Prospective, randomized comparison of ringed and nonringed polytetrafluoroethylene femoropopliteal bypass grafts: A preliminary report Sushil K. Gupta, MD, Frank J. Veith, MD, Harry B. Kram, MD, and Kurt R. Wengerter, MD, New York, N.Y. Kinking and compression with knee flexion are thought to be one cause of failure of below-knee polytetrafluoroethylene femoropopliteal bypass. To prevent this problem polytetrafluoroethylene grafts externally supported with rigid rings have been developed. The present randomized, prospective study compared ringed and nouringed polytetrafluoroethylene grafts in 122 patients who underwent femoropopliteal bypass for severe limb ischemia. Patients were well matched for surgical indications and risk factors. There was no significant difference in the 3-year graft patency rate of ringed versus nonringed polytetrafluoroethylene femoropopliteal bypasses (74% vs 68%, p = 0.5). Similarly, no significant differences were found in the 3-year graft patency rates of ringed versus nonringed above-knee (82% vs 74%, p = 0.5) or below-knee polytetrafluoroethylene femoropopliteal bypasses (68% vs 59%, p = 0.5). The 3-year graft patency rate of all aboveknee polytetrafluoroethylene femoropopliteal bypasses was slightly greater than that of below-knee polytetrafluoroethylene femoropopliteal bypasses (76% vs 62%), but this difference was not statistically significant (p = 0.25). The 3-year limb salvage rate with ringed polytetrafluoroethylene grafts was 92% compared with 79% for nonringed polytetrafluoroethylene grafts, but this difference was not statistically significant (p = 0.25). Data to date from this study fail to support the recommendation that ringed polytetrafluoroethylene grafts be used preferentially over conventional polytetrafluoroethylene grafts in patients who require femoropopliteal bypass with a synthetic graft. (J VASe SURG 1991;13:162-72.)
Femoropopliteal bypass has become well established as a safe and effective operation for the treatment of severe lower extremity ischemia.13 Although autogenous saphenous vein (ASV) provides the most durable conduit for arterial reconstruction below the inguinal ligament, it may be unavailable or inadequate for femoropopliteal bypass in many patients. 4's Furthermore, in some patients use of prosthetic grafts has been recommended because their life expectancy is limited to 2 to 3 years or to limit the operating time because of the poor medical condition of the patient. ~ Accordingly, several artifiFrom the Division of Vascular Surge~, Montefiore Medical Center/Albert Einstein College of Medicine. Supported in part by The Manning Foundation, The Anna S. Brown Trust, The Renate and Allan B. Hunter Surgical Research Fund, and The New York Institute for Vascular Studies. Presented at the Forty-fourth Annual Meeting of the Society for Vascular Surge~, Los Angeles, Calif., June 4-6, 1990. Reprint requests: Sushil K. Gupta, MD, Division of Vascular Surgery, Montefiore Medical Center, 111 E. 210th St., New York, NY 10467. 24/6/25646
cial arterial conduits have been tried as alternatives to ASV for femoropopliteal bypass, with expanded polytetrafluoroethylene (PTFE) grafts having emerged as the current first choice of most vascular surgeons. Polytetrafluoroethylene grafts were first used clinically for femoropopliteal bypass in 1976. 6 Since then, several studies have shown that PTFE grafts inserted in the femoropopliteal position function well, but that the long-term results of below-knee femoropopliteal bypass are inferior to those achieved with above-knee femoropopliteal bypass. 79 It is thought that this may in part be due to diminished bypass flow brought about by kinking and compression of prosthetic grafts with flexion as they cross the knee joint.~° To obviate this factor as a possible cause of prosthetic femoropopliteal bypass failure, several graft modifications have been developed, including external support of the prostheses with fluorinated ethylene propylene (FEP) in ringed, spiral, and double-helix configurations J °-16 Although we have postulated that below-knee 163
164
Journal of VASCULAR SURGERY
Gupta et al.
Table I. Indications for operation Ringed (58)
Nonringed (64)
Indication
No.
%
No.
%
p value
Claudication Rest pain Ulcer/necrosis
1 16 41
2% 28% 70%
3 11 50
5% 17% 78%
0.8 0.2 0.6
No significant differences occurred in the indications for surgery in patients who received ringed and nonringed PTFE grafts. Values listed are numbers of patients.
prosthetic graft failure is multifactorial, we were hopeful that these rigid grafts would improve the results. However, this hypothesis remained unproven, and hence the present prospective, randomized comparison of ringed and nonringed PTFE grafts was tmdertaken in our patients who were candidates for a prosthetic femoropopliteal bypass. Since we had observed kinking caused by knee flexion in some above-knee grafts as well, these cases were also included in the randomization protocol. Furthermore, the inclusion of above-knee grafts in the study also allowed us to compare the behavior of grafts that extended below the knee to those that terminated above the knee. This article reports the initial results of this study, which began on Jan. 1, 1986. MATERIAL A N D M E T H O D S Patient selection
On the basis of a history of prior removal, duplex ultrasonography, or saphenous venography, all patients without an available ipsilateral ASV long enough to serve as a femoropopliteal bypass, as well as those patients whose life expectancy was judged to be less than 3 years, were randomized to receive either ringed or nonringed PTFE grafts. This group was augmented by patients who were scheduled to receive an ASV graft but who at operation werc found to have an unsuitable ASV because of inadequate length, small size, or fibrotic changes. Randomization techniques were similar to those previously used in a randomized trial comparing ASV and PTFE grafts for infrainguinal arterial reconstructions and consisted of selecting a random card from an unsorted deck of cards marked with the choice of graft material. 4's Patients were selected for aboveknee femoropopliteal bypass when the superficial femoral artery was occluded with reconstitution at the above-knee popliteal artery, unless there coexisted a popliteal artery stenosis greater than 40% behind the knee, in which case below-knee femoropopliteal bypass was performed? The presence of a distal popliteal artery occlusion rendering the patent popliteal artery an isolated segment was not a contraindication
for inclusion in the study, and our criteria for use of an isolated popliteal segment were followed.2'~7 However, patients with extensive necrosis requiring sequential grafts to distal arteries were excluded from randomization. Patients who required a femoropopliteal bypass for reasons other than arteriosclerotic occlusive disease were also excluded from the study. Ninety-one (75%) patients were operated on for tissue necrosis or nonhealing foot ulceration; 27 (22%) for rest pain; and 4 (3%) for claudication. No significant differences were found in the frequency of any of the indications for femoropopliteal bypassing in patients who received ringed versus nonringed PTFE grafts (Table I). The mean age and standard deviation (SD) of the patients was 71 +_ 12 years (range, 27 to 93 years). A high incidence of associated medical illness was found in our patients, and many had significant risk factors such as diabetes, hypertension, smoking, cardiac disease, and previous vascular surgery in the same leg. The patient populations in both groups were evenly matched in regard to these risk factors (Table II). The popliteal outflow tract was quantified as to the number of named patent vessels originating from the popliteal artery. If no named tibial or peroneal artery originated directlv from the popliteal artery, the outflow tract was noted to be isolated. No significant difference was found in the ringed and the nonringed groups in regard to the number of patent runoff vessels (Table II). Efforts to achieve maximal improvement in medical status were instituted before the operation. However, in some patients, the urgent need for revascularization made it necessary to proceed with surgery before optimal medical therapy was completed. 2 Technical considerations
Femoropopliteal bypasses originated from the common femoral artery in 89 (73%) patients and the superficial femoral artery, in 33 (27%) patients. Distal (popliteal) anastomoses were performed above the knee in 59 (48%) patients and below the knee in 63 (52%) patients. Incisions were made in the groin for exposure of the femoral arteries and medially above
Volume 13 Number 1 January 1991
Randomized comparison of ringed and nonringed PTFE grafts
165
Table II. Risk factors and the number of popliteal outflow vessels in patients who underwent femoropopliteal bypass Ringed (58)
Risk factors Diabetes Hypertension Smoking Cardiac disease Previous bypass Popliteal runoff vessels Isolated popliteal arteu 1 vessel 2 vessels 3 vessels
Nonringed (64)
No.
%
No.
