- Email: [email protected]
Initial Experience With the Palmaz Stent for Aortoiliac Stenoses
Initial Experience With the Palmaz Stent for Aortoiliac Stenoses Yehuda G. Wolf, MD, Richard A. Schatz, MD, Harry J. Knowles, MD, Mohsin Saeed, MD, Eugene F. Bernstein, MD, PhD, and Ralph B. Dilley, MD, La Jolla, California
The initial 37 consecutive patients to be treated at our institution with the Palmaz stent placed in the aortoiliac arteries were retrospectively reviewed. In these patients, 50 stenoses and six occlusions were treated with 128 stents. Nine patients with combined iliac and common femoral obstruction underwent common femoral endarterectomy and profundaplasty with intraoperative iliac artery angioplasty and stent application. Stenoses were reduced from 57 - 17% to 1 _ 5% (p <0.01), and peak systolic pressure gradients across the lesions were reduced from 45 _ 30 mm Hg to 1.3 _+ 3.4 mm Hg (p <0.01). Symptoms resolved in 27 patients and improved in eight patients. One patient died and four patients were treated nonoperatively for complications. During a mean follow-up of 12 months (6 to 21 months), six patients had recurrence of symptoms (16%) and four patients died of other diseases. Routine arteriograms after 6 months in 19 patients demonstrated recurrent mild to moderate stenoses (9% to 43%) in six patients (32%), but only two were symptomatic (11%). Secondary procedures included reexpansion of aortic and iliac stents in two patients and aortofemoral bypass in two patients. Early results suggest the efficacy of the Palmaz stent in the management of aortoiliac stenoses and its use intraoperatively in conjunction with surgical correction of outflow. Close follow-up of these patients by multidisciplinary groups is warranted. (Ann Vasc Surg 1993;7:254-261.)
The first report of an experimental intravascular metal stent appeared in 1969.1 Several designs were investigated experimentally for both arterial stenosis and dissection. 2-~ The first clinical report of self-expanding intravascular stents in the coronary and iliofemoral arteries appeared in 1987. 6 The Palmaz stent, which was the first balloonexpandable stent, was initially developed as a woven-wire device~ but was subsequently simplified into a slotted stainless steel tube. It is one of several intravascular stents currently in clinical use s'9 and recently was approved by the Food and From the Divisions of Vascular Surgery, Cardiology (R,A.S.), and Radiology (H.J.K. and M.S.), Scripps Clinic and Research Foundation, La Jolla, California. Reprint requests: Ralph B. Dilley, MD, Division of Vascular Surgery, Scripps Clinic and Research Foundation, 10666 N. Torrey Pines Rd., La Jolla, CA 92037.
254
Drug Administration for use in the iliac arteries. We have used it at the Scripps Clinic and Research Foundation since August 1990. This review was undertaken to examine our early experience with this device and to ascertain its role in the m a n a g e m e n t of aortoiliac stenosis. PATIENTS AND METHODS This retrospective review includes the initial 37 consecutive patients with aortoiliac stenoses and occlusions treated with the Palmaz peripheral vascular stent at Scripps Clinic between August 1990 and November 1991. Symptoms included lower extremity ischemic rest pain or nonhealing ulcer in five patients and intermittent claudication in 31 patients. In one patient who had no prior lower extremity symptoms, acute external iliac artery dissection and occlusion occurred dur-
VoL 7, No. 3 1993
ing cardiac catheterization. The lesions were recurrent in 14 patients, new in 22 patients, and acute in one case. The mean age of patients was 63.7 (42 to 83) years. Twenty-one patients were males, nine patients were diabetic, and 24 patients were hypertensive. Twenty-nine patients had a positive smoking history, and in five, hypercholesterolemia was documented. Informed consent was obtained from each patient after the protocol was approved by the Scripps Clinic Institutional Review Board. A diagnostic arteriogram was performed, and all lesions intended for stent placement initially were subjected to standard balloon angioplasty. Indications for stents included recurrent lesions after previous balloon angioplasty, a complex long, irregular, or eccentric external iliac lesion, and complete iliac occlusion after clot dissolution by thrombolysis. In addition, an unsatisfactory arteriographic appearance after balloon dilatation, a particularly significant arterial dissection, or a residual pressure gradient (->5 m m Hg) were taken as indications for stent placement. Additional stents were applied to mild or moderate stenoses at the origin of the c o m m o n iliac artery w h e n they were contralateral to a lesion showing the usual indications for a stent. Common iliac lesions responsive to simple balloon angioplasty were not treated with stents. The Palmaz stent (Johnson & Johnson Interventional Systems, Warren, N.J.), usually the PS-30 (3 x 30 mm), was mounted on a folded balloon catheter, advanced through an introducer sheath to the desired location, and deployed at the site by inflation of the balloon. Treatment with aspirin and dipyridamole was begun 24 hours prior to the procedure and continued for 3 months. Heparin was used during the procedure, but long-term anticoagulation was not routine. Segmental lower limb pressures were obtained noninvasively before treatment in 35 of the 37 patients (94.6%). All patients were reexamined following the procedure after 1, 3, and 6 months. Pelvic or abdominal radiography was performed 1 m o n t h after stent placement to determine stent position. Arteriograms 6 months after stent placement were recommended routinely and were obtained in 19 patients. Stenoses were expressed as the ratio of the diameter at the site of obstruction to the nearest normal distal diameter. The result of intervention was assessed by comparing the degree of stenoses prior to the procedure, immediately thereafter, and 6 months later w h e n such an arteriogram
Experience with Palmaz stent for aortoiliac stenoses 255
was available. The peak systolic pressure gradient across the lesion was measured before and after stent placement. On the 6-month arteriogram, recurrent stenosis was defined as any degree of narrowing at a site of stent implantation as opposed to the commonly used term "restenosis," which usually implies recurrent stenosis of >50%. The stented segments were considered patent w h e n demonstrated on follow-up arteriograms or in the presence of a normal femoral pulse and an ankle-arm index that was normal or improved by > 0.1 compared to baseline. Group data were expressed as mean values _ SD and compared by Student's t test. A p value of < 0.05 was considered significant.
