion criteria. Demographic and risk factors assessed were age; sex; indication; principal versus adjunctive procedure; primary versus secondary operation; presence of proximal arterial reconstruction; inflow vessel; outflow vessel; number of patent runoff vessels; graft size; presence of documented cardiac disease, carotid disease, hypertension, and diabetes mellitus; and history of smoking (a patient who smoked at any time, past or present, was considered a smoker). All terms were defined according to the guidelines suggested by the Ad Hoc Committee on Reporting Standards of the Society for Vascular Surgery and the North American Chapter of the International Society for Cardiovascnlar Surgery. 6
(n = ss)
38)
%
p
63/85 39/72 19/67 37/73 35/82 65/76
74 54 28 51 43 86
0.48 0.13 0.34 0.22 0.31 0.74 0.74 0.30
18/84 48/84 18/84 16/85 15/85 70/81 43/85 31/66 48/84
21 57 21 19 18 86 51 47 57
65.9 - 9.7
0.44 0.16 0.46 i 0.67 ~ 0.26
Patency was assessed by documentation by the operating surgeon (R.C.D.) of a palpable pulse distal to the graft and/or maintenance of the postoperative improvement in pulse volume recordings and Doppler pressures. In some instances patency was documented by angiography or direct intraoperative assessment. In four cases the assessment of another physician was accepted as documentation ofpatency. These patency criteria do not strictly adhere to the guidelines of the Ad Hoc Committee. Strict adherence is difficult in a retrospective series that predates the Committee's report. Every effort was made to follow these guidelines as closely as possible, and in most cases the criteria of the Committee were fulfilled. A graft was considered patent only to the date of the last examination that documented patency. After this date, the graft was considered lost to follow-up. In most instances the date of graft occlusion could be unequivocally determined, became the patient developed acute symptoms that led to prompt evaluation. Less commonly, the graft occluded without symptoms. When an asymptomatic occlusion was diagnosed, the graft was considered to have occluded halfway through the interval between examinations, as recommended by the Ad Hoc C o m m i t t e e . 6 Primary and secondary patency, as defined by the Ad Hoc Committee, are reported. Grafts undergoing thrombectomy or revision within 24 hours of the original operation, with subsequent sustained patency, were considered to be primarily patent. Grafts that occluded more than 24 hours after implantation, in which patency was subsequently reestablished, were considered to be secondarily
62
Journal of VASCULAR SURGERY
Pevec et al.
Table II. Overall patency Interval (too)
Grafts at risk (no.)
Failedgrafts (no.)
Withdrawn grafts (no.)
Interval patency (%)
Cumulative patency (%)
SE (%)
PTFE
0-1 1-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 48-54 54-60
85 72 52 38 30 24 21 18 17 15 12
10 13 13 5 4 2 1 0 0 2 1
3 7 1 3 2 1 2 1 2 1 0
88 82 75 87 87 92 95 100 100 87 92
88 71 54 46 39 36 34 34 34 29 27
3.5 5.1 5.8 5.8 5.8 5.8 5.8 5.8 5.8 5.9 5.9
KNVD
0-1 1-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 48-54 54-60
38 33 27 24 20 19 17 16 14 13 12
5 4 2 3 1 0 1 1 1 0 0
0 2 1 1 0 2 0 1 0 1 0
87 88 93 88 95 100 94 94 93 100 100
87 76 70 61 58 58 55 51 48 48 48
5.5 7.0 7.6 8.2 8.3 8.3 8.5 8.6 8.8 8.8 8.8
patent. Limb salvage, defined as the proportion of grafted limbs maintained without below- or aboveknee amputation, is also reported. Chi-square analysis was used to compare the demographic and risk factors between the two groups of patients. Student's t test was used to compare the ages of the two groups. Percent graft patency and limb salvage rate were determined by the Kaplan-Meier (product limit) method. Patency and limb salvage curves were compared by the log rank test. In addition, Cox Proportional Hazards regression was used to evaluate the influence of graft type (KNVD or PTFE) on primary patency while adjusting simultaneously for the effects of all listed risk factors, according to the method described by Kalbfleisch and Prentice. 7 All statistical computations were performed by use of SAS software (SAS Institute, Cary, N.C.). RESULTS No statistically significant differences occurred in demographic or risk factors between the KNVD and PTFE groups (Table I). One death occurred shortly after operation in the KNVD group (operative mortality rate, 2.8%), and three occurred in the PTFE group (operative mortality rate, 3.8%). Two deaths were due to myocardial infarction, one was due to a perforated gallbladder, and one was due to bowel infarction. Four
complications occurred in the KNVD group (morbidity, 10.5%), and 15 complications occurred in the PTFE group (morbidity, 17.6%). Acute thrombectomy and/or graft revision (less than 24 hours after the original operation), with subsequent sustained patency, were performed in one KNVD (2.6%) and six PTFE (7.1%) grafts. One pseudoaneurysm developed in each group. A KNVD graft developed a pseudoaneurysm at the popliteal anastomosis, became of disruption of the Dacron prosthetic, 8 years and 6 months after implantation. One PTFE graft developed a pseudoaneurysm at the site of the proximal anastomosis to a previously placed aortofemoral prosthesis, 5 years and 10 months after implantation of the femoropopliteal graft. The patency curves for the two graft materials were compared by use of the log rank test. Overall primary patency of the KNVD grafts was superior to the primary patency of the PTFE grafts (p = 0.0096, Table LI, Fig. 1). Five-year primary patency of the Dacron was 48% versus 27% for the PTFE. The log rank test also showed that overall secondary patency was superior for KNVD over PTFE (p = 0.011, Fig. 2). Reporting of secondary patency resulted in slightly improved early graft patency rates, but at 5 years secondary patency was equal to primary patency. Because this was a retrospective study, with a similar but not identical distribution of risk factors in
Volume
16
Number
1
Femoropopliteal reconstruction with Dacron versus PTFE 63
July 1992
i-.