%
p value
35 29 32 26 21
60% 50% 55% 45% 36%
37 28 28 26 22
58% 44% 44% 41% 34%
0.9 0.8 0.2 0.4 0.9
18 22 10 8
31% 38% 17% 14%
23 22 15 4
36% 34% 24% 6%
0.8 0.9 0.5 0.2
Note that there was no significant difference in the incidence of associated risk factors or the distribution of outflow vessels in patients who received ringed and nonringed IrI'FE grafts. Values listed are numbers of patients.
or below the knee for exposure of the popliteal artery. A fascial incision along the anterior border of the sartorius muscle was made to expose the above-knee popliteal artery. The medial head of the gastrocnemius muscle was retracted posteriorly to expose the below-knee popliteal artery. A subsartorial tunnel was created by use of blunt dissection with the fingers or a standard graft tunneling instrument passed from the groin to the popliteal fossa. To reach the belowknee popliteal fossa, an anatomic tunnel was made along the popliteal neurovascular bundle. Meticulous technique and operative adjuncts such as headlight illumination and magnification were used for all bypasses. Grafts of 6 mm PTFE (W. L. Gore & Associates, Inc, Elkton, Md.) were used for all procedures performed in this study. The anastomoses were performed with 6.0 polypropylene or 7.0 IH'FE sutures. Systemic heparin in an initial dose of 100 mg/kg was used and was not reversed in most cases. Aspirin and dipyridamole (Persantine) were started 2 to 3 days before surgery and continued indefinitely if tolerated and the medical condition allowed their use. 2 Special attention was given to fashioning the anastomosis of the PTFE grafts so that they coursed without twisting, redundancy, or angnlation with the knee flexed or extended. This required that the inelastic PTFE grafts be cut to precisely the correct length. All anastomoses were meticulously performed taking care to include all layers of the arterial wall in the anastomosis and to avoid any exposed media or adventitial tags within the anastomotic lumen. 2 In patients who received ringed PTFE grafts, only the most proximal and distal rings were removed from the graft, allowing for external reinforcement
of the prosthesis to extend to within 2 cm of the anastomoses. Patient follow-up Patients were examined every 2 to 4 weeks for the first 3 months after operation and then every 3 months thereafter. Graft patency was confirmed by unequivocal pulse examination, ankle Doppler determined pressure measurements, pulse volume recordings, and duplex examination. Contrast arteriography was performed in patients who had a change in pulse examination, a significant decrease in any o f the noninvasive parameters, or the recurrence ofischemic symptoms. Grafts determined to be in the "failing" state (i.e., to be patent but associated with hemodynamically significant lesions that might cause graft thrombosis) were treated with transluminal angioplasty, local revision, or short extensions proximal or distal to the graft? s'19 However, these grafts were considered failed for the purposes of primary pateneT life-table calculations.
Data analysis Three-year graft pateneT rates were tabulated by the cumulative life-table method. Patency intervals were terminated at the last known follow-up period or when patients died or were irretrievably lost to follow-up. Patients requiring major amputation despite a patent femoropopliteal bypass had their pateneT interval terminated at that rime. Primary pateneT intervals were terminated at the first graft thrombosis or at any reintervention for a failing graft. Results of these reinterventions and the effect of attempted thrombectomy or revisions o f failed grafts are shown by secondary patency calculations. 2,2° For
166
Journal of VASCULAR SURGERY
Gupta et M.
100
•
58
36
80
"''''''''! 60
..........
19 17
1,
40
20
0 0
I
I
I
I
I
6
12
18
24
30
86
MONTHS AFTER OPERATION FP W I T H RINGED PTFE
.....
FP
WITH
RINGED Figure
NONPTF'E
1
Fig. 1. Cumulative 3-year primary, graft patency of ringed and nonringed FTFE femoropopliteal bypass grafts (FP). No significant difference was found between the two groups (p = 0.5).
T a b l e I I I . Cumulative 3-year graft patency o f all ringed and n o n r i n g e d P T F E femoropopliteal bypass grafts (p = 0.5) Interval
Ringed 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo Nonringed 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
A t risk
Failed
Withdrawn
Interval patency
Cumulative patency
Standard error
58 49 36 27 20 17 15
3 2 3 1 1 0 0
6 11 6 6 2 2 2
95 95 91 96 95 100 100
95 90 82 79 74 74 74
2.89 4.03 5.79 6.99 8.41 9.12 9.71
64 55 42 33 24 19 17
3 4 1 2
6 9 8 7
95 92 97 93
95 88 85 79
2.63 4.1 5.05 6.26
2
3
91
72
7.78
1 0
1 3
95 100
68 68
8.82 9.32
calculating limb salvage rates, the end point was a major amputation below or above the knee regardless o f the status o f the patency o f the graft. T h e significance o f the differences in graft patency rates in patients w h o received ringed versus n o n r i n g e d P T F E grafts and above-knee versus below-knee femoropopliteal bypasses was determined by the log rank test, and the standard error for each interval was calculated by means o f the m e t h o d o f Peto et al.20 Statistical significance was assumed at the 95% level
(p < 0.05). T h e chi-square m e t h o d was used to test the significance o f differences in the indications for surgery, the patient risk factors, and the popliteal o u t f l o w vessels. RESULTS Primary patency W h e n above-knee and below-knee femoropopliteal bypasses were considered together, n o significant difference was f o u n d in the 3-year primary patency
Volume 13 Number 1 January 1991
Randomized comparison of ringed and nonringed PTFE grafls
167
19
100 .
1
;v~
9
80
"'**
8
,.,
" t
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
12 11
1(
40
20
0
I
0
6
I
I
I
12
18
24
I
36
30
MONTHS AFTER OPERA T/ON RINGED
AK
.....
NONRINGED
AK
Fig. 2. Cumulative 3-year primary graft paten~ of above-knee (AK) ringed and nonringed PTFE femoropopliteal bypass grafts. No significant difference was found between the two groups (p = 0.5). Table IV. Cumulative 3-year graft patency of above-knee (AK) ringed and nonringed PTFE femoropopliteal bypass grafts (p = 0.5) Interval Ringed above-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo Nonringed above-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
At risk
Failed
Withdrawn
Interval patency
Cumulative patency
Standard error
29 25 19 14 10 9 9
1 1 2 1 1 0 0
3 5 3 4 1 0 1
96 96 89 100 100 100 100
96 92 82 82 82 82 82
3.41 5.18 8.03 9.36 10.07 10.67 10.67
30 26 23 20 16 12 11
1 1 0 1 2 0 0
3 2 3 3 2 1 1
96 96 100 94 87 100 100
96 93 93 87 74 74 74
3.3 4.93 5.24 6.89 9.31 10.25 10.75
rates of ringed and nonringed PTFE grafts (74% vs 68%,p -- 0.5) (Table III; Fig. 1). Similarly, no significant differences were found in the 3-year primary graft patcncy rates of ringed and nonringed PTFE above-knee bypasses (82% vs 74%,p = 0.5) (Table IV; Fig. 2) or below-knee PTFE femoropopliteal bypasses (68% vs 59%, p = 0.5) (Table V; Fig. 3). The 3-year primary graft patency rate of all aboveknee PTFE femoropopliteal bypasses was superior to that of below-knee PTFE femoropopliteal bypasses (76% vs 62%), but this difference was not statistically significant (p = 0.25) (Table VI; Fig. 4).
Secondary patency Improvement occurred in the secondary 3-year patency rates of ringed PTFE femoropopliteal grafts from 72% to 88% (p = 0.25) as a result of secondary procedures performed to restore patency or correct defects causing graft thrombosis (failed grafts) or hemodynamic failure (failing grafts). In contrast, the secondary patency rates of nonringed PTFE grafts were improved only slightly (from 68% to 71%, p = 0.9) after these secondary procedures. Four grafts failed in a total of 10 bypasses requiting reintervention in the tinged group and only one graft
168
Journal of VASCULAR SURGERY
Gupta et al.
100
I~" Z'~ r.
17
13 10
= ~
1T
8
6
5
80
I
"
60
!"1
......... 6
4
40
20
0
i 0
i
6
RINGED
I
I
i
12 18 24 MONTHS A F T E R OPERATION .....
