Author + information
- Received May 24, 1996
- Revision received October 15, 1996
- Accepted October 28, 1996
- Published online February 1, 1997.
- Kenneth Rosenfield, MD, FACCA,*,
- Robert Schainfeld, DOA,
- Ann Pieczek, RNA,
- Laura Haley, BSA and
- Jeffrey M Isner, MD, FACCA
- ↵*Dr. Kenneth Rosenfield, St. Elizabeth’s Medical Center, 736 Cambridge Street, Boston, Massachusetts 02135.
Objectives. We sought to determine the basis for restenosis within superficial femoral arteries (SFAs) and hemodialysis conduits treated with balloon-expandable stents.
Background. Use of stents within coronary and peripheral vessels continues to increase exponentially. The mechanism of restenosis within stents placed at various vascular sites is not well understood. In particular, the implications of deploying a balloon-expandable stent in a compressible site are not well understood.
Methods. After the serendipitous detection of stent deformation during intravascular ultrasound (IVUS) examination of a restenosed dialysis fistula, we evaluated a consecutive series of patients with stents placed in compressible vascular sites, including the SFA (six patients) and hemodialysis fistulae (five patients). Clinical, angiographic and IVUS examinations were performed to evaluate mechanisms of restenosis.
Results. Stent compression was identified as the principal cause of restenosis in all dialysis conduits and SFAs. Stent deformity was not reliably identified by angiography; however, IVUS identified compression of two forms: eccentric deformation, implicating two-point compressive force, and complete circumferential encroachment of stent struts around the catheter, suggesting multidirectional compressive force. Despite redilation, secondary restenosis resulting from recurrent compression recurred in most sites.
Conclusions. Restenosis within balloon-expandable endovascular stents may occur as a result of stent compression, a phenomenon readily detected by IVUS, but often not by angiography. These findings have significant implications for the use of balloon-expandable stents within vascular sites subject to extrinsic compression, such as hemodialysis conduits, the adductor canal segment of the SFA and carotid arteries.
(J Am Coll Cardiol 1997;29:328–38)
It is estimated that endovascular stents will be implanted in more than 100,000 patients in the United States alone in 1996, and this number is expected to double by 1997 (). This rapid growth is the result of studies that have demonstrated that endovascular stents may optimize the results of percutaneous revascularization ([2, 3]). Stents have been shown to be particularly effective in counteracting undesirable consequences of balloon dilation—namely, elastic recoil, flow-limiting dissection and abrupt occlusion ([4–8]). The ability to scaffold the disrupted vessel wall using a stent may both salvage a suboptimal result after balloon dilation and preserve long term the acute gain in lumen dimensions. Randomized trials have, in fact, demonstrated that for coronary arteries the use of stents may reduce restenosis ([2, 3]). Reports of nonrandomized trials suggest a similar benefit applies to the aorta and iliac arteries ([8–13]). Not surprisingly, given these encouraging results, attempts have been made to extend the use of stents to other sites of the vascular tree in which percutaneous revascularization is complicated by a high frequency of restenosis. These include aortocoronary saphenous vein grafts (), renal arteries (), femoral arteries ([12, 16–18]) and dialysis conduits ([19–21]). Stents are likewise being employed in clinical trials of carotid artery balloon angioplasty ([22–25]). The incidence of stent restenosis for these newer sites remains incompletely defined. In most cases of restenosis, however, the mechanism of restenosis has been inferred to result from intimal hyperplasia ([12, 14, 18]).
Failure of stents placed in the coronary or iliac arteries to prevent restenosis appears to be the result of neointimal proliferation within the stent ([4, 26–30]). Little consideration has been given to mechanical factors that might contribute to late stent failure.
Between September 1992 and September 1995, we identified the presence of stent compression as the principal cause of restenosis in a series of patients treated with balloon-expandable stents for vascular obstruction involving two unrelated vascular sites, hemodialysis conduits and the adductor canal segment of the superficial femoral artery (SFA). Although unrelated anatomically, these two vascular segments have in common the feature of being susceptible to external compressive forces. This report describes the features of stent compression diagnosed by intravascular ultrasound (IVUS), fluoroscopy and angiography, and the potential implications of these findings for the application of stents in vascular segments subject to unfavorable extrinsic or intrinsic, or both, mechanical forces.
