Randomized comparison of percutaneous Viabahn stent grafts vs prosthetic femoral-popliteal bypass in the treatment of superficial femoral arterial occlusive disease

Results 

Between March 2004 and May 2005, 50 limbs in 40 patients were treated percutaneously with the stent graft, and 50 limbs in 46 patients were treated surgically with femoral–AK popliteal artery bypass.

Pretreatment clinical categories of chronic limb ischemia using Rutherford’s classification3 for the treated limbs are shown in Table II. No significant difference in pretreatment clinical grades between the two treatment groups was noted.

Table II. Pretreatment distribution of chronic limb ischemia Rutherford categories3

	Clinical grade	Stent graft limbs (n = 50)	Surgical bypass limbs (n = 50)
	0	0	0
	1	2	1
	2	23	20
	3	16	10
	4	4	10
	5	4	7
	6	1	2

Generalized Fisher’s exact test, P = .3676.

By following the TASC grading system4 for femoropopliteal lesions, each limb in both treatment groups was assigned a TASC classification as summarized in Table III. The TASC classifications between the two treatment groups were not significantly different.

Table III. Lesion TransAtlantic Inter-Society Consensus (TASC) classification per limb

	TASC	Stent graft (n = 50)	Surgical bypass (n = 50)
	A	2	1
	B	6	8
	C	37	31
	D	5	10

Generalized Fisher’s exact test, P = .4739.

Stent graft placement was technically successful in 100% of the treated limbs in the stent graft treatment group. A total of 114 devices were implanted in the 50 limbs, with a mean of 2.3 stent grafts placed per limb. The mean diameter of the stent grafts was 5.7 mm (range, 5 to 7 mm). The mean total length of artery covered with the stent graft was 25.6 ± 15 cm.

After treatment, 37 (93%) of 40 patients in the stent graft group took clopidogrel and aspirin for a minimum of 3 months. One patient claimed an allergy to clopidogrel, and two others refused to take it. These three patients did take aspirin, however.

Femoral–AK popliteal artery bypass was successfully performed in 100% of the treated limbs in the surgical treatment group. Dacron grafts were used in 32 limbs (64%), and ePTFE was used in 18 limbs (36%). The mean diameter of the synthetic bypass grafts was 7.4 mm (range, 7 to 8 mm).

In the surgical bypass group, 24 (52%) of 46 patients were on a clopidogrel and aspirin regimen post-treatment. Seventeen patients were taking aspirin only as recommended by the treating surgeon. The remaining five patients were taking warfarin preoperatively and were continued only on this regimen postoperatively.

Immediate procedure-related and early postoperative, nonthrombotic complications were observed in four (8%) of 50 limbs (40 patients) treated with the stent graft. In one patient, a dissection was created in the SFA during passage of a guidewire. The stent graft was used to exclude the dissection along with the atherosclerotic lesion. Another patient experienced transient mild leg edema in the treated limb. Deep venous thrombosis was ruled out, and the edema resolved. One patient reported severe thigh pain in the treated limb that required readmission to the hospital for pain management. The pain resolved ≤24 hours without any identifiable pathology. One final patient had a small groin hematoma that resolved without intervention.

In the surgical bypass group, early postoperative complications were observed in three (6%) of 50 limbs (46 patients). These three patients developed a groin lymphocele, a groin seroma, and a small superficial groin wound dehiscence, respectively. The patient with a groin lymphocele was returned to the operating room for washout and reclosure of the wound. The other two patients were managed nonoperatively.

An improvement in grade of the Rutherford classification occurred in all limbs (100%) in the stent graft treatment group and in 46 limbs (92%) in the surgical bypass group. The overall mean improvement was 2.4 clinical grades in both groups.

Median follow-up duration was 18 months for both treatment groups. Follow-up was available for 31 (78%) of 40 patients in the stent graft group. Four patients died during the study period from conditions unrelated to infrainguinal disease, and all but one of these patients had tissue loss preoperatively. Five patients were lost to follow-up.

