Author + information
- Received October 19, 2005
- Revision received January 2, 2006
- Accepted January 16, 2006
- Published online June 20, 2006.
- David J. Clark, MD⁎,†,
- Sara Lessio, MD⁎,1,
- Margaret O’Donoghue, MD⁎,
- Con Tsalamandris, MD†,
- Robert Schainfeld, DO⁎ and
- Kenneth Rosenfield, MD, FACC⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Kenneth Rosenfield, Cardiology Division, Massachusetts General Hospital, 55 Fruit Street, Gray/Bigelow 800, Mailstop 843, Boston, Massachusetts 02114.
Objectives The aim of this study was to determine the mechanisms and predictors of carotid artery restenosis after carotid artery stenting (CAS) using serial intravascular ultrasound (IVUS) imaging.
Background Carotid artery stenting is increasingly used to treat high-grade obstructive carotid disease, but our knowledge of carotid in-stent restenosis and remodeling remains limited.
Methods Post-procedural and 6-month (median 6 months) follow-up quantitative carotid angiography and IVUS were performed after self-expanding stent deployment in 50 internal carotid arteries (ICA). The IVUS measurements at multiple designated sites included minimal luminal diameter, lumen area, stent area (SA), and neointimal hyperplasia area (NIH).
Results Late stent enlargement at follow-up was found at all segments, and the percentage increase was greatest at the ICA lesion site (mean ± SD, 48.9 ± 35.3%). The NIH, expressed as a percentage of SA, was seen within all segments of the stent and was greatest at the ICA lesion site (37.3 ± 23.3%). There was a strong positive correlation between the amount of NIH and late stent enlargement (r = 0.64; p < 0.001). Immediate post-procedural minimum ICA SA (r = −0.37; p < 0.01) and stent expansion (r = −0.44; p = 0.001) correlated negatively with the percentage restenotic area at follow-up.
Conclusions Although self-expanding carotid stents generate considerable neointimal hyperplasia, the process is balanced by marked late stent enlargement. Small stent dimensions immediately post-procedure were associated with a higher risk of restenosis.
Carotid artery stenting (CAS) has proved to be a safe and effective alternative to endarterectomy to treat severe carotid stenosis in patients at high-risk for surgery (1–4). Although the risk of symptomatic restenosis after CAS is very low, up to 10% of patients develop >50% stenosis by angiography or carotid duplex (4–7). In some patient subgroups, such as women and the elderly, it approaches 20% (6).
Serial intravascular ultrasound (IVUS) studies have been fundamental to our understanding of restenosis after balloon-expandable, self-expanding, and drug-eluting stent deployment in coronary arteries (8–13). Self-expanding coronary stents were found to enlarge significantly over time, but also induced greater neointima formation (10–12). Furthermore, for balloon-expandable coronary stents, smaller immediate post-procedural lumen dimensions, as measured by IVUS, have been strong predictors of restenosis (8,9).
Ultrasound follow-up studies after CAS have been limited to externally applied carotid duplex (14,15). Intravascular ultrasound offers a more direct and detailed method to assess the carotid arterial response to CAS. Intravascular ultrasound provides images from within the vessel, has greater resolution (axially 80 and laterally 200 to 250 μm), does not have to penetrate extravascular soft tissues, and can assess the vessel in three dimensions (14,16). The aim of this study was to use serial IVUS imaging to understand the mechanisms and determine predictors of restenosis after CAS in patients undergoing CAS due to high surgical risk for endarterectomy.
Carotid artery stenting (CAS), with adjunctive IVUS imaging (before and after stent deployment in all procedures), was performed in 82 consecutive patients and 87 arteries at a single center between June 1995 and December 2000. Of these, follow-up angiography and IVUS was performed in 58 patients and 61 arteries (71%) at a median of 6 months (range 4 to 12 months) after stent deployment. Of these, 11 arteries were excluded from analysis because: 1) the stent was balloon-expandable (6 arteries); 2) the stented segment did not involve the internal carotid artery (ICA) or distal common carotid artery (CCA) bifurcation (4 arteries with aorto-ostial CCA lesion location); or 3) the IVUS imaging was suboptimal quality (1 artery). The final study group comprised 48 patients and 50 arteries treated with self-expanding stents involving the ICA or CCA/ICA bifurcation.