BK
NONRINGED
30
36
BK
Fig. 3. Cumulative 3-year primary graft paten~ of below-knee (BK) ringed and nonringed PTFE femoropopliteal bypass grafts. No significant difference was found between the two groups (p = 0.5). Table V. Cumulative 3-year graft patency of below-knee (BK) ringed and nonringed PTFE femoropopliteal bypass grafts (p = 0.5) Interval Ringed below-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 m o 18-24 mo 24-30 mo 30-36 mo Nonringed below-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
A t risk
Failed
Withdrawn
Interval patency
Cumulative patency
Standard error
29 24 17 13 10 8 6
2 1 1 1 1 0 0
3 6 3 2 1 2 1
93 95 94 92 89 100 100
93 88 83 76 68 68 68
4.6 6.1 8.3 9.8 10.1 12.1 12.4
34 29 19 13 8 7 6
2 3 1 1 0 1 0
3 7 5 4 1 0 2
94 88 94 91 100 85 100
94 82 77 70 70 59 59
3.9 6.3 8.4 9.8 10.6 11.9 13.1
failed of 13 requiring reintervention in the nonringed group. Limb salvage
Life-table 3-year limb salvage rates were 92% for patients with ringed PTFE grafts and 79% for those with nonringed grafts (p = 0.25) (Table VII). The 3-year limb salvage rates of above-knee and belowknee ringed grafts were 93% and 91%, respectively. The 3-year limb salvage rates of above-knee and below-knee nonringed grafts were 83% and 73%,
respectively (p = 0.5). No statistically significant difference was found between ringed abovc-knce and nonringed above-knee grafts (p = 0.5) and/or between ringed below-knee and nonringed below-knee femoropopliteal bypasses (p = 0.1). However, the number of patients available for comparison in these subgroups was small. In each group of patients who received ringed and nonringed grafts, one postoperative death occurred (within 30 days), resulting in an operative mortality rate of less than 2%.
Volume 13 Number 1 January. 1991
Randomized comparison of ringed and nonringed PTFE grafts
100
34
169
26 21
20
1
f
80 X.
"i! ......... ......... l- ........
60
15
_L
I
[
40
20
0
I
0
I
6
ALL
I
12 18 24 MONTHS AFTER OPERA T/ON AK
PTFE
.....
FEM POPS
ALL
BK
30
36
PTFE
FEM POPS
Fig. 4. Cumulative 3-year primaqr graft patcn~ of all above-knee (AK) and all below-knee (BK) PTFE femoropopliteal bypass grafts (FEM POPS). The 3-year patency rate of aboveknee PTFE femoropopliteal bypass grafts was marginally better than that of below-knee PTFE femoropoplitea! bypass grafts (p = 0.25).
Table VI. Cumulative 3-year graft patency of all above-knee and all below-knee PTFE femoropopliteal bypass grafts (p = 0.25) Interval Above-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo Below-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
At risk
Failed
Withdrawn
Interval patency
Cumulative patency
Standard error
59 51 42 34 26 21 20
2 2 2 1 2 0 0
6 7 6 7 3 1 2
96 96 95 97 92 100 100
96 92 88 85 76 76 76
2.3 3.5 4.7 5.6 7.1 7.9 8.1
63 53 36 26 18 15 12
4 4 2 2 1 1 0
6 13 8 6 2 2 3
93 91 93 91 94 93 100
93 85 80 73 68 62 62
3.1 4.5 5.6 7.4 9.1 9.9 10.3
DISCUSSION
Although ASV graft patency is superior to PTFE graft patency after 2 years, a substantial proportion of patients undergoing femoropopliteal bypass will require a prosthetic graft because the ipsilateral vein is absent or inadequate or because the patient has limited life expectancy, s Moreover, some surgeons use PTFE grafts preferentially for femoropopliteal bypasses for a variety of reasons. 7'22
Kinking and compression of prosthetic grafts with knee flexion has been postulated to be one cause of graft thrombosis and a major reason that belowknee prosthetic grafts have poorer long-term patency then those inserted above the knee. Externally supported Dacron grafts were first introduced by Kenney et al. 1° as a method of preventing kinking and compression. Since then the concept of external support to reduce kinking and compression of PTFE
170
Journal of VASCULAR SURGERY
Gupta et al.
Table VII. Cumulative 3-year limb salvage rates for ringed and nonringed PTFE femoropopliteal bypass grafts (p = 0.25) Interval Ringed 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo Nonringed 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
A t risk
Amputations
Withdrawn
Interval paten~
Cumulative paten~
Standard error
58 49 36 26 19 17 15
1 1 1 0 0 0 0
8 12 9 7 2 2 2
98 98 97 100 100 100 100
98 96 92 92 92 92 92
1.7 2.7 4.1 4.8 5.7 6.1 6.4
64 55 42 33 24 19 17
3 5 0 0 1 0 0
6 8 9 9 4 2 3
95 90 100 100 95 100 100
95 86 86 86 79 79 79
2,6 4.3 4.9 5.6 7.1 8,1 8,5
grafts has been introduced and is used widely, particularly when grafts are placed across areas of flexion or bony prominences. Several clinical reports suggesting better PTFE graft paten~ rates with the use of externally supported grafts have appeared in the literature, l°'11.~s'~s However, these reports are inconclusive for several reasons including too few patients, l° limited foUow-up, ~ or no control groups.13'15 Data from the present randomized, prospective study fail to document a clear advantage of externally supported ringed PTFE grafts as compared to unsupported, nonringed PTFE grafts when used for femoropopliteal bypass. Our findings are supported by two experimental canine studies, which compared ringed and nonringed PTFE grafts and failed to document increased graft patency rates of ringed grafts.~2"~4Moreover, Geiger et a1.12 showed by postoperative arteriography that graft kinking and constriction could still occur in PTFE grafts that were externally supported with FEP in double-helix and ring configurations. External support of PTFE grafts with rings or other configurations, does not necessarily prevent graft kinking, compression, and reduced flow. Moreover, several other studies failed to show a decrease in bypass flow when below-knee femoropopliteal bypass grafts constructed with conventional PTFE prostheses were subjected to high degrees of knee flexion.23a4 It is interesting to note that the patency rates of above-knee grafts (76%) were consistently superior to all those terminating below the knee (62%) (Table VI, Fig. 4). However,
this difference was not statistically significant (p = 0.25) because of the fewer numbers of patients in each group and the relatively short duration of follow-up. It is difficult to speculate why above-knee PTFE grafts typically perform better than belowknee PTFE grafts. Since the paten~ rate of ringed above-knee grafts (84%) (Table IV) was superior to ringed below-knee grafts (68%) (Table V), avoidance of kinking and compression by adding rings does not appear to make a significant difference. We and others have always believed that fcmoropopliteal bypass graft failure is caused by multiple factors, and that it cannot be prevented by simply obviating a single factor, 5"9"17"19"25Because of this and because there are disadvantages to using ringed grafts, we thought it essential to determine how important external support was before using it routinely. Although no clear advantage can yet be demonstrated for externally supported ringed grafts, our data may still be too prelimina~ to show it. We will therefore continue our study with further enrollment of patients and additional follow-up ofthosc already enrolled. However, the results to date suggest that PTFE grafts should not be routinely externally supported. Moreover, the potential advantages of ringed grafts used in other positions should be clearly shown by adequate controlled studies before they are adopted for routine use. This opinion is reinforccd by the fact that ringed grafts have intrinsic disadvantages. They are technically more difficult to inscrt because they are less malleable. This property can actually increase their tendency to kink at the junction
Volume 13 Number 1 January 1991
Randomized comparison of ringed and nonringed PTFE grafts 171
between the ringed and nonringed portion of the graft. This mandates that the surgeon carry the ringed portion of the graft to within 1 to 2 cm of the anastomosis if the latter is in proximity to a joint. Another disadvantage to ringed grafts is that patency is more difficult to document by graft pulse palpation. This may make it necessary to use noninvasive testing more frequently to detect failing grafts before they thrombose. Such detection has been shown to be important in achieving the ultimate objective of limb salvage in many of these patients. 18,~9The difficulty of palpating pulses in ringed PTFE grafts also makes these grafts more difficult to find and dissect out if an extension or a reoperation on a failing graft is required. Despite these considerations the possibility that ringed grafts offer major advantages when used for other ~pes of bypasses or minor advantages when they are used for certain femoropopliteal bypasses demands further study. It is certainly possible that the externally supported prosthetic grafts may be advantageous when used for axillofemoral or other extraanatomic bypasses in which a bypass may be subjected to external compression against an unyielding chest wall or bony structure. 26 In summary the present study comparing ringed and nonringed PTFE grafts for femoropopliteal bypass failed to document a significant advantage to the former in terms of 3-year graft patency rates in either above-knee or below-knee positions. At the present time, ringed PTFE grafts cannot be preferentially recommended over conventional PTFE grafts for routine use in patients who require a prosthetic femoropopliteal bypass. Further controlled trials comparing ringed and nonringed PTFE grafts in patients who require other arterial reconstructions with synthetic grafts are indicated to clarify the real advantages, if any, of these externally supported prostheses.