100
0 0
8O
>. o to
~
I1.
u E
,'I
60
!ol 3o
1
T
,__
,.
I
I
4o
[
13
1 t
--
12 p = 0.0096
15
11
2o
"c
13_
0
I
0
I
12
I
24
I
36
I
48
60
Months
Fig. 1. Overall primary patency. Solid line, KNVD, n = 38; broken line, PTFE, n = 85.
" r® 0
'>.
o
r-
I00
80 I I
60
--
o
a_
--
1
24
14
-- - - I
40
20
12
12
>. L
0
-o tO 0 (D O3
p = 0.011
I---
20 0
0
I
I
I
I
I
12
24
36
48
60
Months
Fig. 2. Overall secondary patency. Solid line, KNVD, n = 38; broken line, PTFE, n = 85.
the K N V D and PTFE groups, a Cox Proportional Hazards regression was applied. This indicated a significant association between graft type and primary patency, favoring K N V D over PTFE (p = 0.039), after simultaneous adjustment for the effects of smoking, indication, diabetes, runoff, prior proximal reconstruction, principal versus adjunctive procedure, primary versus secondary bypass, and above- versus below-knee outflow. N o significant difference was observed in limb salvage rates for limbs with grafts of K N V D or PTFE (to = 0.079 by log rank, Fig. 3). The 5-year limb salvage rate for K N V D was 79%, and for PTFE it was 61%. (The number o f limbs in Fig. 3 is smaller than the number of grafts in Figs. 1 and 2, because three limbs had more than one graft implanted.)
DISCUSSION
In this series of femoropopliteal reconstructions, the overall patency was better for grafts o f K N V D than for grafts o f P T F E . Chi-square analysis showed the two populations to be similar, suggesting that log rank comparison of the patency curves was appropriate for this retrospective series. To further control for population differences, a Cox Proportional Hazards regression was also applied. This statistical test is designed to simultaneously correct for the disproportionate distribution ofmuhiple risk factors in two populations. The Cox regression confirmed the significant advantage in patency for K N V D over PTFE. The rate of limb salvage was not significantly different for the two graft materials. This is not
64
Journalof VASCULAR SURGERY
Pevec et al.
-,(o
100
T 8O
19
T
- - , --
--
~
(D CL
--
--I 37--
--
T 14
7 I
24
I ----~----~
60 03 0 >
T 16
p = 0.079 19
4O
C~ (f) ..0
2O
E
. - -
--J
0 12
24
36
48
60
Monfhs
Fig. 3. Limb salvage. Solid line, KNVD, n = 37; broken line, PTFE, n = 83. surprising, because limb salvage is mostly dependent on the adequacy of collateral circulation, which is difficult to quantitate and stratify. Comparison with other series of femoropopliteal reconstructions can be misleading, since population characteristics are quite variable, as are results. But such comparisons may be useful to gain a proper perspective regarding the results in this report. The best designed study of prosthetic femoropopliteal reconstructions to date is the prospective, randomized, multicenter study reported by Veith et al.s in 1986. The overall 5-year primary patency for PTFE grafts was 38%, better than the rate of 27% in this series. However, Veith et al. found that in patients without an adequate saphenous vein, the situation that pertained to most of the patients in the present series, the 4-year patency of PTFE was 29%, very similar to the results reported here. Other comparable retrospective series report 5-year primary patency ofPTFE femoropopliteal grafts ranging from 11% to 5 7 % . 9-12
No prospective evaluation of Dacron femoropopliteal grafts was found on review of the literature. A comparable retrospective series of knitted, external velour Dacron grafts had a 5-year secondary patency of 50%, and another series of externally supported Dacron grafts had a 4-year patency of 76%. 13'~ The results of the former study are comparable to, and the latter study are better than, the 48% 5-year patency in the present series. The only previous study directly comparing Dacron and PTFE as femoropopliteal grafts was reported by Rosenthal et al.s in 1990. The type of Dacron used was not specified. The series included only patients receiving above-knee grafts for inter-