RESULTS Findings and Procedures Symptomatic stenoses were present in the lower abdominal aorta in three patients, in the c o m m o n iliac artery in 30 patients, and in the external iliac artery in 25 patients. Significant bilateral lesions were found in 22 patients. Concomitant infrainguinal arterial obstruction was present in 22 limbs in 16 patients. Six complete iliac occlusions were recanalized initially with urokinase, after which stents were placed (Fig. 1). Stents were inserted bilaterally in 19 instances, and contralateral balloon dilatation without stenting was performed in three cases. Stents were employed only for the specific indications mentioned above, and during the same time period, 20 aortofemoral bypass procedures and about 60 balloon angioplasty procedures on the iliac arteries were performed. The average degree of maximal stenoses was 57 + 17% (23% to 99%) as measured by reduction of diameter on an arteriogram (Table I). In patients w h o received bilateral stents the average stenosis was 53% on the right and 57% on the left. The m e a n resting initial peak systolic pressure gradient was 45 + 30 m m Hg (4 to 100 m m Hg). Gradients < I0 m m Hg were measured in three patients. In one patient, thrombolysis and balloon dilatation were used initially, and in the other two, stenoses were contralateral to a proximal c o m m o n iliac lesion treated with a stent. A total of 128 stents were implanted in 37 patients (3.5 stents/patient; range 1 to 14). In one patient, six stents were implanted in each iliac system, and in addition, two stents were applied at stenosed distal anastomoses of bilateral femoropopliteal bypass grafts.
Combined femoral endarterectomy and iliac stent
256
Annals of Vascular Surgery
Wolf et al.
A
Fig. 1. A 73-year-old w o m a n w i t h incapacitating left thigh claudication. A, Arteriogram demonstrates occlusion of the left c o m m o n and external iliac arteries. B, Follow-up arteriogram 1 year after thrornbolysis and p l a c e m e n t of stents in the left iliac system.
Table
I. R e s u l t s a f t e r p l a c e m e n t
No. of patients No. of stenotic lesions Stenosis (%) Peak systolic pressure gradient (ram Hg) Upper thigh pressure index (UTPI) Ankle-brachial index (ABI) Claudication (No. of limbs) lschemia at rest (No, of limbs) Deaths
of aortoiliac stents Prior to intervention
Immediate result
6- and 21-month follow-up
37 54 57 +_ 17 (23-99) 45 ± 30 (4-100) 0.82 ± 0.21 0.71 ± 0.19 31 5 0
37 2 1 - 5t (0-30) 1.6 ---+3.4 t (0-20) 0.97 +- 0.04 t 0.89 - 0.15 t 9# 0 1
32* 9 (9-43)
NS = not significantly different from previous column. *Follow-up arteriograms were performed routinely on 19 patients. tp < 0.01, ¢All but one patient improved and infrainguinal obstruction was present in 7 patients,
Not measured 0.96 ± 0.05 (NS} 0.90 ± 0.14 (NS) 7 0 4
Vol. 7, No. 3 1993
Experience with Palmaz stent far aortoiliac stenoses 257
B
A
Fig. 2. A 75-year-old m a n with bilateral intermittent claudication and severe ischemic heart disease. A, Arteriogram demonstrates bilateral common and external iliac artery stenoses and occlusion of the right common femoral artery. B, Follow-up arteriogram 6 months after right common femoral endarterectomy and profundaplasty with simultaneous application of stents to both common itiac arteries, The right superficial femoral artery is occluded,
placement was performed in nine patients with significant common femoral and/or profunda stenoses and ipsilateral iliac artery disease. In three patients, femoral lesions were bilateral, and iliac artery stenoses were bilateral in all patients. Although these patients would have been candidates for aortofemoral bypass, they had severe underlying systemic disease that increased the risk of an intra-abdominal operation. Femoral endarterectomy and profundaplasty were performed initially and iliac stents were placed through the reconstructed femoral artery. These operations were performed in an operating suite with arteriographic capability. In five instances a stent was placed percutaneously at the time of the contralateral femoral endarterectomy (Fig. 2).
Early R e s u l t s Stenoses were reduced to 0% in all but two patients (p < 0.01). In one patient a distal aortic stent was inadequately expanded to 12 mm, which resulted in a residual stenosis of 30%. No additional manipulation was performed at that
time. In the second patient a 23% residual stenosis in a large common iliac artery and a 6 m m Hg pressure gradient were accepted. Peak systolic pressure gradients across all the lesions were reduced to a mean of 1.6 + 3.4 m m Hg (0 to 20 m m Hg) (p < 0.001). Complete elimination of the pressure gradient was achieved in 41 of 56 arteries treated with a stent. A residual gradient of over 10 m m Hg occurred in only one patient across six stents placed in the common and external iliac arteries, which were previously occluded. No acute occlusion or embolization occurred, and the average hospital stay was 2.7 days ( 1 to 6 days) for the whole group and 4 days for those undergoing both femoral endarterectomy and stent placement. Complications included one death attributed to cardiac arrhythmia 1 day after an uncomplicated placement of an iliac stent with no evidence of local stent-related complications. Another patient developed upper gastrointestinal bleeding, which stopped after the anticoagulants were discontinued. Two patients who developed groin hematomas and one patient with a femoral
258
Annals of Vascular Surgery
Wolfet al.
Table II. Recurrences after placement of aortoiliac stents Sites of stents (no.) Symptomatic* A o r t a (1), Rt CIA (1) Rt CIA (2); Lt CIA (1) ? Rt EIA (1); Lt CIA (1) Rt CIA, EIA (4); Lt EIA (2) Rt EIA (4) Rt CIA, EIA (5); Lt CIA, EIA (4) ? A s y m p t o m a t i c ~: Rt CIA, EIA (2); Lt CIA, EIA (2) Rt CIA (2); Lt CIA (1) Rt CIA (2); Lt CIA, EIA (3) Rt CIA, EIA (4); Lt CIA (1)
Time to recurrence (mo) 5 15 13 7 6 8 6 6 6 6
Length of follow-up (mo)
Site of recurrence and degree
Secondary procedures
Aorta, 41% Lt CIA, 40% Rt EIA, 23%; Lt CIA, 9% :~ Outside s t e n t sites O u t s i d e s t e n t sites I n s t e n t a n d i n CFA
19 17 16 18 11 10
R e e x p a n s i o n of s t e n t None None A o r t o f e m o r a l bypass A o r t o f e m o r a l bypass E x p a n s i o n s of s t e n t s a n d PTA
Lt Rt Rt Lt
8 13 11 16
None None None None
EIA, CIA, CIA, CIA,
27% 10% 23%; Lt CIA, 20% 43%
CIA = common iliac artery; EIA = external iliac artery; CFA = common femoral artery; PTA = percutaneous transluminal angioplasty. *All symptomatic recurrences manifested with intermittent claudication. ?Done in conjuction with common femoral endarterectomy. 1:Identified on routine arteriogram after 6 months.