REFERENCES 1. Maini BS, Mannick JA. Effect of arterial reconstruction on limb salvage: a ten-year appraisal. Arch Surg 1978; 113:1297304. 2. Veith FJ, Gupta SK, Samson RH, et al. Progress in limb salvage by reconstructive arterial surgery combined with new or improved adjunctive procedures. Ann Surg 1981; 194:386401. 3. Veith FJ, Gupta SK, Wengerter KR, et al. Changing arteriosclerotic disease patterns and management strategies in lower limb threatening ischemia. Ann Surg 1990;212:402-14. 4. Bergan JJ, Veith FJ, Bernhard VM, et al. Randomization
of autogenous vein and polytetrafluoroethylene grafts in femoral-distal reconstruction. Surgery, 1982;92:921-30. 5. Veith FJ, Gupta SK, Ascer E, et al. Six-year prospective multicenter randomized comparison of autologous saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J VASCSURG 1986;3:104-14. 6. Campbell CD, Brooks DH, Webster MW, et al. The use of expanded microporous polytetrafluoroethylene for limb salvage. A preliminary report. Surgery 1976;79:485-91. 7. Quinones-Baldrich WJ, Martin-Paredero V, Baker JD, et al. Polytetrafluoroethylene grafts as the first-choice arterial substitute in femoropopliteal revascularization. Arch Surg 1984; 119:1238-47. 8. Hallett JW, Brewster DC, Darling RC. The limitations of polytetrafluoroethylene in the reconstruction of femoropopliteal and tibial arteries. Surg Gynecol Obstet 1981; 152:81921. 9. O'Donnell TF, Farber SP, Richmand DM, et al. Above-knee polytetrafluoroethylene femoropopliteal bypass graft: is it a reasonable alternative to the below-knee reversed autogenous vein graft? Surgery 1983;94:26-36. 10. Kennev DA, Sauvage LR, Wood SJ, et al. Comparison of noncrimped, externally supported (EXS) and crimped, nonsupported Dacron prostheses for axillofemoral and aboveknee femoropopliteal bypass. Surgery 1982;92:931-46. 11. Ruhland D, Cramer B, Arndt M, et al. Arterial rcconstruction with new polytetrafluoroethylene (PTFE) spiral-wrap prosthesis. Angiology 1980; 1:57-66. 12. Geiger G, Hoevels J, Storz L, et al. Vascular grafts ill belowknee femoropopliteal bypass. A comparative studv. J Cardiovasc Surg 1984;25:523-9. 13. Takenaka K, Kholoussy AM, Yang Y, et al. The influence of supporting rings on the healing of the ringed expanded polytetrafluoroethylene vascular graft. Vase Surg 1985;19: 323-8. 14. Hurwitz RL, Johnson JM, Hufnagel CE. Femoropopliteal bypass using externally supported polytetrafluoroethylenc grafts. Early results in a multiinstitutional study. Am J Surg 1985;150:574-6. 15. Dunn MM, Robinette DR, Peoples lB. Comparison between externally stented and unstented PTFE vascular grafts. Am Surg 1988;54:324-5. 16. Tuchmann A, Dinstl K. Below-knee femoropopliteal bypass using externally supported polytetrafluoroethylenc (PTFE) grafts. Ann Vase Surg 1989;3:177-80. 17. Veith FJ, Gupta SK, Daly V. Femoropopliteal bypass to the isolated popliteal segment: Is pol.vtetrafluoroethylene graft acceptable? Surgery 1981;89:296-303. 18. Veith FJ, Weiser RK, Gupta SK, et al. Diagnosis and management of failing lower extremity arterial reconstructions prior to graft occlusion. J Cardiovasc Surg 1984;25:381-4. 19. Veith FJ, Gupta SK, Ascer E, Rivers SP, Wengerter KR. Improved strategies for secondary operations on infrainguinal arteries. Ann Vase Surg 1990;4:85-93. 20. Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized trials requiring prolonged observations of each patient. II. Analysis and examples. Br J Cancer 1977;35:139. 21. Rutherford RB, Flanigan DP, Gupta SK, ct al. Suggested standards for reports dealing with lower extremity ischemia. J VASC SURG 1986;4:80-94. 22. Haimov H, Giron F, Jacobson JH. The expanded polytetra
172
Gupta et al.
Journal of VASCULAR SURGERY
fluoroethylene graft - three years' experience with 362 grafts. Arch Surg 1979;114:673-7. 23. Burnham SJ, Flanigan DP, Goodreau JJ, et al. Nonvein bypass in below-knee reoperation for lower limb ischemia. Surge~ 1978;84:417-24. 24. Kempczinski RF. Physical characteristics of implanted polytetrafluoroethylene grafts. A prelimina~ report. Arch Surg 1979; 114:917-9.
25. Hiatt JG, Raviola C, Baker JD, ct al. The limitations of predictabili~ of success of femoro-popliteal bypass grafts. J VAse SURG 1984;1:617-22. 26. Harris EJ, Taylor LM, McConneU DB, Moneta GL, Yeager RA, Porter JM. Improved modern results of a~xillobifemoral bypass using externally supported VI'FE. J VAsc SURG 1990;12:416-21.
DISCUSSION
almost never find an above-knee femoropopliteal bypass possible in patients who need bypass surgery for limb salvage. Do you have any idea why this extremely large percentage of patients having above-knee bypasses appeared in your series? The critical problem I have with this paper is one o f statistics. To detect a paten~ improvement with a ringed graft above knee with a power of 0.8, according to the calculations of our biostatisticians, will require randomization of 988 patients. Your randomization of 30 patients in each group gives a power of about 0.05. To detect a patency improvement with a ringed graft below knee with a power of 0.8, would require a total entry of about 350 patients. Your entry of 29 and 34 patients gives a power estimated at 0.15. Thus your likelihood of a type II error in this stud),, or failure to detect a difference when one really does exist, is about 95%. Clearly, many more patients will be required before clear conclusions can be reached. Dr. Gupta. Thank you, Dr. Porter, for your generous comments. I will try to answer vour questions to the best of my ability. You commented on the fact that your results with ringed PTFE grafts for axillofemoral bypass were better than those for the historical controls. However, your study was not performed in a prospective, randomized fashion. If we had not prospectively randomized our patients, it is possible that we might have reached a similar conclusion. This exemplifies our belief that patient randomization is critical to the performance of an accurate comparative study, no matter what kind of graft or bypass is being evaluated. You also raised several points about the statistical analysis used in our study. Our statistician calculated that if only 60 patients were entered in each group, then a 34% patency rate for nonringed grafts and a 50% patency rate for ringed grafts would be required to show a significant difference between these two groups. We will continue to enroll patients in this study, and we hope we will be able to provide more information in the next 2 or 3 years.