1.1 Stents in hemodialysis conduits.
Between September 1992 and January 1994, five hemodialysis patients undergoing revascularization at St. Elizabeth’s Medical Center in Boston for failing dialysis access sites were treated with one or more Palmaz stents (Johnson & Johnson Interventional Systems). One patient was treated with stents at two separate sites. In all cases, the stents were used on a “compassionate” basis to salvage a suboptimal result after percutaneous transluminal angioplasty (PTA) or directional atherectomy (Devices for Vascular Intervention). Each of the six stent sites was followed up longitudinally to evaluate patency.
1.1.2 Initial stent deployment.
Patients were treated initially with Palmaz balloon-expandable peripheral arterial or biliary stents (Johnson & Johnson Interventional Systems). Standard technique for stent deployment was used (). Two of the five dialysis patients were treated with long-term oral anticoagulation after initial stent deployment. Balloon size for stent deployment was selected to match the diameter, measured by IVUS, of an adjacent vascular segment of normal caliber (“reference site”). After initial deployment, stents were evaluated by IVUS for the presence of underexpansion, defined by either 1) the presence of struts protruding freely into the lumen (nonapposition with the vessel wall), or 2) persistent narrowing at the lesion site relative to the adjacent reference site. Underexpanded stents were redilated with a larger balloon or higher inflation pressure.
1.1.3 Clinical follow-up and detection of restenosis.
Dialysis patients were followed closely after initial stent deployment to determine the efficacy of dialysis and relief of symptoms and physical findings (e.g., arm edema). Dialysis efficacy was assessed at monthly intervals by measuring the urea reduction ratio and venous return pressures while on dialysis. If symptoms returned or dialysis efficacy deteriorated, then angiographic and IVUS examinations were repeated.
1.1.4 Angiographic evaluation.
Contrast angiography at the time of follow-up examination was performed in anteroposterior and oblique projections, attempting to identify irregular stent geometry. Quantitative analysis was performed using previously validated, automated edge-detection software (); percent lumen diameter and cross-sectional area (CSA) narrowing were calculated in comparison to the reference site. Qualitatively, stent sites were examined for nonuniform expansion deformity and stent narrowing.
1.1.5 Intravascular ultrasound examination.
This was performed during initial stent deployment and with each subsequent intervention. The IVUS instrument consisted of a 3.5F monorail catheter (Sonicath, Boston Scientific) housing a single-element, 20-MHz transducer, which was rotated at 1,800 rpm. Images were viewed on an IVUS Sonos console (Hewlett Packard) and transferred to standard VHS videotape for off-line analysis. Minimal (minor) and maximal (major) lumen diameters, as well as lumen CSA, were measured at each lesion. Uniformity of stent expansion was determined at the time of each IVUS examination. A “uniformity index,” defined as the ratio of minor to major lumen diameter, was calculated for each site. A uniformity index of 1.0 indicated symmetric circumferential expansion of the stent into a circular geometric configuration. Eccentric stent compression was arbitrarily determined to be present if the uniformity index was ≤0.6.
1.1.6 Treatment for restenosis.
Restenotic stent sites were redilated and additional Palmaz stents or Wallstents (Schneider USA) were then placed at the discretion of the interventionalist to buttress the original stent or to treat new areas of stenosis, or both. Retreatment for second and third episodes of restenosis was performed as indicated.
1.2 Stents in the SFA.
Between July 1994 and March 1995, patients at St. Elizabeth’s Medical Center presenting with claudication due to stenosis or occlusion of the SFA were enrolled in the Femoral Artery Stenting Trial (FAST) under an Investigational Device Exemption granted to Johnson & Johnson Interventional Systems and approved by the Human Institutional Review Board at St. Elizabeth’s Medical Center. Patients were prospectively randomized to treatment with either conventional balloon angioplasty or PTA with adjunctive Palmaz stent deployment. Subsequently, patients were followed at regular intervals with office visits and noninvasive evaluation. Repeat angiographic and IVUS examinations to evaluate patency of the stent sites were performed 6 months after initial stent deployment. Of the seven patients randomized to treatment with stents who have undergone 6-month follow-up angiographic and IVUS examinations, six were observed to have stent compression; these six constitute the SFA cohort for the current report.