pseudoaneurysm were treated nonoperatively. No w o u n d complications occurred in the surgical subgroup. Lower limb segmental pressures were obtained noninvasively before and after the procedure in 52 limbs in 34 patients. The mean resting upper thigh pressure index (UTPI) was 0.82 _+ 0.21 initially and increased to 0.97 _+ 0.04 (p < 0.01) after intervention. In 21 limbs the initial UTPI was > 0.9 and it increased by at least 0.1 in 28 limbs. The mean resting ankle-brachial pressure index (ABI) was 0.71 +_ 0.19 initially and increased to 0.89 _ 0.15 (p <0.01) after intervention. In 10 limbs the initial ABI was > 0.90. Three of these were treated because of a contralateral stent, one for dissection during cardiac catheterization, two for aortic lesions, two for symptomatic recurrent common iliac stenosis, and two for primary iliac stenosis. The index increased by 0.10 or more in 31 limbs. Two patients underwent an early scheduled operation after stent placement. In one patient with complete obstruction of the left iliac artery a femorofemoral by~ass was constructed 1 day after placement of right iliac stents. In another, femoral endarterectomy and profundaplasty with intraoperative superficial femoral arthrectomy were performed 3 days after placement of an aortic stent.
Follow-Up The 36 patients w h o were discharged from the hospital were followed for an average of 12
T a b l e III. Outcome of aortoiliac stents by site
Site
No. of stents
Symptomatic recurrent stenosis
Asymptomatic stenosis ( > 25%)
Aorta CIA EIA
3 52 73
1 1 2
0 1 1
recurrent
CIA = common iliac artery; EIA = external iliac artery.
months (6 to 21 months). Four patients died during follow-up of myocardial infarction, lung cancer, and multiple myeloma. Symptoms resolved completely in 27 patients (73%), improved in eight patients (22%), and remained unchanged in one patient (3%). Of nine patients with incomplete relief of symptoms, seven had arteriographically documented infrainguinal arterial obstruction. In one patient, inadequate expansion of a stent in the distal abdominal aorta resulted in early recurrence of symptoms. No stent migration was identified on plain radiograms obtained 1 month after the procedure. Routine arteriograms were performed 6 months after stent placement in 19 patients with 55 stents (Figs. 1, /3, and 2, B), and recurrent stenoses were found at the site of eight stents in six patients. No stenosis was > 50%, the mean diameter reduction was 24% (9% to 43%), and only two patients were symptomatic (Tables II and III). In three patients, stents crossed the origins of two internal iliac arteries and an inferior mesen-
Vol, 7, No, 3 1993
Experience with Palmaz stent for aortoiliac stenoses 259
teric artery, and these branches have remained patent on the 6-month arteriogram. During a mean follow-up of 12 months, clinical improvement was maintained in 30 (86%) patients. The increase in the UTPI was maintained
in 23 of 27 limbs (0.956 _ 0.052) and in the ABI in 26 of 31 limbs (0.899 ± 0.142), and both mean values were not significantly different from the immediate studies after stent placement (Table I). The six patients who developed recurrent symptoms (16%) and four asymptomatic patients (11%) in w h o m recurrent stenoses were found on routine arteriograms are listed in Table II. Recurrent symptoms were intermittent claudication in all cases, and in two patients with mild recurrent claudication, no additional procedures were performed. One patient with an incompletely expanded aortic stent improved after further expansion of the stent. After bilateral femoral endarterectomy and iliac stent placement one patient developed high-grade recurrent stenoses within the stent sites as well as just beyond the termination of the most distal stent, at the inguinal ligament, bilaterally (Fig. 3). This was treated by further expansion of the stents and balloon dilatation of the distal external iliac arteries with the expectation that surgical reconstruction will eventually be required. Two patients with progressive stenoses at sites different from the stent locations were treated with aortofemoral bypass. No stent occluded during follow-up, but w h e n symptomatic recurrence > 40% (one patient), recurrent stenosis requiring redilatation (two patients), and construction of aortofemoral bypass are considered failures, the cumulative patency at 1 year was 94% (Table IV).
DISCUSSION
Fig. 3. Arteriogram of right groin 8 months after bilateral femoral endarterectomy and iliac stent placement. The patient developed recurrent claudication, and recurrent stenosis is evident beyond the distal right external iliac stent.
The recent development of the intravascular stent has increased our ability to deal with difficult vascular stenoses nonoperatively, and the present assessment of this new intervention confirms its early efficacy. Although this initial experience at a single institution over a 15-month period is small and the follow-up is limited, we believe sufficient data are available to permit several conclusions.
Table IV. Life-table analysis of success rate following aortoiliac stent placement Interval (mo)
No. of limbs*
Failed"t"
Duration/loss
0-6 6-12 12-18
56 52 28
2 1 3
0 22 20
Death
Interval patency
Cumulative patency (%)
SE (%)
2 1 4
0.96 0.98 0.81
100 96 94
0 2.7 4.4
*Analysis by limbs treated; aortic stents alone are counted as one limb. tNo graft occlusion occurred. Failures include symptomatic restenosis (> 40%); stents requiring repeat dilatation and construction of an aortofemoral bypass in two patients (counted as three limbs).