Dr. John Porter (Portland, Ore.). We have been interested in using externally supported PTFE grafts primarily for axillobifemoral application. We have recently performed a retrospective review of our axillofemoral experience. Before 1983 we had performed 34 axillobifemoral bypasses using unsupported PTFE. Between 1983 and 1989, we performed 76 axillobifemoral bypasses using supported prostheses. The groups were approximately comparable. The dramatic difference is the difference in patency. The primary patency of these two groups has been radically different in our experience. The first group had the usual primary, patency o f approximately 45% at 3 vears for the unsupported axiUobifemoral grafts. The externally supported group has had the remarkably improved primaq, patency of 85% at 3 years. Based on this experience, we are clearly using many more axillobifemoral bypasses at the present time, especially in the older, more debilitated patients. I have several questions. I am somewhat puzzled bv your indication for patients entry into this study. You state that patients were randomized on the presumption that they were going to need a prosthetic bypass. Indications for randomization included patients who had a history of prior ipsilateral vein removal and duplex ultrasound or saphenous phlebography indicating an inadequate ipsilateral saphenous vein, as well as people whose life expectancy, was judged to be less than 3 years. This was augmented by additional randomization, if at the time of surgery a patient was found to have an inadequate ipsilateral greater saphenous vein. The obvious question is what percentage of your patient population was eligible for randomization based on these rather generous criteria? And, what percentage was not? Your liberal entry criteria clearly permitted significant patient presetection for entry into this study. Only 4 of 122 patients in your series, or 3%, were operated on for claudication. It is interesting to note that you were able to perform an above-knee femoropopliteal bypass in 48% of these patients. In our experience we
Femoropopliteal bypass has become well established as a safe and effective operation for the treatment of severe lower extremity ischemia.13 Although autogenous saphenous vein (ASV) provides the most durable conduit for arterial reconstruction below the inguinal ligament, it may be unavailable or inadequate for femoropopliteal bypass in many patients. 4's Furthermore, in some patients use of prosthetic grafts has been recommended because their life expectancy is limited to 2 to 3 years or to limit the operating time because of the poor medical condition of the patient. ~ Accordingly, several artifiFrom the Division of Vascular Surge~, Montefiore Medical Center/Albert Einstein College of Medicine. Supported in part by The Manning Foundation, The Anna S. Brown Trust, The Renate and Allan B. Hunter Surgical Research Fund, and The New York Institute for Vascular Studies. Presented at the Forty-fourth Annual Meeting of the Society for Vascular Surge~, Los Angeles, Calif., June 4-6, 1990. Reprint requests: Sushil K. Gupta, MD, Division of Vascular Surgery, Montefiore Medical Center, 111 E. 210th St., New York, NY 10467. 24/6/25646
cial arterial conduits have been tried as alternatives to ASV for femoropopliteal bypass, with expanded polytetrafluoroethylene (PTFE) grafts having emerged as the current first choice of most vascular surgeons. Polytetrafluoroethylene grafts were first used clinically for femoropopliteal bypass in 1976. 6 Since then, several studies have shown that PTFE grafts inserted in the femoropopliteal position function well, but that the long-term results of below-knee femoropopliteal bypass are inferior to those achieved with above-knee femoropopliteal bypass. 79 It is thought that this may in part be due to diminished bypass flow brought about by kinking and compression of prosthetic grafts with flexion as they cross the knee joint.~° To obviate this factor as a possible cause of prosthetic femoropopliteal bypass failure, several graft modifications have been developed, including external support of the prostheses with fluorinated ethylene propylene (FEP) in ringed, spiral, and double-helix configurations J °-16 Although we have postulated that below-knee 163
164
Journal of VASCULAR SURGERY
Gupta et al.
Table I. Indications for operation Ringed (58)
Nonringed (64)
Indication
No.
%
No.
%
p value
Claudication Rest pain Ulcer/necrosis
1 16 41
2% 28% 70%
3 11 50
5% 17% 78%
0.8 0.2 0.6
No significant differences occurred in the indications for surgery in patients who received ringed and nonringed PTFE grafts. Values listed are numbers of patients.
prosthetic graft failure is multifactorial, we were hopeful that these rigid grafts would improve the results. However, this hypothesis remained unproven, and hence the present prospective, randomized comparison of ringed and nonringed PTFE grafts was tmdertaken in our patients who were candidates for a prosthetic femoropopliteal bypass. Since we had observed kinking caused by knee flexion in some above-knee grafts as well, these cases were also included in the randomization protocol. Furthermore, the inclusion of above-knee grafts in the study also allowed us to compare the behavior of grafts that extended below the knee to those that terminated above the knee. This article reports the initial results of this study, which began on Jan. 1, 1986. MATERIAL A N D M E T H O D S Patient selection
On the basis of a history of prior removal, duplex ultrasonography, or saphenous venography, all patients without an available ipsilateral ASV long enough to serve as a femoropopliteal bypass, as well as those patients whose life expectancy was judged to be less than 3 years, were randomized to receive either ringed or nonringed PTFE grafts. This group was augmented by patients who were scheduled to receive an ASV graft but who at operation werc found to have an unsuitable ASV because of inadequate length, small size, or fibrotic changes. Randomization techniques were similar to those previously used in a randomized trial comparing ASV and PTFE grafts for infrainguinal arterial reconstructions and consisted of selecting a random card from an unsorted deck of cards marked with the choice of graft material. 4's Patients were selected for aboveknee femoropopliteal bypass when the superficial femoral artery was occluded with reconstitution at the above-knee popliteal artery, unless there coexisted a popliteal artery stenosis greater than 40% behind the knee, in which case below-knee femoropopliteal bypass was performed? The presence of a distal popliteal artery occlusion rendering the patent popliteal artery an isolated segment was not a contraindication
for inclusion in the study, and our criteria for use of an isolated popliteal segment were followed.2'~7 However, patients with extensive necrosis requiring sequential grafts to distal arteries were excluded from randomization. Patients who required a femoropopliteal bypass for reasons other than arteriosclerotic occlusive disease were also excluded from the study. Ninety-one (75%) patients were operated on for tissue necrosis or nonhealing foot ulceration; 27 (22%) for rest pain; and 4 (3%) for claudication. No significant differences were found in the frequency of any of the indications for femoropopliteal bypassing in patients who received ringed versus nonringed PTFE grafts (Table I). The mean age and standard deviation (SD) of the patients was 71 +_ 12 years (range, 27 to 93 years). A high incidence of associated medical illness was found in our patients, and many had significant risk factors such as diabetes, hypertension, smoking, cardiac disease, and previous vascular surgery in the same leg. The patient populations in both groups were evenly matched in regard to these risk factors (Table II). The popliteal outflow tract was quantified as to the number of named patent vessels originating from the popliteal artery. If no named tibial or peroneal artery originated directlv from the popliteal artery, the outflow tract was noted to be isolated. No significant difference was found in the ringed and the nonringed groups in regard to the number of patent runoff vessels (Table II). Efforts to achieve maximal improvement in medical status were instituted before the operation. However, in some patients, the urgent need for revascularization made it necessary to proceed with surgery before optimal medical therapy was completed. 2 Technical considerations
Femoropopliteal bypasses originated from the common femoral artery in 89 (73%) patients and the superficial femoral artery, in 33 (27%) patients. Distal (popliteal) anastomoses were performed above the knee in 59 (48%) patients and below the knee in 63 (52%) patients. Incisions were made in the groin for exposure of the femoral arteries and medially above
Volume 13 Number 1 January 1991
Randomized comparison of ringed and nonringed PTFE grafts
165
Table II. Risk factors and the number of popliteal outflow vessels in patients who underwent femoropopliteal bypass Ringed (58)
Risk factors Diabetes Hypertension Smoking Cardiac disease Previous bypass Popliteal runoff vessels Isolated popliteal arteu 1 vessel 2 vessels 3 vessels
Nonringed (64)
No.
%
No.