Initial stent deployment in the SFA was carried out using standard technique () using Johnson & Johnson Interventional Systems Palmaz biliary stents, premounted on polyethylene balloons (Meditech). Balloon size was selected based on dimensions of the reference site, as measured by IVUS and angiography. If more than one stent was required to span the lesion, then additional stents were deployed in tandem, overlapping adjacent stents by 1 to 3 mm. After initial deployment, high pressure inflation (12 to 15 atm) was performed with a noncompliant balloon. Subsequently, IVUS was used to confirm the presence of complete and symmetrical stent expansion and adequate strut apposition. Underexpanded stent sites were retreated until deployment was deemed to be satisfactory by IVUS. All patients with SFA stents received anticoagulation to an international normalized ratio of 2.0 to 2.5 for 3 months and were treated with aspirin (325 mg/day) for the duration of the follow-up period.
1.2.2 Clinical, angiographic and IVUS follow-up in patients with SFA stents.
Patients with SFA stents were followed with clinical and noninvasive evaluation at 1, 3 and 6 months. Patients were monitored for recurrence of claudication. Ankle-brachial index (ABI) was measured at rest at 1, 3 and 6 months after stent implantation. Exercise treadmill performance, time to onset of claudication and ABI immediately after graded exercise ([32, 33]) were measured at 1 and 6 months after stent implantation.
Patients underwent follow-up angiographic and IVUS examinations at 6 months after stent implantation. Two or more oblique angiographic views were obtained to evaluate the vascular lumen for patency and the stent geometry for evidence of asymmetry. Automated quantitative angiographic analysis was performed to assess the degree of lumen narrowing.
Intravascular ultrasound examination was systematically performed in SFA stent sites at the time of 6-month angiographic evaluation. Lumen dimensions, percent diameter and CSA narrowing and uniformity index were all measured. Strut apposition and thickness of neointimal lining were assessed qualitatively.
1.2.3 Retreatment of SFA stent sites.
Restenotic sites were redilated in symptomatic patients, after which angiography and IVUS confirmed uniform reexpansion of stents and restoration of lumen dimensions. Patients were monitored thereafter with clinical evaluation.
2.1 Dialysis conduits.
Six vascular sites in five patients with dialysis conduit stenosis or occlusion were treated with Palmaz stents. All six stent sites developed repeated episodes of restenosis, each associated with stent compression, despite the employment of aggressive strategies to maintain patency (Fig. 1and Fig. 2). These strategies included the use of progressively larger balloons or a higher inflation pressure to maximize stent expansion, placement of additional balloon-expandable Palmaz or self-expanding Wallstents and the use of anticoagulation. Table 1summarizes the clinical course of these five dialysis patients, including intervals between episodes of restenosis and the percent angiographic narrowing observed at each episode. At the time of first or second restenosis, five additional Palmaz stents and two self-expanding Wallstents were placed in subclavian vein sites.
2.2 Intravascular ultrasound examination in dialysis conduits: evidence of stent compression.
Intravascular ultrasound measurements of minor and major lumen dimension, CSA and uniformity index for each lesion site are shown in Table 2. The uniformity index immediately after initial stent deployment was 0.89 ± 0.25 (mean ± SEM), indicating nearly concentric geometry and uniform expansion. In contrast, at the initial episode of restenosis, the ratio of minor to major lumen dimension fell to 0.37 ± 0.35, indicating eccentric stent compression. Repeat PTA subsequently restored the circular geometry of the stents (uniformity index 0.90 ± 0.27).