During this period, 13 of the stent grafts failed secondary to thrombosis. An early graft thrombosis occurred in the recovery room the same day of the procedure in one patient. One stent graft thrombosis occurred within the first month after stent graft implantation. The other 11 stent graft thromboses were detected after a mean period of 5.4 ± 3 months after stent graft placement.

Of the 13 grafts that thrombosed, five (38%) were successfully declotted with mechanical balloon thrombectomy and one was successfully recanalized with intra-arterial tissue plasminogen activator–mediated lysis. In six (46%), attempts at thrombectomy or lysis were unsuccessful, and these patients eventually underwent open surgical bypass (five to the AK popliteal artery). Finally, one of the patients with a thrombosed stent graft was found to have heparin-induced thrombocytopenia and amputation eventually was performed owing to progressive tissue loss. This patient had tissue loss preoperatively.

In two of the patients who underwent thrombectomy of their clotted stent grafts, an underlying stenosis was detected after post-thrombectomy angiography. One patient had a high-grade stenosis just distal to the stent graft. The other patient had a proximal stenosis. In both cases, the stenotic area was treated with angioplasty and stent graft placement.

In one patient (two limbs), both lower extremities thrombosed simultaneously. This patient was found to have bilateral lower extremity embolization after new onset atrial fibrillation. Thirty days before occlusion, a routine follow-up arterial duplex demonstrated no stenosis within either limb and the patient had palpable bilateral pedal pulses. One limb was eventually opened with thrombectomy, but the other limb required femoral–below-knee bypass for limb salvage.

Another patient in the stent graft group with a patent graft was found to have a short segment proximal arterial stenosis detected at follow-up color flow duplex sonography imaging, which was confirmed at angiography to be approximately 30% to 40%. This segment was treated with angioplasty alone. Overall, 14 interventions had to be performed in the stent graft treatment group during 12 months.

None of the three patients in the stent graft group that were not taking clopidogrel after treatment had a thrombosed stent graft. Of the 13 thrombosed stent grafts, 11 (84%) were TASC C lesions. Of the other two thrombosed grafts, one (8%) was a TASC A lesion and one (8%) was a TASC D lesion.

Follow-up was available for 37 (80%) of 46 patients in the surgical bypass group. Four patients died owing to conditions unrelated to their infrainguinal disease. Five patients were lost to follow-up.

During this period, 10 incidences of synthetic graft thrombosis occurred, of which seven (70%) were TASC C lesions and three (30%) were TASC D lesions; none were TASC A or B lesions. One occurred within the first month after implantation. The other nine were detected a mean of 6.9 ± 4 months after graft placement. Four of the 10 thrombosed synthetic grafts were successfully declotted with mechanical balloon thrombectomy. Three patients underwent below knee popliteal artery bypass with great saphenous vein conduit after thrombectomy failed.

In three separate instances, ischemia from clotted grafts eventually led to below knee amputation. In two additional cases, progressive tissue loss despite patent grafts led to two additional amputations. All instances of limb amputation occurred in patients that had tissue loss preoperatively.

One patient in the surgical group with a patent graft was found on follow-up duplex ultrasound to have a distal native arterial stenosis. This was confirmed at angiography, and the stenosis was treated with balloon angioplasty and stent placement. As already mentioned, one patient with a groin lymphocele had an operative intervention. Overall, 12 interventions had to be performed in the surgical bypass group during 12 months.

Cumulative primary and secondary patency rates were calculated with use of the life-table method. At follow-up at 3, 6, 9 and 12 months, the respective primary patency rates were 84.0%, 82.0%, 75.6%, and 73.5%, for the stent graft group and 90.0%, 81.8%, 79.7%, and 74.2% for the surgical bypass group (Table IV). Secondary patency at the 12-month follow-up was 83.7% for the stent graft group and 83.9% for the surgical bypass group (Table V). There was no significant difference in primary patency (P = .895) or secondary patency (P = .861) between the two treatment groups. Limb salvage at the 12-month follow-up was not significantly different at 98.0% for the stent graft patients and 89.6% for surgical bypass patients (P = .094) (Table VI). Baseline ABIs were were 0.57 ± 0.19 for the stent graft group and 0.46 ± 0.22 for the surgical bypass group. At 12 months, the mean improvement in ABI was 0.23 for the stent graft treatment group and 0.37 for the surgical bypass group (P = .113).