The baseline characteristics of the 48 patients and 50 arteries are shown in Table 1.These patients were considered high risk for carotid endarterectomy (CEA) and most would have been excluded from the North American Symptomatic Carotid Endarterectomy Trial (NASCET) (17). Almost half had known coronary artery disease. Heavy calcification was frequently present, one-quarter were restenotic after CEA, and a quarter had contralateral carotid occlusion.
The study was approved by the institutional ethics committee of St. Elizabeth’s Medical Center of Boston, and all patients provided written informed consent.
Carotid procedure and IVUS imaging protocol
The CAS and IVUS were performed by one experienced operator, and procedures were undertaken before the routine use of distal protection devices. Procedural data are summarized in Table 1. The Wallstent (Boston Scientific Corp., Natick, Massachusetts) was used in the majority of cases, and the median self-expanding stent diameter was 8 mm. The mean pressure used to post-dilate the ICA stent was approximately 9 atm and the average ratio of post-dilation balloon diameter/distal reference vessel diameter (by IVUS) was 1.16.
Initial IVUS imaging was successfully performed before CAS and immediately after an optimal angiographic result was obtained. In 5 of 50 (10%) of procedures, the findings by IVUS led to further intervention: further post-dilation to distal stent edge (3 arteries) or lesion site ICA (1 artery) and the placement of an additional stent (1 artery). In these cases, IVUS was repeated at the end of the procedure. Follow-up IVUS and selective carotid angiography was performed at a median of six months. Intracarotid nitroglycerine (dose 100 to 200 μg) was administered to all patients before insertion of the IVUS catheter. Heparin was administered to maintain an activated clotting time between 200 s and 250 s. The transducer was positioned in the non-tapered distal segment of the ICA, and then systematically pulled back through the stented portion of the ICA and CCA and into the guiding catheter or sheath.
The IVUS studies were performed using a commercially available system (Boston Scientific Corp.) with 2.6- to 3.5-F monorail catheters and 20- to 40-MHz transducers rotated at 1,800 rpm. Motorized transducer pullback was used at a speed of 0.5 mm/s. Studies were recorded on high-resolution s-VHS tape for off-line analysis.
The in-hospital and 30-day outcomes from our experience with IVUS-guided CAS have been previously reported and are in line with those reported in worldwide registries of CAS (2,3). There were no neurologic events (transient ischemic attacks or stroke) at the time of follow-up angiography and IVUS.
Quantitative angiography and IVUS measurements
Figures 1Ato 1C illustrate the specific sites where measurements were made with both IVUS and quantitative carotid angiography (QCA). These included the lesion site, the distal non-tapered ICA reference, and the proximal CCA reference. Sites within the stent included the original lesion site, the minimum luminal diameter (MLD) of the ICA and the CCA, and reference sites within the ICA and CCA.
Multiple angulated views were taken to define the narrowest lumen diameter before and after stenting and at six months. A metallic washer, 12.75 mm in diameter, was taped to the skin immediately posterior to the point of carotid bifurcation to ensure accurate calibration (Fig. 1). Off-line analysis was performed using digital calipers. The percentage ICA stenosis was defined as: (1 − [MLD/non-tapered portion of the distal carotid reference]) × 100, as originally described by the NASCET collaborators (17). Qualitative assessment included analysis for the presence, location and degree of carotid calcification.