At the time of second restenosis, three of six sites were unable to be imaged by IVUS, because of the inability to pass the catheter through the occluded stent. However, the three sites that were visualized again demonstrated stent compression (uniformity index 0.43 ± 0.44). Repeat PTA, with or without additional stent deployment, again restored a circular configuration.
Eccentric stent deformation, with an associated decrease in lumen CSA, recurred at all lesion sites, despite repeated attempts to restore the desired circular configuration by more aggressive PTA. Moreover, the placement of additional stents within these sites, in an attempt to augment the radial strength of the stents, failed to protect the stents from compression and restenosis.
2.3 Fluoroscopic and angiographic imaging of dialysis conduits.
Angiographic percent lumen diameter narrowing is shown in Table 1. Before intervention percent diameter stenosis was 75 ± 5% (mean ± SEM). After initial stent deployment, lumen diameter was typically greater than that of the reference site. At the first episode of restenosis, lumen narrowing was reduced to 48 ± 5%, but improved to 14 ± 4% after repeat dilation. At the second episode of restenosis, lumen diameter narrowing was 87 ± 4%; for the four vessels that were again able to be redilated, residual narrowing was 4 ± 5%.
Importantly, examination of the stents by fluoroscopy and angiography was unreliable in identifying the presence or absence of sent compression. Radiographic images in the standard anteroposterior projection disclosed stent deformity in only one case; however, even in this case, stent compression was underestimated fluoroscopically, compared with the corresponding IVUS images. In certain instances, oblique views disclosed stent asymmetry, suggesting possible stent compression. In most cases, however, particularly in subclavian vein sites (Fig. 2), even oblique fluoroscopic views failed to show stent deformation.
2.4 Superficial femoral artery sites.
Table 3summarizes the clinical course, results of noninvasive evaluation and percent angiographic stenosis before and after treatment and at 6 months follow-up for the six patients in whom a stent was delivered to the SFA. In all patients lumen caliber post-stent deployment (mean 2.8 ± 0.9 stents/patient) was greater than or equal to that of the reference vessel. Symptoms of claudication (), rest ABI and postexercise ABI improved after initial treatment.
At 6-month angiographic and IVUS follow-up examinations, restenosis at one or more stent sites was documented in all six patients. Patient symptoms, rest ABI or postexercise ABI, or all of these, were also consistent with restenosis.
2.5 Intravascular ultrasound examination of SFA stent sites.
Table 4summarizes the IVUS findings for SFA stents. Percent lumen CSA stenosis was improved from 84 ± 3% before PTA to −2 ± 5% after stent deployment, but returned to 72 ± 4% at the 6-month IVUS examination. In contrast to uniform stent expansion observed after initial deployment, IVUS examination at 6 months demonstrated compression of one or more stents in each of the six patients (Fig. 3). Compression was eccentric (uniformity index ≤0.6) in patients 2, 3, 5 and 6, and concentric in patients 1 and 4. All six sites of SFA stent compression were located within the adductor canal.
Five of the six patients were treated successfully with repeat PTA. No new stents were deployed. Of note, the stents in patients 1 and 5 reexpanded in an eccentric fashion, leaving an ovoid lumen configuration and raising the possibility of a persistent compressive force surrounding the SFA at these sites.
2.6 Radiographic imaging of SFA stent sites: evidence of compression.
The results of quantitative angiographic analysis for SFA stent compression are shown in Table 3. During initial treatment and stent deployment, percent diameter stenosis, using the anteroposterior view, was reduced from 67 ± 3% to −5 ± 3% at lesion sites. The identical sites at 6-month follow-up angiography were narrowed angiographically by 71 ± 5%.
Conventional anteroposterior radiographic imaging of SFA stent sites consistently failed to reveal the presence of stent deformity or compression (Fig. 3). In selected instances, extremely angulated, in particular, lateral, views disclosed eccentric stent deformation.