260
Wolf et aL
The stent has proved extremely effective in improving the immediate arteriographic result following iliac artery balloon angioplasty. Stenoses were completely corrected in all but two patients, and the peak systolic pressure gradients were completely eliminated in the majority of lesions and reduced to < 10 m m Hg in all but one patient. When compared with balloon angioplasty alone, which is associated with an immediate failure rate of up to 10% and residual pressure gradients > 1 0 m m Hg in 22% of cases, ~° the addition of a stent on a selective basis seems reasonable. The short-term follow-up of patients with the Palmaz stent appears promising. Clinical improvement was achieved in 34 of 36 patients and was maintained over a mean follow-up of 12 months in 30 (86%) patients. This compares favorably with the results of common and external iliac artery balloon angioplasty, which are associated with a 1-year success rate of 78% and 73%, respectively)' It is important to note that our patients were selected for unfavorable lesions such as recurrent or complex external iliac lesions, complete occlusions, and unsatisfactory result after angioplasty. These patients would be expected to do worse with angioplasty alone. After 6 months routine arteriograms demonstrated in all patients minimal nonobstructive intimal hyperplasia along the stent, which covered the exposed stents and isolated the stent from the blood. If this process is more proliferative, recurrent stenoses within the stent lumen will result, and this occurred in six patients at eight stent sites. Unfortunately, factors that influence the degree of intimal hyperplasia at the stent site remain unknown. In a large multicenter study Palmaz et al. ~2 found intimal hyperplasia measuring 0.79 _+ 0.52 m m and a patency rate of 100% on 6-month arteriographic follow-up of 45 patients. Results with other stents have recently been reported. In one study in which the Wallstent was used in 26 iliac arteries in 18 patients, one occlusion and four stenoses developed over a mean follow-up of 16 months. ~3 In 31 patients with Strecker stents in the iliac artery, one symptomatic restenosis was found after 2V2 months, ~4 and in a series reporting placement of both Wallstents and Strecker stents in 49 patients, primary patency at 1 year was 85.3% and secondary patency was 96.1%./~ Results with the Palmaz stent compare favorably with these series, and the feasibility of further stent expansion in cases of restenosis is
Annals of Vascular Surgery
appealing, but longer follow-up is clearly necessary. The long-term patency rate of stents in the aortoiliac location will probably be lower than the potency rate for aortofemoral bypass grafts. First, there is restenosis within the stents, and in addition, occlusive disease can progress in native vessels proximal and distal to the stented segments, as demonstrated in two of our patients who ultimately underwent aortofemoral bypass. As a temporizing measure in patients with limited life expectancies, the success rate of iliac stents appears acceptable and does not compromise the possibility of ultimate aortofemoral bypass. In nine patients we performed a combined procedure of surgical outflow reconstruction and intraoperative placement of iliac stents through the reconstructed femoral artery. These patients had severe underlying cardiac or pulmonary disease and were considered at high risk for intra-. abdominal operations. In each instance excellent early results were obtained, and our concerns about the possibility of w o u n d infection did not materialize. One patient who underwent bilateral femoral endarterectomy and placement of a total of nine stents in both common and external iliac arteries developed symptomatic recurrence after 8 months. In addition to stenoses within the stent sites, bilateral stenoses just beyond the distal stent in the proximal common femoral artery at the level of the inguinal ligament were identified arteriographically (Fig. 3). In our experience, early recurrent stenosis after femoral endarterectomy is u n c o m m o n at this location, and we postulate that it m a y be related to the proximity of a stent to the mobile segment of the femoral artery in the groin. In the eight other patients, combined femoral endarterectomy with iliac stent placement has produced satisfactory results, and it may be an alternative to extra-anatomic bypass in selected high-risk patients. Although our early experience is encouraging, m a n y unanswered questions remain. In this series some degree of recurrent stenosis occurred at 11 stent sites in eight patients who were studied arteriographically; thus the stent is not immune to the development of intimal hyperplasia. Furthermore, stenosis proximal or distal to the stents may progress to hemodynamic significance, as shown in three of our patients. Complications occurred in five patients, including one death. Although the problems were handled nonoperatively, the incidence is high. In addition, the stent implantation procedures require 1 to 3 hours and
Vol. 7, No. 3 1993
in two instances resulted in blood loss necessitating transfusion of 2 units of blood. Finally, these procedures are expensive, and although hospital stay is short, the cost closely approaches the more standard therapeutic approach of aortofemoral bypass. Because of recent approval by the Food and Drug Administration of the use of the stent in the iliac system, we anticipate greater application of this modality and its attendant risks and recurrences.
CONCLUSION The Palmaz intravascular stent appears to confer benefit in the treatment of aortoiliac occlusive disease and may be combined with femoral endarterectomy and profundaplasty. Close follow-up and study of these patients will be required. The participation of vascular surgeons in the evaluation of these patients and in the assessment of the therapeutic results is important. REFERENCES 1. Dotter CT. Transtuminally placed coil-spring endarterial tube grafts: Long-term patency in canine popliteal artery. Invest Radiol 1969;4:329-332. 2. Cragg A, Lund G, Rysavy J, et al. Nonsurgical placement of arterial endoprosthesis: A new technique using nitinol wire. Radiology 1983;147:261. 3. Maass D, Zollikofer CL, Largiader F, et at. Radiological follow-up of transluminally inserted vascular endoprostheses: An experimental study using expanding spirals. Radiology 1983;152:659-663.
Experience with Palmaz stent for aortoiliac stenoses 261
4. Wright KC, Wallace S, Charnsangavej C, et al. Percutaneous vascular stents: An experimental evaluation. Radiology 1985; 156:69-72. 5. Trent MS, Parsonnet V, Shoenfeld R, et al. A balloon-expandable intravascular stent for obliterating experimental aortic dissection. J Vasc Surg 1990;11:707-717. 6. Sigwart U, Puel J, Mirkovitch V, et al. Intravascular stents to prevent occlusion and restenosis after transluminal angioplasty. N Engl J Med 1987;316:701-706. 7. Palmaz JC, Sibbitt RR, Tio FO, et al. Expandable intraluminal vascular graft: A feasibility study. Surgery 1986;99:199-205. 8. Schatz RA. A view of vascular stems. Circulation 1989;79: 445-457. 9. Penn IM, Levine SL, Schatz RA. Intravascular stents as an adjunct to endovascular intervention: In Moore WS, Ahn SS, eds. Endovascular Surgery. Philadelphia: WB Sannders, 1989, pp 258-277. 10. Cambria RP, Faust G, Gusberg R, et al. Percutaneous angioplasty for peripheral arterial occlusive disease. Arch Surg 1987; 122:283-287. 11. Johnston KW, Rae M, Hogg-Johnston S, et al. Five-year results of a prospective study of percutaneous transluminal angioplasty. Ann Surg 1987;206:403-413. 12. Palmaz JC, Garcia OJ, Schatz RA, et al. Placement of balloonexpandable intraluminal stents in iliac arteries: First 171 procedures. Radiology 1990;174:969-975. 13. Zollikofer CL, Antonucci F, Pfyffer M, et al. Arterial stem placement with use of the Wallstent: Midterm results of clinical experience. Radiology 1991;179:449-456. 14. Gunther WW, Vorwerk 13, Bohndorf K, et al. Iliac and femoral artery stenoses and occlusions: Treatment with intravascular stents. Radiology 1989;172:725-730. 15. Long AL, Page PE, Raynaud AC, et al. Percutaneous iliac artery stent: Angiographic long-term follow-up. Radiology I991;180:771-778.