%
p value
35 29 32 26 21
60% 50% 55% 45% 36%
37 28 28 26 22
58% 44% 44% 41% 34%
0.9 0.8 0.2 0.4 0.9
18 22 10 8
31% 38% 17% 14%
23 22 15 4
36% 34% 24% 6%
0.8 0.9 0.5 0.2
Note that there was no significant difference in the incidence of associated risk factors or the distribution of outflow vessels in patients who received ringed and nonringed IrI'FE grafts. Values listed are numbers of patients.
or below the knee for exposure of the popliteal artery. A fascial incision along the anterior border of the sartorius muscle was made to expose the above-knee popliteal artery. The medial head of the gastrocnemius muscle was retracted posteriorly to expose the below-knee popliteal artery. A subsartorial tunnel was created by use of blunt dissection with the fingers or a standard graft tunneling instrument passed from the groin to the popliteal fossa. To reach the belowknee popliteal fossa, an anatomic tunnel was made along the popliteal neurovascular bundle. Meticulous technique and operative adjuncts such as headlight illumination and magnification were used for all bypasses. Grafts of 6 mm PTFE (W. L. Gore & Associates, Inc, Elkton, Md.) were used for all procedures performed in this study. The anastomoses were performed with 6.0 polypropylene or 7.0 IH'FE sutures. Systemic heparin in an initial dose of 100 mg/kg was used and was not reversed in most cases. Aspirin and dipyridamole (Persantine) were started 2 to 3 days before surgery and continued indefinitely if tolerated and the medical condition allowed their use. 2 Special attention was given to fashioning the anastomosis of the PTFE grafts so that they coursed without twisting, redundancy, or angnlation with the knee flexed or extended. This required that the inelastic PTFE grafts be cut to precisely the correct length. All anastomoses were meticulously performed taking care to include all layers of the arterial wall in the anastomosis and to avoid any exposed media or adventitial tags within the anastomotic lumen. 2 In patients who received ringed PTFE grafts, only the most proximal and distal rings were removed from the graft, allowing for external reinforcement
of the prosthesis to extend to within 2 cm of the anastomoses. Patient follow-up Patients were examined every 2 to 4 weeks for the first 3 months after operation and then every 3 months thereafter. Graft patency was confirmed by unequivocal pulse examination, ankle Doppler determined pressure measurements, pulse volume recordings, and duplex examination. Contrast arteriography was performed in patients who had a change in pulse examination, a significant decrease in any o f the noninvasive parameters, or the recurrence ofischemic symptoms. Grafts determined to be in the "failing" state (i.e., to be patent but associated with hemodynamically significant lesions that might cause graft thrombosis) were treated with transluminal angioplasty, local revision, or short extensions proximal or distal to the graft? s'19 However, these grafts were considered failed for the purposes of primary pateneT life-table calculations.
Data analysis Three-year graft pateneT rates were tabulated by the cumulative life-table method. Patency intervals were terminated at the last known follow-up period or when patients died or were irretrievably lost to follow-up. Patients requiring major amputation despite a patent femoropopliteal bypass had their pateneT interval terminated at that rime. Primary pateneT intervals were terminated at the first graft thrombosis or at any reintervention for a failing graft. Results of these reinterventions and the effect of attempted thrombectomy or revisions o f failed grafts are shown by secondary patency calculations. 2,2° For
166
Journal of VASCULAR SURGERY
Gupta et M.
100
•
58
36
80
"''''''''! 60
..........
19 17
1,
40
20
0 0
I
I
I
I
I
6
12
18
24
30
86
MONTHS AFTER OPERATION FP W I T H RINGED PTFE
.....
FP
WITH
RINGED Figure
NONPTF'E
1
Fig. 1. Cumulative 3-year primary, graft patency of ringed and nonringed FTFE femoropopliteal bypass grafts (FP). No significant difference was found between the two groups (p = 0.5).
T a b l e I I I . Cumulative 3-year graft patency o f all ringed and n o n r i n g e d P T F E femoropopliteal bypass grafts (p = 0.5) Interval
Ringed 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo Nonringed 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
A t risk
Failed
Withdrawn
Interval patency
Cumulative patency
Standard error
58 49 36 27 20 17 15
3 2 3 1 1 0 0
6 11 6 6 2 2 2
95 95 91 96 95 100 100
95 90 82 79 74 74 74
2.89 4.03 5.79 6.99 8.41 9.12 9.71
64 55 42 33 24 19 17
3 4 1 2
6 9 8 7
95 92 97 93
95 88 85 79
2.63 4.1 5.05 6.26
2
3
91
72
7.78
1 0
1 3
95 100
68 68
8.82 9.32
calculating limb salvage rates, the end point was a major amputation below or above the knee regardless o f the status o f the patency o f the graft. T h e significance o f the differences in graft patency rates in patients w h o received ringed versus n o n r i n g e d P T F E grafts and above-knee versus below-knee femoropopliteal bypasses was determined by the log rank test, and the standard error for each interval was calculated by means o f the m e t h o d o f Peto et al.20 Statistical significance was assumed at the 95% level
(p < 0.05). T h e chi-square m e t h o d was used to test the significance o f differences in the indications for surgery, the patient risk factors, and the popliteal o u t f l o w vessels. RESULTS Primary patency W h e n above-knee and below-knee femoropopliteal bypasses were considered together, n o significant difference was f o u n d in the 3-year primary patency
Volume 13 Number 1 January 1991
Randomized comparison of ringed and nonringed PTFE grafls
167
19
100 .
1
;v~
9
80
"'**
8
,.,
" t
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
12 11
1(
40
20
0
I
0
6
I
I
I
12
18
24
I
36
30
MONTHS AFTER OPERA T/ON RINGED
AK
.....
NONRINGED
AK
Fig. 2. Cumulative 3-year primary graft paten~ of above-knee (AK) ringed and nonringed PTFE femoropopliteal bypass grafts. No significant difference was found between the two groups (p = 0.5). Table IV. Cumulative 3-year graft patency of above-knee (AK) ringed and nonringed PTFE femoropopliteal bypass grafts (p = 0.5) Interval Ringed above-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo Nonringed above-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
At risk
Failed
Withdrawn
Interval patency
Cumulative patency
Standard error
29 25 19 14 10 9 9
1 1 2 1 1 0 0
3 5 3 4 1 0 1
96 96 89 100 100 100 100
96 92 82 82 82 82 82
3.41 5.18 8.03 9.36 10.07 10.67 10.67
30 26 23 20 16 12 11
1 1 0 1 2 0 0
3 2 3 3 2 1 1
96 96 100 94 87 100 100
96 93 93 87 74 74 74
3.3 4.93 5.24 6.89 9.31 10.25 10.75
rates of ringed and nonringed PTFE grafts (74% vs 68%,p -- 0.5) (Table III; Fig. 1). Similarly, no significant differences were found in the 3-year primary graft patcncy rates of ringed and nonringed PTFE above-knee bypasses (82% vs 74%,p = 0.5) (Table IV; Fig. 2) or below-knee PTFE femoropopliteal bypasses (68% vs 59%, p = 0.5) (Table V; Fig. 3). The 3-year primary graft patency rate of all aboveknee PTFE femoropopliteal bypasses was superior to that of below-knee PTFE femoropopliteal bypasses (76% vs 62%), but this difference was not statistically significant (p = 0.25) (Table VI; Fig. 4).
Secondary patency Improvement occurred in the secondary 3-year patency rates of ringed PTFE femoropopliteal grafts from 72% to 88% (p = 0.25) as a result of secondary procedures performed to restore patency or correct defects causing graft thrombosis (failed grafts) or hemodynamic failure (failing grafts). In contrast, the secondary patency rates of nonringed PTFE grafts were improved only slightly (from 68% to 71%, p = 0.9) after these secondary procedures. Four grafts failed in a total of 10 bypasses requiting reintervention in the tinged group and only one graft
168
Journal of VASCULAR SURGERY
Gupta et al.
100
I~" Z'~ r.
17
13 10
= ~
1T
8
6
5
80
I
"
60
!"1
......... 6
4
40
20
0
i 0
i
6
RINGED
I
I
i
12 18 24 MONTHS A F T E R OPERATION .....
BK
NONRINGED
30
36
BK
Fig. 3. Cumulative 3-year primary graft paten~ of below-knee (BK) ringed and nonringed PTFE femoropopliteal bypass grafts. No significant difference was found between the two groups (p = 0.5). Table V. Cumulative 3-year graft patency of below-knee (BK) ringed and nonringed PTFE femoropopliteal bypass grafts (p = 0.5) Interval Ringed below-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 m o 18-24 mo 24-30 mo 30-36 mo Nonringed below-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
A t risk
Failed
Withdrawn
Interval patency
Cumulative patency
Standard error
29 24 17 13 10 8 6
2 1 1 1 1 0 0
3 6 3 2 1 2 1
93 95 94 92 89 100 100
93 88 83 76 68 68 68
4.6 6.1 8.3 9.8 10.1 12.1 12.4
34 29 19 13 8 7 6
2 3 1 1 0 1 0
3 7 5 4 1 0 2
94 88 94 91 100 85 100
94 82 77 70 70 59 59
3.9 6.3 8.4 9.8 10.6 11.9 13.1
failed of 13 requiring reintervention in the nonringed group. Limb salvage
Life-table 3-year limb salvage rates were 92% for patients with ringed PTFE grafts and 79% for those with nonringed grafts (p = 0.25) (Table VII). The 3-year limb salvage rates of above-knee and belowknee ringed grafts were 93% and 91%, respectively. The 3-year limb salvage rates of above-knee and below-knee nonringed grafts were 83% and 73%,
respectively (p = 0.5). No statistically significant difference was found between ringed abovc-knce and nonringed above-knee grafts (p = 0.5) and/or between ringed below-knee and nonringed below-knee femoropopliteal bypasses (p = 0.1). However, the number of patients available for comparison in these subgroups was small. In each group of patients who received ringed and nonringed grafts, one postoperative death occurred (within 30 days), resulting in an operative mortality rate of less than 2%.