Table IV. Primary patency in femoral-popliteal bypass group and stent graft group

	Time post-treatment (months)	N at risk at start of interval	N events during interval⁎	N censored during interval⁎	% Free from loss of patency	95% CI
	Operative (day 0-30)
	Fem-pop bypass	50	2(2)	0(0)	0.960	(0.849,0.990)
	Stent graft	50	2(2)	0(0)	0.960	(0.849,0.990)
	3 months (day 31-136)
	Fen-pop bypass	48	3(5)	1(1)	0.900	(0.776,0.957)
	Stent graft	48	6(8)	1(1)	0.840	(0.705,0.917)
	6 months (day 137-227)
	Fem-pop bypass	44	4(9)	2(3)	0.818	(0.680,0.901)
	Stent graft	41	1(9)	1(2)	0.820	(0.682,0.902)
	9 months (day 228-319)
	Fem-pop bypass	38	1(10)	8(11)	0.797	(0.655,0.885)
	Stent graft	39	3(12)	1(3)	0.756	(0.611,0.854)
	12 months (day 320-456)
	Fem-pop bypass	29	2(12)	8(19)	0.742	(0.587,0.846)
	Stent graft	35	1(13)	7(10)	0.735	(0.587,0.837)

CI, Confidence interval. Log-rank P = .895.

⁎ Number in parenthesis represents cumulative events or censored observations through end of interval.

Table V. Secondary patency in femoropopliteal bypass group and stent graft group

	Time post-treatment (months)	N at risk at start of interval	N events during interval⁎	N censored during interval⁎	% Free from loss of patency	95% CI
	Operative (day 0-30)
	Fem-pop bypass	50	2(2)	0(0)	0.960	(0.849,0.990)
	Stent graft	50	0(0)	0(0)	1.000	(1.000,1.000)
	3 months (day 31-136)
	Fem-pop bypass	48	2(4)	1(1)	0.920	(0.801,0.969)
	Stent graft	50	6(6)	1(1)	0.880	(0.752,0.944)
	6 months (day 137-227)
	Fem-pop bypass	45	3(7)	2(3)	0.859	(0.726,0.930)
	Stent graft	43	1(7)	1(2)	0.860	(0.728,0.930)
	9 months (day 228-319)
	Fem-pop bypass	40	1(8)	8(11)	0.837	(0.701,0.915)
	Stent graft	41	1(8)	1(3)	0.839	(0.703,0.916)
	12 months (day 320-456)
	Fem-pop bypass	31	0(8)	11(22)	0.837	(0.701,0.915)
	Stent graft	39	0(8)	9(12)	0.839	(0.703,0.916)

CI, Confidence interval. Log-rank P value = .861.

⁎ Number in parenthesis represents cumulative events or censored observations through end of interval.

Table VI. Limb salvage in femoropopliteal bypass group and stent graft group

	Time post-treatment (months)	N at risk at start of interval	N events during interval⁎	N censored during interval⁎	% Free from loss of patency	95% CI
	Group: femoropopliteal bypass
	Operative (day 0-30)
	Fem-pop bypass	50	2(2)	0(0)	0.960	(0.849,0.990)
	Stent graft	50	0(0)	0(0)	1.000	(1.000,1.000)
	3 months (day 31-136)
	Fem-pop bypass	48	1(3)	2(2)	0.940	(0.825,0.980)
	Stent graft	50	1(1)	4(4)	0.980	(0.866,0.997)
	6 months (day 137-227)
	Fem-pop bypass	45	1(4)	4(6)	0.919	(0.799,0.969)
	Stent graft	45	0(1)	1(5)	0.980	(0.866,0.997)
	9 months (day 228-319)
	Fem-pop bypass	40	1(5)	8(14)	0.896	(0.768,0.956)
	Stent graft	44	0(1)	2(7)	0.980	(0.866,0.997)
	12 months (day 320-456)
	Fem-pop bypass	31	0(5)	11(25)	0.896	(0.768,0.956)
	Stent graft	42	0(1)	11(18)	0.980	(0.866,0.997)

CI, Confidence interval. Log-rank P = .094.