The initial 37 consecutive patients to be treated at our institution with the Palmaz stent placed in the aortoiliac arteries were retrospectively reviewed. In these patients, 50 stenoses and six occlusions were treated with 128 stents. Nine patients with combined iliac and common femoral obstruction underwent common femoral endarterectomy and profundaplasty with intraoperative iliac artery angioplasty and stent application. Stenoses were reduced from 57 - 17% to 1 _ 5% (p <0.01), and peak systolic pressure gradients across the lesions were reduced from 45 _ 30 mm Hg to 1.3 _+ 3.4 mm Hg (p <0.01). Symptoms resolved in 27 patients and improved in eight patients. One patient died and four patients were treated nonoperatively for complications. During a mean follow-up of 12 months (6 to 21 months), six patients had recurrence of symptoms (16%) and four patients died of other diseases. Routine arteriograms after 6 months in 19 patients demonstrated recurrent mild to moderate stenoses (9% to 43%) in six patients (32%), but only two were symptomatic (11%). Secondary procedures included reexpansion of aortic and iliac stents in two patients and aortofemoral bypass in two patients. Early results suggest the efficacy of the Palmaz stent in the management of aortoiliac stenoses and its use intraoperatively in conjunction with surgical correction of outflow. Close follow-up of these patients by multidisciplinary groups is warranted. (Ann Vasc Surg 1993;7:254-261.)
The first report of an experimental intravascular metal stent appeared in 1969.1 Several designs were investigated experimentally for both arterial stenosis and dissection. 2-~ The first clinical report of self-expanding intravascular stents in the coronary and iliofemoral arteries appeared in 1987. 6 The Palmaz stent, which was the first balloonexpandable stent, was initially developed as a woven-wire device~ but was subsequently simplified into a slotted stainless steel tube. It is one of several intravascular stents currently in clinical use s'9 and recently was approved by the Food and From the Divisions of Vascular Surgery, Cardiology (R,A.S.), and Radiology (H.J.K. and M.S.), Scripps Clinic and Research Foundation, La Jolla, California. Reprint requests: Ralph B. Dilley, MD, Division of Vascular Surgery, Scripps Clinic and Research Foundation, 10666 N. Torrey Pines Rd., La Jolla, CA 92037.
254
Drug Administration for use in the iliac arteries. We have used it at the Scripps Clinic and Research Foundation since August 1990. This review was undertaken to examine our early experience with this device and to ascertain its role in the m a n a g e m e n t of aortoiliac stenosis. PATIENTS AND METHODS This retrospective review includes the initial 37 consecutive patients with aortoiliac stenoses and occlusions treated with the Palmaz peripheral vascular stent at Scripps Clinic between August 1990 and November 1991. Symptoms included lower extremity ischemic rest pain or nonhealing ulcer in five patients and intermittent claudication in 31 patients. In one patient who had no prior lower extremity symptoms, acute external iliac artery dissection and occlusion occurred dur-
VoL 7, No. 3 1993
ing cardiac catheterization. The lesions were recurrent in 14 patients, new in 22 patients, and acute in one case. The mean age of patients was 63.7 (42 to 83) years. Twenty-one patients were males, nine patients were diabetic, and 24 patients were hypertensive. Twenty-nine patients had a positive smoking history, and in five, hypercholesterolemia was documented. Informed consent was obtained from each patient after the protocol was approved by the Scripps Clinic Institutional Review Board. A diagnostic arteriogram was performed, and all lesions intended for stent placement initially were subjected to standard balloon angioplasty. Indications for stents included recurrent lesions after previous balloon angioplasty, a complex long, irregular, or eccentric external iliac lesion, and complete iliac occlusion after clot dissolution by thrombolysis. In addition, an unsatisfactory arteriographic appearance after balloon dilatation, a particularly significant arterial dissection, or a residual pressure gradient (->5 m m Hg) were taken as indications for stent placement. Additional stents were applied to mild or moderate stenoses at the origin of the c o m m o n iliac artery w h e n they were contralateral to a lesion showing the usual indications for a stent. Common iliac lesions responsive to simple balloon angioplasty were not treated with stents. The Palmaz stent (Johnson & Johnson Interventional Systems, Warren, N.J.), usually the PS-30 (3 x 30 mm), was mounted on a folded balloon catheter, advanced through an introducer sheath to the desired location, and deployed at the site by inflation of the balloon. Treatment with aspirin and dipyridamole was begun 24 hours prior to the procedure and continued for 3 months. Heparin was used during the procedure, but long-term anticoagulation was not routine. Segmental lower limb pressures were obtained noninvasively before treatment in 35 of the 37 patients (94.6%). All patients were reexamined following the procedure after 1, 3, and 6 months. Pelvic or abdominal radiography was performed 1 m o n t h after stent placement to determine stent position. Arteriograms 6 months after stent placement were recommended routinely and were obtained in 19 patients. Stenoses were expressed as the ratio of the diameter at the site of obstruction to the nearest normal distal diameter. The result of intervention was assessed by comparing the degree of stenoses prior to the procedure, immediately thereafter, and 6 months later w h e n such an arteriogram
Experience with Palmaz stent for aortoiliac stenoses 255
was available. The peak systolic pressure gradient across the lesion was measured before and after stent placement. On the 6-month arteriogram, recurrent stenosis was defined as any degree of narrowing at a site of stent implantation as opposed to the commonly used term "restenosis," which usually implies recurrent stenosis of >50%. The stented segments were considered patent w h e n demonstrated on follow-up arteriograms or in the presence of a normal femoral pulse and an ankle-arm index that was normal or improved by > 0.1 compared to baseline. Group data were expressed as mean values _ SD and compared by Student's t test. A p value of < 0.05 was considered significant.