Volume 13 Number 1 January. 1991
Randomized comparison of ringed and nonringed PTFE grafts
100
34
169
26 21
20
1
f
80 X.
"i! ......... ......... l- ........
60
15
_L
I
[
40
20
0
I
0
I
6
ALL
I
12 18 24 MONTHS AFTER OPERA T/ON AK
PTFE
.....
FEM POPS
ALL
BK
30
36
PTFE
FEM POPS
Fig. 4. Cumulative 3-year primaqr graft patcn~ of all above-knee (AK) and all below-knee (BK) PTFE femoropopliteal bypass grafts (FEM POPS). The 3-year patency rate of aboveknee PTFE femoropopliteal bypass grafts was marginally better than that of below-knee PTFE femoropoplitea! bypass grafts (p = 0.25).
Table VI. Cumulative 3-year graft patency of all above-knee and all below-knee PTFE femoropopliteal bypass grafts (p = 0.25) Interval Above-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo Below-knee grafts 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
At risk
Failed
Withdrawn
Interval patency
Cumulative patency
Standard error
59 51 42 34 26 21 20
2 2 2 1 2 0 0
6 7 6 7 3 1 2
96 96 95 97 92 100 100
96 92 88 85 76 76 76
2.3 3.5 4.7 5.6 7.1 7.9 8.1
63 53 36 26 18 15 12
4 4 2 2 1 1 0
6 13 8 6 2 2 3
93 91 93 91 94 93 100
93 85 80 73 68 62 62
3.1 4.5 5.6 7.4 9.1 9.9 10.3
DISCUSSION
Although ASV graft patency is superior to PTFE graft patency after 2 years, a substantial proportion of patients undergoing femoropopliteal bypass will require a prosthetic graft because the ipsilateral vein is absent or inadequate or because the patient has limited life expectancy, s Moreover, some surgeons use PTFE grafts preferentially for femoropopliteal bypasses for a variety of reasons. 7'22
Kinking and compression of prosthetic grafts with knee flexion has been postulated to be one cause of graft thrombosis and a major reason that belowknee prosthetic grafts have poorer long-term patency then those inserted above the knee. Externally supported Dacron grafts were first introduced by Kenney et al. 1° as a method of preventing kinking and compression. Since then the concept of external support to reduce kinking and compression of PTFE
170
Journal of VASCULAR SURGERY
Gupta et al.
Table VII. Cumulative 3-year limb salvage rates for ringed and nonringed PTFE femoropopliteal bypass grafts (p = 0.25) Interval Ringed 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo Nonringed 0-1 mo 1-6 mo 6-12 mo 12-18 mo 18-24 mo 24-30 mo 30-36 mo
A t risk
Amputations
Withdrawn
Interval paten~
Cumulative paten~
Standard error
58 49 36 26 19 17 15
1 1 1 0 0 0 0
8 12 9 7 2 2 2
98 98 97 100 100 100 100
98 96 92 92 92 92 92
1.7 2.7 4.1 4.8 5.7 6.1 6.4
64 55 42 33 24 19 17
3 5 0 0 1 0 0
6 8 9 9 4 2 3
95 90 100 100 95 100 100
95 86 86 86 79 79 79
2,6 4.3 4.9 5.6 7.1 8,1 8,5
grafts has been introduced and is used widely, particularly when grafts are placed across areas of flexion or bony prominences. Several clinical reports suggesting better PTFE graft paten~ rates with the use of externally supported grafts have appeared in the literature, l°'11.~s'~s However, these reports are inconclusive for several reasons including too few patients, l° limited foUow-up, ~ or no control groups.13'15 Data from the present randomized, prospective study fail to document a clear advantage of externally supported ringed PTFE grafts as compared to unsupported, nonringed PTFE grafts when used for femoropopliteal bypass. Our findings are supported by two experimental canine studies, which compared ringed and nonringed PTFE grafts and failed to document increased graft patency rates of ringed grafts.~2"~4Moreover, Geiger et a1.12 showed by postoperative arteriography that graft kinking and constriction could still occur in PTFE grafts that were externally supported with FEP in double-helix and ring configurations. External support of PTFE grafts with rings or other configurations, does not necessarily prevent graft kinking, compression, and reduced flow. Moreover, several other studies failed to show a decrease in bypass flow when below-knee femoropopliteal bypass grafts constructed with conventional PTFE prostheses were subjected to high degrees of knee flexion.23a4 It is interesting to note that the patency rates of above-knee grafts (76%) were consistently superior to all those terminating below the knee (62%) (Table VI, Fig. 4). However,
this difference was not statistically significant (p = 0.25) because of the fewer numbers of patients in each group and the relatively short duration of follow-up. It is difficult to speculate why above-knee PTFE grafts typically perform better than belowknee PTFE grafts. Since the paten~ rate of ringed above-knee grafts (84%) (Table IV) was superior to ringed below-knee grafts (68%) (Table V), avoidance of kinking and compression by adding rings does not appear to make a significant difference. We and others have always believed that fcmoropopliteal bypass graft failure is caused by multiple factors, and that it cannot be prevented by simply obviating a single factor, 5"9"17"19"25Because of this and because there are disadvantages to using ringed grafts, we thought it essential to determine how important external support was before using it routinely. Although no clear advantage can yet be demonstrated for externally supported ringed grafts, our data may still be too prelimina~ to show it. We will therefore continue our study with further enrollment of patients and additional follow-up ofthosc already enrolled. However, the results to date suggest that PTFE grafts should not be routinely externally supported. Moreover, the potential advantages of ringed grafts used in other positions should be clearly shown by adequate controlled studies before they are adopted for routine use. This opinion is reinforccd by the fact that ringed grafts have intrinsic disadvantages. They are technically more difficult to inscrt because they are less malleable. This property can actually increase their tendency to kink at the junction
Volume 13 Number 1 January 1991
Randomized comparison of ringed and nonringed PTFE grafts 171
between the ringed and nonringed portion of the graft. This mandates that the surgeon carry the ringed portion of the graft to within 1 to 2 cm of the anastomosis if the latter is in proximity to a joint. Another disadvantage to ringed grafts is that patency is more difficult to document by graft pulse palpation. This may make it necessary to use noninvasive testing more frequently to detect failing grafts before they thrombose. Such detection has been shown to be important in achieving the ultimate objective of limb salvage in many of these patients. 18,~9The difficulty of palpating pulses in ringed PTFE grafts also makes these grafts more difficult to find and dissect out if an extension or a reoperation on a failing graft is required. Despite these considerations the possibility that ringed grafts offer major advantages when used for other ~pes of bypasses or minor advantages when they are used for certain femoropopliteal bypasses demands further study. It is certainly possible that the externally supported prosthetic grafts may be advantageous when used for axillofemoral or other extraanatomic bypasses in which a bypass may be subjected to external compression against an unyielding chest wall or bony structure. 26 In summary the present study comparing ringed and nonringed PTFE grafts for femoropopliteal bypass failed to document a significant advantage to the former in terms of 3-year graft patency rates in either above-knee or below-knee positions. At the present time, ringed PTFE grafts cannot be preferentially recommended over conventional PTFE grafts for routine use in patients who require a prosthetic femoropopliteal bypass. Further controlled trials comparing ringed and nonringed PTFE grafts in patients who require other arterial reconstructions with synthetic grafts are indicated to clarify the real advantages, if any, of these externally supported prostheses.