The cases presented in this report illustrate a specific mechanism—namely, stent compression—as the principal cause for restenosis of stents deployed in dialysis conduits and the SFA. The possibility that stents deployed at certain sites might be liable to compression was a concern from the early days of stent development. In designing his balloon-expandable stent, Palmaz considered the possibility that the circumferential elasticity, or “hoop stress,” of a given vessel may produce enough force to overcome the radial strength of the stent opposing it, resulting in inward collapse of the stent struts (). Hoop stress may vary, depending on vessel diameter, wall thickness, type of vessel (muscular versus elastic), plaque thickness, plaque content, degree of calcification and characteristics of surrounding periadventitial tissues.
Isolated instances of stent compression have been reported previously ([36, 37]). Intrinsic elastic recoil has also been considered to be a factor potentially contributing to stent restenosis ([14, 38, 39]). Bonner et al. () and Kimura et al. (), in serial angiographic studies after stent deployment, observed recoil of ∼15% within the first week after the procedure. Painter et al. (), however, observed no recoil in stented coronary arteries studied by IVUS and suggested that previous reports of lumen loss attributed to stent recoil may actually have been due to stent underexpansion, a feature not detected until IVUS imaging was performed.
Despite early concerns during stent development about the possibility of compression, and notwithstanding isolated case reports, extrinsic compression has not been perceived to be a predominant cause of stent failure or restenosis. Instead, most reports have focused on the contribution of neointimal proliferation within the stents ([4, 26–29]).
3.1 Stent compression in dialysis conduits.
Endovascular stents have been used in an attempt to enhance the long-term patency of dialysis conduits ([19, 36, 42, 43]). Reports regarding the use of balloon-expandable (Palmaz) stents in this setting indicate a high rate of restenosis, but few reports have addressed the responsible mechanism. Elson et al. () alluded to the possibility of stent compromise due to placement in compressible sites. In one of their patients, a stent deployed in the right subclavian vein would not expand fully during initial placement, presumably owing to “extrinsic compression at [the] first costo-clavicular junction.” A second patient developed occlusion of a stent placed in the common iliac vein within two weeks, due to “two-point compressive forces”; the computed tomographic scan depicted an ovoid stent configuration similar to that seen by IVUS in our patients.
Two sites of stent compression in the present series of patients were located in the proximal upper extremity, where the brachial and cephalic veins have little surrounding tissue (Fig. 1). It is likely that the stents within these vessels were deformed by application of external compressive force on the upper arm. Among the four patients with subclavian vein stents, compression likely occurred between the first rib and the clavicle at the costoclavicular junction (Fig. 2). Notably, stent compression has not been identified as a significant problem in coronary, renal or aortoiliac vessels. These vessels may be relatively “protected” from extrinsic compression by their location and surrounding support structures.
3.2 Stent compression in the SFA.
Primary patency rates after conventional PTA for stenotic or occlusive SFA disease are at best as high as 40% to 70% at 3 to 5 years ([44, 45]). Several investigators have demonstrated the feasibility of placing stents in the SFA, in hopes of improving long-term patency ([12, 16–18]). Henry et al. () reported a stent thrombosis rate of 22% and a 6-month restenosis rate of 18.5% for 27 Palmaz stents placed in the adductor canal segment of the SFA. All six of our patients with SFA stent compression had the stent deployed within the adductor canal. Therefore, much as the SFA demonstrates a predilection to develop both primary disease and restenosis after conventional PTA within the adductor canal (), it appears that this poor prognosis extends to stenting in that same location. Presumably, stent failure at this site is a consequence of extrinsic compression from the heavy musculature and fibrous aponeurosis through which the SFA traverses.