RESULTS Findings and Procedures Symptomatic stenoses were present in the lower abdominal aorta in three patients, in the c o m m o n iliac artery in 30 patients, and in the external iliac artery in 25 patients. Significant bilateral lesions were found in 22 patients. Concomitant infrainguinal arterial obstruction was present in 22 limbs in 16 patients. Six complete iliac occlusions were recanalized initially with urokinase, after which stents were placed (Fig. 1). Stents were inserted bilaterally in 19 instances, and contralateral balloon dilatation without stenting was performed in three cases. Stents were employed only for the specific indications mentioned above, and during the same time period, 20 aortofemoral bypass procedures and about 60 balloon angioplasty procedures on the iliac arteries were performed. The average degree of maximal stenoses was 57 + 17% (23% to 99%) as measured by reduction of diameter on an arteriogram (Table I). In patients w h o received bilateral stents the average stenosis was 53% on the right and 57% on the left. The m e a n resting initial peak systolic pressure gradient was 45 + 30 m m Hg (4 to 100 m m Hg). Gradients < I0 m m Hg were measured in three patients. In one patient, thrombolysis and balloon dilatation were used initially, and in the other two, stenoses were contralateral to a proximal c o m m o n iliac lesion treated with a stent. A total of 128 stents were implanted in 37 patients (3.5 stents/patient; range 1 to 14). In one patient, six stents were implanted in each iliac system, and in addition, two stents were applied at stenosed distal anastomoses of bilateral femoropopliteal bypass grafts.
Combined femoral endarterectomy and iliac stent
256
Annals of Vascular Surgery
Wolf et al.
A
Fig. 1. A 73-year-old w o m a n w i t h incapacitating left thigh claudication. A, Arteriogram demonstrates occlusion of the left c o m m o n and external iliac arteries. B, Follow-up arteriogram 1 year after thrornbolysis and p l a c e m e n t of stents in the left iliac system.
Table
I. R e s u l t s a f t e r p l a c e m e n t
No. of patients No. of stenotic lesions Stenosis (%) Peak systolic pressure gradient (ram Hg) Upper thigh pressure index (UTPI) Ankle-brachial index (ABI) Claudication (No. of limbs) lschemia at rest (No, of limbs) Deaths
of aortoiliac stents Prior to intervention
Immediate result
6- and 21-month follow-up
37 54 57 +_ 17 (23-99) 45 ± 30 (4-100) 0.82 ± 0.21 0.71 ± 0.19 31 5 0
37 2 1 - 5t (0-30) 1.6 ---+3.4 t (0-20) 0.97 +- 0.04 t 0.89 - 0.15 t 9# 0 1
32* 9 (9-43)
NS = not significantly different from previous column. *Follow-up arteriograms were performed routinely on 19 patients. tp < 0.01, ¢All but one patient improved and infrainguinal obstruction was present in 7 patients,
Not measured 0.96 ± 0.05 (NS} 0.90 ± 0.14 (NS) 7 0 4
Vol. 7, No. 3 1993
Experience with Palmaz stent far aortoiliac stenoses 257
B
A
Fig. 2. A 75-year-old m a n with bilateral intermittent claudication and severe ischemic heart disease. A, Arteriogram demonstrates bilateral common and external iliac artery stenoses and occlusion of the right common femoral artery. B, Follow-up arteriogram 6 months after right common femoral endarterectomy and profundaplasty with simultaneous application of stents to both common itiac arteries, The right superficial femoral artery is occluded,
placement was performed in nine patients with significant common femoral and/or profunda stenoses and ipsilateral iliac artery disease. In three patients, femoral lesions were bilateral, and iliac artery stenoses were bilateral in all patients. Although these patients would have been candidates for aortofemoral bypass, they had severe underlying systemic disease that increased the risk of an intra-abdominal operation. Femoral endarterectomy and profundaplasty were performed initially and iliac stents were placed through the reconstructed femoral artery. These operations were performed in an operating suite with arteriographic capability. In five instances a stent was placed percutaneously at the time of the contralateral femoral endarterectomy (Fig. 2).
Early R e s u l t s Stenoses were reduced to 0% in all but two patients (p < 0.01). In one patient a distal aortic stent was inadequately expanded to 12 mm, which resulted in a residual stenosis of 30%. No additional manipulation was performed at that
time. In the second patient a 23% residual stenosis in a large common iliac artery and a 6 m m Hg pressure gradient were accepted. Peak systolic pressure gradients across all the lesions were reduced to a mean of 1.6 + 3.4 m m Hg (0 to 20 m m Hg) (p < 0.001). Complete elimination of the pressure gradient was achieved in 41 of 56 arteries treated with a stent. A residual gradient of over 10 m m Hg occurred in only one patient across six stents placed in the common and external iliac arteries, which were previously occluded. No acute occlusion or embolization occurred, and the average hospital stay was 2.7 days ( 1 to 6 days) for the whole group and 4 days for those undergoing both femoral endarterectomy and stent placement. Complications included one death attributed to cardiac arrhythmia 1 day after an uncomplicated placement of an iliac stent with no evidence of local stent-related complications. Another patient developed upper gastrointestinal bleeding, which stopped after the anticoagulants were discontinued. Two patients who developed groin hematomas and one patient with a femoral
258
Annals of Vascular Surgery
Wolfet al.
Table II. Recurrences after placement of aortoiliac stents Sites of stents (no.) Symptomatic* A o r t a (1), Rt CIA (1) Rt CIA (2); Lt CIA (1) ? Rt EIA (1); Lt CIA (1) Rt CIA, EIA (4); Lt EIA (2) Rt EIA (4) Rt CIA, EIA (5); Lt CIA, EIA (4) ? A s y m p t o m a t i c ~: Rt CIA, EIA (2); Lt CIA, EIA (2) Rt CIA (2); Lt CIA (1) Rt CIA (2); Lt CIA, EIA (3) Rt CIA, EIA (4); Lt CIA (1)
Time to recurrence (mo) 5 15 13 7 6 8 6 6 6 6
Length of follow-up (mo)
Site of recurrence and degree
Secondary procedures
Aorta, 41% Lt CIA, 40% Rt EIA, 23%; Lt CIA, 9% :~ Outside s t e n t sites O u t s i d e s t e n t sites I n s t e n t a n d i n CFA
19 17 16 18 11 10
R e e x p a n s i o n of s t e n t None None A o r t o f e m o r a l bypass A o r t o f e m o r a l bypass E x p a n s i o n s of s t e n t s a n d PTA
Lt Rt Rt Lt
8 13 11 16
None None None None
EIA, CIA, CIA, CIA,
27% 10% 23%; Lt CIA, 20% 43%
CIA = common iliac artery; EIA = external iliac artery; CFA = common femoral artery; PTA = percutaneous transluminal angioplasty. *All symptomatic recurrences manifested with intermittent claudication. ?Done in conjuction with common femoral endarterectomy. 1:Identified on routine arteriogram after 6 months.