REFERENCES 1. Maini BS, Mannick JA. Effect of arterial reconstruction on limb salvage: a ten-year appraisal. Arch Surg 1978; 113:1297304. 2. Veith FJ, Gupta SK, Samson RH, et al. Progress in limb salvage by reconstructive arterial surgery combined with new or improved adjunctive procedures. Ann Surg 1981; 194:386401. 3. Veith FJ, Gupta SK, Wengerter KR, et al. Changing arteriosclerotic disease patterns and management strategies in lower limb threatening ischemia. Ann Surg 1990;212:402-14. 4. Bergan JJ, Veith FJ, Bernhard VM, et al. Randomization
of autogenous vein and polytetrafluoroethylene grafts in femoral-distal reconstruction. Surgery, 1982;92:921-30. 5. Veith FJ, Gupta SK, Ascer E, et al. Six-year prospective multicenter randomized comparison of autologous saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J VASCSURG 1986;3:104-14. 6. Campbell CD, Brooks DH, Webster MW, et al. The use of expanded microporous polytetrafluoroethylene for limb salvage. A preliminary report. Surgery 1976;79:485-91. 7. Quinones-Baldrich WJ, Martin-Paredero V, Baker JD, et al. Polytetrafluoroethylene grafts as the first-choice arterial substitute in femoropopliteal revascularization. Arch Surg 1984; 119:1238-47. 8. Hallett JW, Brewster DC, Darling RC. The limitations of polytetrafluoroethylene in the reconstruction of femoropopliteal and tibial arteries. Surg Gynecol Obstet 1981; 152:81921. 9. O'Donnell TF, Farber SP, Richmand DM, et al. Above-knee polytetrafluoroethylene femoropopliteal bypass graft: is it a reasonable alternative to the below-knee reversed autogenous vein graft? Surgery 1983;94:26-36. 10. Kennev DA, Sauvage LR, Wood SJ, et al. Comparison of noncrimped, externally supported (EXS) and crimped, nonsupported Dacron prostheses for axillofemoral and aboveknee femoropopliteal bypass. Surgery 1982;92:931-46. 11. Ruhland D, Cramer B, Arndt M, et al. Arterial rcconstruction with new polytetrafluoroethylene (PTFE) spiral-wrap prosthesis. Angiology 1980; 1:57-66. 12. Geiger G, Hoevels J, Storz L, et al. Vascular grafts ill belowknee femoropopliteal bypass. A comparative studv. J Cardiovasc Surg 1984;25:523-9. 13. Takenaka K, Kholoussy AM, Yang Y, et al. The influence of supporting rings on the healing of the ringed expanded polytetrafluoroethylene vascular graft. Vase Surg 1985;19: 323-8. 14. Hurwitz RL, Johnson JM, Hufnagel CE. Femoropopliteal bypass using externally supported polytetrafluoroethylenc grafts. Early results in a multiinstitutional study. Am J Surg 1985;150:574-6. 15. Dunn MM, Robinette DR, Peoples lB. Comparison between externally stented and unstented PTFE vascular grafts. Am Surg 1988;54:324-5. 16. Tuchmann A, Dinstl K. Below-knee femoropopliteal bypass using externally supported polytetrafluoroethylenc (PTFE) grafts. Ann Vase Surg 1989;3:177-80. 17. Veith FJ, Gupta SK, Daly V. Femoropopliteal bypass to the isolated popliteal segment: Is pol.vtetrafluoroethylene graft acceptable? Surgery 1981;89:296-303. 18. Veith FJ, Weiser RK, Gupta SK, et al. Diagnosis and management of failing lower extremity arterial reconstructions prior to graft occlusion. J Cardiovasc Surg 1984;25:381-4. 19. Veith FJ, Gupta SK, Ascer E, Rivers SP, Wengerter KR. Improved strategies for secondary operations on infrainguinal arteries. Ann Vase Surg 1990;4:85-93. 20. Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized trials requiring prolonged observations of each patient. II. Analysis and examples. Br J Cancer 1977;35:139. 21. Rutherford RB, Flanigan DP, Gupta SK, ct al. Suggested standards for reports dealing with lower extremity ischemia. J VASC SURG 1986;4:80-94. 22. Haimov H, Giron F, Jacobson JH. The expanded polytetra
172
Gupta et al.
Journal of VASCULAR SURGERY
fluoroethylene graft - three years' experience with 362 grafts. Arch Surg 1979;114:673-7. 23. Burnham SJ, Flanigan DP, Goodreau JJ, et al. Nonvein bypass in below-knee reoperation for lower limb ischemia. Surge~ 1978;84:417-24. 24. Kempczinski RF. Physical characteristics of implanted polytetrafluoroethylene grafts. A prelimina~ report. Arch Surg 1979; 114:917-9.
25. Hiatt JG, Raviola C, Baker JD, ct al. The limitations of predictabili~ of success of femoro-popliteal bypass grafts. J VAse SURG 1984;1:617-22. 26. Harris EJ, Taylor LM, McConneU DB, Moneta GL, Yeager RA, Porter JM. Improved modern results of a~xillobifemoral bypass using externally supported VI'FE. J VAsc SURG 1990;12:416-21.
DISCUSSION
almost never find an above-knee femoropopliteal bypass possible in patients who need bypass surgery for limb salvage. Do you have any idea why this extremely large percentage of patients having above-knee bypasses appeared in your series? The critical problem I have with this paper is one o f statistics. To detect a paten~ improvement with a ringed graft above knee with a power of 0.8, according to the calculations of our biostatisticians, will require randomization of 988 patients. Your randomization of 30 patients in each group gives a power of about 0.05. To detect a patency improvement with a ringed graft below knee with a power of 0.8, would require a total entry of about 350 patients. Your entry of 29 and 34 patients gives a power estimated at 0.15. Thus your likelihood of a type II error in this stud),, or failure to detect a difference when one really does exist, is about 95%. Clearly, many more patients will be required before clear conclusions can be reached. Dr. Gupta. Thank you, Dr. Porter, for your generous comments. I will try to answer vour questions to the best of my ability. You commented on the fact that your results with ringed PTFE grafts for axillofemoral bypass were better than those for the historical controls. However, your study was not performed in a prospective, randomized fashion. If we had not prospectively randomized our patients, it is possible that we might have reached a similar conclusion. This exemplifies our belief that patient randomization is critical to the performance of an accurate comparative study, no matter what kind of graft or bypass is being evaluated. You also raised several points about the statistical analysis used in our study. Our statistician calculated that if only 60 patients were entered in each group, then a 34% patency rate for nonringed grafts and a 50% patency rate for ringed grafts would be required to show a significant difference between these two groups. We will continue to enroll patients in this study, and we hope we will be able to provide more information in the next 2 or 3 years.
Dr. John Porter (Portland, Ore.). We have been interested in using externally supported PTFE grafts primarily for axillobifemoral application. We have recently performed a retrospective review of our axillofemoral experience. Before 1983 we had performed 34 axillobifemoral bypasses using unsupported PTFE. Between 1983 and 1989, we performed 76 axillobifemoral bypasses using supported prostheses. The groups were approximately comparable. The dramatic difference is the difference in patency. The primary patency of these two groups has been radically different in our experience. The first group had the usual primary, patency o f approximately 45% at 3 vears for the unsupported axiUobifemoral grafts. The externally supported group has had the remarkably improved primaq, patency of 85% at 3 years. Based on this experience, we are clearly using many more axillobifemoral bypasses at the present time, especially in the older, more debilitated patients. I have several questions. I am somewhat puzzled bv your indication for patients entry into this study. You state that patients were randomized on the presumption that they were going to need a prosthetic bypass. Indications for randomization included patients who had a history of prior ipsilateral vein removal and duplex ultrasound or saphenous phlebography indicating an inadequate ipsilateral saphenous vein, as well as people whose life expectancy, was judged to be less than 3 years. This was augmented by additional randomization, if at the time of surgery a patient was found to have an inadequate ipsilateral greater saphenous vein. The obvious question is what percentage of your patient population was eligible for randomization based on these rather generous criteria? And, what percentage was not? Your liberal entry criteria clearly permitted significant patient presetection for entry into this study. Only 4 of 122 patients in your series, or 3%, were operated on for claudication. It is interesting to note that you were able to perform an above-knee femoropopliteal bypass in 48% of these patients. In our experience we