⁎ Number in parenthesis represents cumulative events or censored observations through end of interval.

Length of hospital stay was analyzed for both groups. The mean hospital stay was 0.9 ± 0.8 days for the stent graft group was and 3.1 ± 1.8 days for the surgical group. This difference proved to be significant (t test, P < .001).

Discussion 

Initial attempts to treat SFA atherosclerotic disease percutaneously were made with transluminal angioplasty (PTA).5 Although effective in short-segment stenosis, PTA has been disappointing as a primary treatment for longer SFA stenosis.6 Further investigation led to the use of angioplasty, followed by stent placement. The results of these studies were equivocal to PTA patency,7 most likely due to neointimal hyperplasia.

The patency of percutaneously placed stent grafts for SFA disease was published by Lammer et al in 2000.1 This multicenter study demonstrated 1-year primary and secondary patency rates of 79% and 93% in 80 limbs. Bauermeister et al8 treated 35 patients and achieved primary and secondary patency rates of 73.2% and 82.6% at 12 months. In 2003, Jahnke et al9 reported primary assisted patency rates at 12 and 24 months of 82.4% and 80.3%, and secondary patency was 88.3% and 83.2%.

Our study is the first, to our knowledge, to prospectively randomize patients to percutaneous and surgical arms in treatment of disease in the SFA and then directly compare the outcomes. The prior 11 studies of stent grafts have relied on previously published reports of AK bypass patency and historical data for comparison. Historical controls are useful, but prospective, randomized trials allow direct comparison with the most current therapy.

Our technical success rate of 100% mirrors that of many other published trials. Complications were seen in 8% of the patient limbs in the percutaneously treated group. One patient had to be readmitted to the hospital for severe thigh pain. Saxon et al10 noted the phenomenon of thigh pain after stent graft placement. The pain is thought to be secondary to overdilation of the SFA and may be avoided by not oversizing more than 10% of the native diameter.10 Minor complications occurred in two other patients, and the symptoms of both patients resolved with observation alone.

Arterial occlusions in the stent graft group were seen within all TASC classifications. Interestingly, two of these cases showed there were underlying lesions at the proximal and distal end of the stent graft that became evident after removal of the thrombus. They were successfully treated by angioplasty and stenting. Deutschman et al11 described a similar phenomenon in their series of 17 patients. The increased rate of hyperplasia at the proximal or distal ends of the graft may be due to overdilation with angioplasty and failure to cover the treatment area sufficiently with the covered stent.8

Most of the 13 thrombosed stent grafts (84%) were TASC C lesions. This finding is not surprising and supports the finding of others that the higher the TASC classification the lower the patency. Only one of our TASC D lesions treated thrombosed, but with only five in the group, definitive conclusions cannot reliably be drawn.

All patients treated by endovascular means required only outpatient care and in most instances were discharged on the same day, offering various cost savings in their care and earlier return to normal activities. This is supported by our length-of-stay analysis that shows a decrease of required hospital stay of nearly two thirds. The actual cost analysis was not performed in the current study, but follow-up analysis is planned. All but one of the patients in the endovascular group returned to work and normal daily activities ≤48 hours, and most were able to return ≤24 hours.

Conclusion 

The choice of open surgical bypass vs percutaneously placed covered stents in the SFA has yet to be fully delineated. Nevertheless, our study shows no significant difference in synthetic bypass material placed in an open vs an endovascular fashion. We believe the results demonstrated in this study represent a significant step in defining endovascular treatment of disease in the SFA. The stent graft appears to be a viable option for primary SFA revascularization, particularly when vein is not available or if the patient is a poor candidate for conventional bypass. Correct graft sizing, avoiding overdilation, and not performing angioplasty beyond the area of planned stent coverage seem to be critically important technical facets of successful placement. We plan to continue to follow our cohort to assess longer-term outcomes.

Author contributions