pseudoaneurysm were treated nonoperatively. No w o u n d complications occurred in the surgical subgroup. Lower limb segmental pressures were obtained noninvasively before and after the procedure in 52 limbs in 34 patients. The mean resting upper thigh pressure index (UTPI) was 0.82 _+ 0.21 initially and increased to 0.97 _+ 0.04 (p < 0.01) after intervention. In 21 limbs the initial UTPI was > 0.9 and it increased by at least 0.1 in 28 limbs. The mean resting ankle-brachial pressure index (ABI) was 0.71 +_ 0.19 initially and increased to 0.89 _ 0.15 (p <0.01) after intervention. In 10 limbs the initial ABI was > 0.90. Three of these were treated because of a contralateral stent, one for dissection during cardiac catheterization, two for aortic lesions, two for symptomatic recurrent common iliac stenosis, and two for primary iliac stenosis. The index increased by 0.10 or more in 31 limbs. Two patients underwent an early scheduled operation after stent placement. In one patient with complete obstruction of the left iliac artery a femorofemoral by~ass was constructed 1 day after placement of right iliac stents. In another, femoral endarterectomy and profundaplasty with intraoperative superficial femoral arthrectomy were performed 3 days after placement of an aortic stent.
Follow-Up The 36 patients w h o were discharged from the hospital were followed for an average of 12
T a b l e III. Outcome of aortoiliac stents by site
Site
No. of stents
Symptomatic recurrent stenosis
Asymptomatic stenosis ( > 25%)
Aorta CIA EIA
3 52 73
1 1 2
0 1 1
recurrent
CIA = common iliac artery; EIA = external iliac artery.
months (6 to 21 months). Four patients died during follow-up of myocardial infarction, lung cancer, and multiple myeloma. Symptoms resolved completely in 27 patients (73%), improved in eight patients (22%), and remained unchanged in one patient (3%). Of nine patients with incomplete relief of symptoms, seven had arteriographically documented infrainguinal arterial obstruction. In one patient, inadequate expansion of a stent in the distal abdominal aorta resulted in early recurrence of symptoms. No stent migration was identified on plain radiograms obtained 1 month after the procedure. Routine arteriograms were performed 6 months after stent placement in 19 patients with 55 stents (Figs. 1, /3, and 2, B), and recurrent stenoses were found at the site of eight stents in six patients. No stenosis was > 50%, the mean diameter reduction was 24% (9% to 43%), and only two patients were symptomatic (Tables II and III). In three patients, stents crossed the origins of two internal iliac arteries and an inferior mesen-
Vol, 7, No, 3 1993
Experience with Palmaz stent for aortoiliac stenoses 259
teric artery, and these branches have remained patent on the 6-month arteriogram. During a mean follow-up of 12 months, clinical improvement was maintained in 30 (86%) patients. The increase in the UTPI was maintained
in 23 of 27 limbs (0.956 _ 0.052) and in the ABI in 26 of 31 limbs (0.899 ± 0.142), and both mean values were not significantly different from the immediate studies after stent placement (Table I). The six patients who developed recurrent symptoms (16%) and four asymptomatic patients (11%) in w h o m recurrent stenoses were found on routine arteriograms are listed in Table II. Recurrent symptoms were intermittent claudication in all cases, and in two patients with mild recurrent claudication, no additional procedures were performed. One patient with an incompletely expanded aortic stent improved after further expansion of the stent. After bilateral femoral endarterectomy and iliac stent placement one patient developed high-grade recurrent stenoses within the stent sites as well as just beyond the termination of the most distal stent, at the inguinal ligament, bilaterally (Fig. 3). This was treated by further expansion of the stents and balloon dilatation of the distal external iliac arteries with the expectation that surgical reconstruction will eventually be required. Two patients with progressive stenoses at sites different from the stent locations were treated with aortofemoral bypass. No stent occluded during follow-up, but w h e n symptomatic recurrence > 40% (one patient), recurrent stenosis requiring redilatation (two patients), and construction of aortofemoral bypass are considered failures, the cumulative patency at 1 year was 94% (Table IV).
DISCUSSION
Fig. 3. Arteriogram of right groin 8 months after bilateral femoral endarterectomy and iliac stent placement. The patient developed recurrent claudication, and recurrent stenosis is evident beyond the distal right external iliac stent.
The recent development of the intravascular stent has increased our ability to deal with difficult vascular stenoses nonoperatively, and the present assessment of this new intervention confirms its early efficacy. Although this initial experience at a single institution over a 15-month period is small and the follow-up is limited, we believe sufficient data are available to permit several conclusions.
Table IV. Life-table analysis of success rate following aortoiliac stent placement Interval (mo)
No. of limbs*
Failed"t"
Duration/loss
0-6 6-12 12-18
56 52 28
2 1 3
0 22 20
Death
Interval patency
Cumulative patency (%)
SE (%)
2 1 4
0.96 0.98 0.81
100 96 94
0 2.7 4.4
*Analysis by limbs treated; aortic stents alone are counted as one limb. tNo graft occlusion occurred. Failures include symptomatic restenosis (> 40%); stents requiring repeat dilatation and construction of an aortofemoral bypass in two patients (counted as three limbs).