3.3 Detection of stent compression: radiographic versus IVUS imaging.
In each of our patients, IVUS provided the decisive diagnosis of stent compression. Intravascular ultrasound imaging after initial stent deployment confirmed the presence of complete apposition of the struts against the endolumen surface and circular lumen/stent geometry, with the initial uniformity index close to unity for both dialysis conduits (0.89 ± 0.25) and SFA sites (0.91 ± 0.28). Despite this, restenosis or reocclusion occurred at least once in all SFA sites and repeatedly in all dialysis patients. Although contrast angiography readily identified the recurrence of lumen narrowing, IVUS proved to be instrumental in identifying stent compression as the mechanism of restenosis. Findings on IVUS implicating this mechanism included an ovoid or “slit-like” configuration of the stent and lumen (uniformity index ≤0.6) and, in several instances, the presence of struts protruding into the lumen, abutting the IVUS catheter. These findings were not immediately apparent from fluoroscopic and angiographic examinations. In fact, in several instances, angiography inaccurately implicated the presence of “conventional” restenosis because of intimal thickening. In these same instances, IVUS depicted struts compressed around the catheter, with a relative paucity of neointimal growth interior to the struts. When stent deformation was demonstrated fluoroscopically, it was only after steep angulated views were obtained to confirm the results of the IVUS examination. The decisive detection of stent compression by IVUS in the face of equivocal radiographic findings thus underscores the utility of this imaging technique to elucidate mechanisms of restenosis in vivo.
3.4 Implications of the current study.
Stents are effective in reducing restenosis, in part because they maximize the initial gain in lumen CSA (). In contrast, when placed in a vascular segment subject to compressible forces, a stent may actually act as a catalyst for further lumen compromise, on a mechanical, rather than proliferative, basis. Indeed, as demonstrated in the present series, restenosis may occur in the absence of any obvious neointimal hyperplasia. The phenomenon of stent compression seen in these patients thus illustrates an extreme form of vascular remodeling (). Experience in this series suggests that the use of balloon-expandable prostheses in vessel sites subject to extrinsic compression or extreme flexion is undesirable. The clinical trial, FAST, in which all SFA stent patients described herein were enrolled, has, in fact, been prematurely terminated because of concerns of the manufacturer and investigators regarding SFA stent compression. Similar concern may extend to other sites, such as the carotid artery ([22–25, 48]), where the consequences of stent compression are potentially devastating. Whether restenosis due to stent compression may be eliminated by the use of self-expanding stents or novel stent designs that incorporate more resistance to compression remains to be determined.
We gratefully acknowledge the expert administrative assistance of Susan Panzica.
☆ This study was supported in part by an Academic Award in Vascular Medicine (HL02824) and Grant HL53354 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland.
- ankle-brachial index
- cross-sectional area
- Femoral Artery Stent Trial
- intravascular ultrasound
- percutaneous transluminal angioplasty
- superficial femoral artery
- Received May 24, 1996.
- Revision received October 15, 1996.
- Accepted October 28, 1996.
- The American College of Cardiology
- Winslow R
- ↵Serruys PW, Strauss BH, van der Giessen WJ. Stenting of coronary arteries: has a modern Pandora\'s box been opened? J Am Coll Cardiol 1991;17:143–54B.
- Roubin GS,
- Cannon AD,
- Agrawal SK
- Becker GJ
- Yadov SS,
- Roubin GS,
- Iyer SS,
- et al.
- Diethrich EB,
- Rodriguez-Lopez J,
- Lopez-Galarza L
- Kuntz RE,
- Gibson CM,
- Nobuyoshi M
- van Beusekom HM,
- van der Giessen WJ,
- van Suylen RJ,
- Bos E,
- Bosman FT,
- Serruys PW
- Strauss BH,
- Umans VA,
- van Suylen RJ,
- et al.
- Ellis SG,
- Savage M,
- Fischman D,
- et al.
- Kimura T,
- Nosaka H,
- Yokoi H,
- Iwabuchi M,
- Nobuyoshi M
- LeFree MT,
- Simon SB,
- Mancini GB,
- Vogel RA
- Hiatt NR,
- Nawaz D,
- Regensteiner JG,
- et al.
- Palmaz JC
- Elson JD,
- Becker GJ,
- Wholey MH,
- Ehrman KO
- Ueda Y,
- Sakai A,
- Misaki N,
- et al.
- Isner JM,
- Rosenfield K,
- Losordo DW,
- et al.
- Bonner RF,
- Keren G,
- Douek PC
- Gunther RW,
- Vorwerk D,
- Bohndorf K,
- et al.
- Isner JM,
- Rosenfield K
- Isner JM
- Yadav SS,
- Roubin GS,
- Iyer SS,
- et al.