mittent claudication. The patency of Dacron and PTFE was similar at 5 years (57% versus 65%) and at 10 years (32% versus 31%). In contrast, the current study showed KNVD to be superior to PTFE in all patients, including those with above-knee grafts and those bypassed for claudication (log rank comparison of primary patency for these subgroups showed KNVD to be superior to PTFE, with p < 0.05, but the number of individuals in each subgroup is small). The KNVD graft was designed to fidfill several theoretic requirements. It was believed that high graft porosity was desirable, because porosity would allow ingrowth of arterioles and other host tissues, resulting in complete graft incorporation and greater durability. However, high porosity makes effective preclotting very difficult and often results in leakage from the graft. Other manufacturers thus decreased porosity by adding velour to the graft surface. However, the process of adding velour increased the amount of foreign material, and was thought to increase the permeability of the Dacron fibers to water, resulting in the eventual decomposition of the graft. The Golaski Labs developed a process to manufacture a graft of slightly lower porosity without velour, theoretically allowing both effective predotting and tissue ingrowth, without creating a susceptibility to graft decomposition. Whether any of these theoretic considerations is truly operative in the good patency and durability described in this report is purely speculative. Although the data in this report suggest that KNVD is superior to IrFFE as a graft material for femoropopliteal reconstruction, this is a small, retrospective study. As such, it is subject to potential
Volume 16 Number 1
july 1992
biases and inaccuracies.Is An attempt was made to minimize the biases by stratifying the patients according to the appropriate risk factors and by applying the Cox Proportional Hazards regression to adjust for the influence of risk factors that may be disproportionately distributed between the two groups. An attempt was made to minimize inaccuracies by adherence to the recommendations of the Ad Hoc Committee on Reporting Standards to the greatest extent possible. Appropriate and rigorous statistical methods were applied. The apparent advantage in patency of KNVD over PTFE in femoropopliteal reconstruction is great enough to justify a prospective, randomized trial comparing these two materials. REFERENCES 1. Haimov H, Giron F, Jacobson JH. The expanded polytetrafluoroethylene graft Arch Surg 1979;114:673-7. 2. Johnson WC. Preliminary experience with expanded polytetrafluoroethylene grafts. Surgery 1979;85:123-8. 3. Raithel D, Groid H. Small artery reconstruction with a new vascular prosthesis. World J Surg 1980;4:223-30. 4. Pradhan DJ, Juanteguy JM, Michelson E, Abosch O. Results of polytetrafluoroethylene grafts in the femoropopliteal region: two-year evaluation of 146 bypass procedures. Am Surg 1981;47:355-8. 5. Rosenthal D, Evans D, McKinsey J, et al. Prosthetic above-knee femoropopliteal bypass for intermittent claudication. J Cardiovasc Surg (Torino) 1990;31:462-8. 6. Rutherford RB, Flanigan DP, Gupta SK, et al. Suggested
Femoropopliteal reconstruction with Dacron versus PTFE
7. 8.
9,
10. 11. 12. 13. 14.
15.
65
standards for reports dealing with lower extremity ischemia. J VAsc SURG 1986;4:80-94. KalbfleischJD, Prentice RL. The statistical analysis of failure time data. New York: John Wiley and Sons, 1980. Veith FJ, Gupta SK, Ascer E, et al. Six-year prospective multicenter randomized comparison of autologons saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J VAsc SuR~ 1986;3:104-14. Harris JI', O'Brien CJ, Stephen MS, Sheil AGR, May J. Should polytetrafluoroethylene grafts be used in preference to saphenous vein for femoropopliteal arterial bypass? Aust N Z J Surg 1985;55:579-83. Quinones-Baldrich WJ, Busuttil RW, Baker JD, et al. Is the preferential use of polytetrafluoroethylene grafts for femoropopliteal bypass justified? J VASC SuRe 1988;8:219-28. Budd JS, Brennan J, Beard JD, Warren H, Burton PR, Bell PRF. Infralnguinal bypass surgery: factors determining late graft patency. Br J Surg 1990;77:1382-7. McAuley CE, Steed DL, Webster MW. Seven-yearfollow-up of expanded polytetrafluoroethylene (PTFE) femoropopliteal bypass grafts. Ann Surg 1984;199:57-60. Mosley JG, Marson A. A 5 year follow-up of Dacron femoropopliteal bypass grafts. Br J Surg 1986;73:24-7. Matsubara J, Nagasue M, Tsuchishima S, Nakatani B, Shimizu T. Clinical results of femoropopliteal bypass using externally supported (EXS) Dacron grafts: with a comparison of above- and below-knee anastomosis. J Cardiovasc Surg (Torino) 1990;31:731-4. Goldman M, Kenchington G. Prospective and retrospective data compared in patients with PTFE femoropopliteal bypass grafts. Ann R Coil Surg Engl 1989;71:243-4.
Submitted Sept. 10, 1991; accepted Nov. 29, 1991.