260
Wolf et aL
The stent has proved extremely effective in improving the immediate arteriographic result following iliac artery balloon angioplasty. Stenoses were completely corrected in all but two patients, and the peak systolic pressure gradients were completely eliminated in the majority of lesions and reduced to < 10 m m Hg in all but one patient. When compared with balloon angioplasty alone, which is associated with an immediate failure rate of up to 10% and residual pressure gradients > 1 0 m m Hg in 22% of cases, ~° the addition of a stent on a selective basis seems reasonable. The short-term follow-up of patients with the Palmaz stent appears promising. Clinical improvement was achieved in 34 of 36 patients and was maintained over a mean follow-up of 12 months in 30 (86%) patients. This compares favorably with the results of common and external iliac artery balloon angioplasty, which are associated with a 1-year success rate of 78% and 73%, respectively)' It is important to note that our patients were selected for unfavorable lesions such as recurrent or complex external iliac lesions, complete occlusions, and unsatisfactory result after angioplasty. These patients would be expected to do worse with angioplasty alone. After 6 months routine arteriograms demonstrated in all patients minimal nonobstructive intimal hyperplasia along the stent, which covered the exposed stents and isolated the stent from the blood. If this process is more proliferative, recurrent stenoses within the stent lumen will result, and this occurred in six patients at eight stent sites. Unfortunately, factors that influence the degree of intimal hyperplasia at the stent site remain unknown. In a large multicenter study Palmaz et al. ~2 found intimal hyperplasia measuring 0.79 _+ 0.52 m m and a patency rate of 100% on 6-month arteriographic follow-up of 45 patients. Results with other stents have recently been reported. In one study in which the Wallstent was used in 26 iliac arteries in 18 patients, one occlusion and four stenoses developed over a mean follow-up of 16 months. ~3 In 31 patients with Strecker stents in the iliac artery, one symptomatic restenosis was found after 2V2 months, ~4 and in a series reporting placement of both Wallstents and Strecker stents in 49 patients, primary patency at 1 year was 85.3% and secondary patency was 96.1%./~ Results with the Palmaz stent compare favorably with these series, and the feasibility of further stent expansion in cases of restenosis is
Annals of Vascular Surgery
appealing, but longer follow-up is clearly necessary. The long-term patency rate of stents in the aortoiliac location will probably be lower than the potency rate for aortofemoral bypass grafts. First, there is restenosis within the stents, and in addition, occlusive disease can progress in native vessels proximal and distal to the stented segments, as demonstrated in two of our patients who ultimately underwent aortofemoral bypass. As a temporizing measure in patients with limited life expectancies, the success rate of iliac stents appears acceptable and does not compromise the possibility of ultimate aortofemoral bypass. In nine patients we performed a combined procedure of surgical outflow reconstruction and intraoperative placement of iliac stents through the reconstructed femoral artery. These patients had severe underlying cardiac or pulmonary disease and were considered at high risk for intra-. abdominal operations. In each instance excellent early results were obtained, and our concerns about the possibility of w o u n d infection did not materialize. One patient who underwent bilateral femoral endarterectomy and placement of a total of nine stents in both common and external iliac arteries developed symptomatic recurrence after 8 months. In addition to stenoses within the stent sites, bilateral stenoses just beyond the distal stent in the proximal common femoral artery at the level of the inguinal ligament were identified arteriographically (Fig. 3). In our experience, early recurrent stenosis after femoral endarterectomy is u n c o m m o n at this location, and we postulate that it m a y be related to the proximity of a stent to the mobile segment of the femoral artery in the groin. In the eight other patients, combined femoral endarterectomy with iliac stent placement has produced satisfactory results, and it may be an alternative to extra-anatomic bypass in selected high-risk patients. Although our early experience is encouraging, m a n y unanswered questions remain. In this series some degree of recurrent stenosis occurred at 11 stent sites in eight patients who were studied arteriographically; thus the stent is not immune to the development of intimal hyperplasia. Furthermore, stenosis proximal or distal to the stents may progress to hemodynamic significance, as shown in three of our patients. Complications occurred in five patients, including one death. Although the problems were handled nonoperatively, the incidence is high. In addition, the stent implantation procedures require 1 to 3 hours and
Vol. 7, No. 3 1993
in two instances resulted in blood loss necessitating transfusion of 2 units of blood. Finally, these procedures are expensive, and although hospital stay is short, the cost closely approaches the more standard therapeutic approach of aortofemoral bypass. Because of recent approval by the Food and Drug Administration of the use of the stent in the iliac system, we anticipate greater application of this modality and its attendant risks and recurrences.
CONCLUSION The Palmaz intravascular stent appears to confer benefit in the treatment of aortoiliac occlusive disease and may be combined with femoral endarterectomy and profundaplasty. Close follow-up and study of these patients will be required. The participation of vascular surgeons in the evaluation of these patients and in the assessment of the therapeutic results is important. REFERENCES 1. Dotter CT. Transtuminally placed coil-spring endarterial tube grafts: Long-term patency in canine popliteal artery. Invest Radiol 1969;4:329-332. 2. Cragg A, Lund G, Rysavy J, et al. Nonsurgical placement of arterial endoprosthesis: A new technique using nitinol wire. Radiology 1983;147:261. 3. Maass D, Zollikofer CL, Largiader F, et at. Radiological follow-up of transluminally inserted vascular endoprostheses: An experimental study using expanding spirals. Radiology 1983;152:659-663.
Experience with Palmaz stent for aortoiliac stenoses 261
4. Wright KC, Wallace S, Charnsangavej C, et al. Percutaneous vascular stents: An experimental evaluation. Radiology 1985; 156:69-72. 5. Trent MS, Parsonnet V, Shoenfeld R, et al. A balloon-expandable intravascular stent for obliterating experimental aortic dissection. J Vasc Surg 1990;11:707-717. 6. Sigwart U, Puel J, Mirkovitch V, et al. Intravascular stents to prevent occlusion and restenosis after transluminal angioplasty. N Engl J Med 1987;316:701-706. 7. Palmaz JC, Sibbitt RR, Tio FO, et al. Expandable intraluminal vascular graft: A feasibility study. Surgery 1986;99:199-205. 8. Schatz RA. A view of vascular stems. Circulation 1989;79: 445-457. 9. Penn IM, Levine SL, Schatz RA. Intravascular stents as an adjunct to endovascular intervention: In Moore WS, Ahn SS, eds. Endovascular Surgery. Philadelphia: WB Sannders, 1989, pp 258-277. 10. Cambria RP, Faust G, Gusberg R, et al. Percutaneous angioplasty for peripheral arterial occlusive disease. Arch Surg 1987; 122:283-287. 11. Johnston KW, Rae M, Hogg-Johnston S, et al. Five-year results of a prospective study of percutaneous transluminal angioplasty. Ann Surg 1987;206:403-413. 12. Palmaz JC, Garcia OJ, Schatz RA, et al. Placement of balloonexpandable intraluminal stents in iliac arteries: First 171 procedures. Radiology 1990;174:969-975. 13. Zollikofer CL, Antonucci F, Pfyffer M, et al. Arterial stem placement with use of the Wallstent: Midterm results of clinical experience. Radiology 1991;179:449-456. 14. Gunther WW, Vorwerk 13, Bohndorf K, et al. Iliac and femoral artery stenoses and occlusions: Treatment with intravascular stents. Radiology 1989;172:725-730. 15. Long AL, Page PE, Raynaud AC, et al. Percutaneous iliac artery stent: Angiographic long-term follow-up. Radiology I991;180:771-778.