All IVUS measurements and morphologic analysis follow current American College of Cardiology guidelines (16). Analysis was performed solely by an interventional cardiologist, not involved in the stent procedure, as part of a one-year IVUS fellowship (2). At all the designated sites (Fig. 1D), the MLD, lumen area (LA), and stent area (SA) were measured using computer planimetry (Tape Measure Indec Systems, Palo Alto, California). Identification of the lesion site pre and post-procedure and then at follow-up was achieved by measuring its distance from the external carotid artery. Although measurements of the external elastic membrane (EEM) dimensions were performed, acoustic shadowing from the stent and the presence of heavy calcification frequently obscured the EEM border and rendered measurements unreliable; these data were excluded from the analysis. Carotid calcification was classified as superficial or deep. The arc of superficial calcium was graded as subtending 1, 2, 3, or 4 quadrants.
Immediate post-procedure percentage stent expansion by IVUS was defined as (1 − (minimum ICA SA/ICA distal reference]) × 100 and (1 − [minimum CCA SA/CCA reference]) × 100 for the internal and common carotid arteries, respectively.
The following IVUS calculations were made at each of the designated ICA and CCA stent segments. Values were determined in both absolute amounts (mm2) and percentage change (%).
1. A. Neointimal hyperplasia area (NIH) = SA follow-up − LA follow-up (mm2).
B. % NIH = (NIH/SA follow-up) × 100 (%).
2. A. Late stent enlargement at follow-up (Δ stent) = SA follow-up − SA post-procedure (mm2).
B. % Δ stent = (Δ stent/SA post-procedure) × 100 (%).
3. A. Late lumen loss (Δ lumen) = LA follow-up − LA post-procedure (mm2).
B. % Δ lumen = (Δ lumen/LA post-procedure) × 100 (%).
The IVUS, instead of QCA, was used to assess the degree of restenosis. Percentage restenotic area of the ICA stent at follow-up was defined as (1 − [minimum LA ICA/ICA distal reference LA]) × 100.
Statistical analysis was performed using StatView (SAS Institute, Cary, North Carolina). Differences in continuous variables were compared using ttests. Correlations were analyzed using Pearson correlation, except where indicated. Multivariable analysis used analysis of covariance (ANCOVA) to adjust associations for additional variables. A p value of <0.05 was considered to be “statistically significant,” and 95% confidence limits were used. Results are presented as mean ± SD (continuous variables) except where indicated.
Serial IVUS quantitative data
Table 2summarizes the IVUS measurements immediately post-stent deployment and at six-month (median six months) follow-up for the 50 arteries. The minimum post-procedure ICA SA was significantly less than the distal ICA reference (12.4 ± 4.3 mm2vs. 17.9 ± 5.2 mm2; p < 0.001). Post-procedure ICA and CCA stent expansion by IVUS was 70.6 ± 18.7% and 73.0 ± 24.0%, respectively. There was a small but significant (p < 0.05) reduction in both the proximal and distal reference vessel lumen at 6-month follow-up.
Figure 2illustrates the relative contribution of late stent enlargement (Δ stent) and NIH to late lumen loss (Δ lumen) at the defined segments within the carotid stent. The changes are expressed both in absolute amounts (mm2) and as a percentage change relative to the reference vessel (Fig. 2).
There was statistically significant (p < 0.05) enlargement of the stent at all segments of the ICA and CCA at follow-up. The greatest enlargement occurred at the ICA lesion site (p < 0.001 vs. other stent segments) with a 48.9 ± 35.3% increase in SA at follow-up.
Neointimal hyperplasia was also found within all segments of the stent. The greatest amount of neointima, as a percentage of the SA, was at the lesion site of the ICA (37.3 ± 23.3%; p < 0.001 vs. other segments). Figure 3shows that the absolute amount of NIH (mm2) at the ICA lesion site correlated closely (r = 0.64; p < 0.0001) with the degree of late stent enlargement, Δ stent (mm2).
The mean late lumen loss as a percentage of the SA was similar (p = NS) across all sites within the stent (mean ± SEM) (ICA lesion = −6.4 ± 5.7%; ICA reference = −6.3 ± 4.4%, CCA MLD = −11.2 ± 5.9%; and CCA reference = −4.7 ± 4.7%).
Clinical, procedural and angiographic predictors of restenosis
The angiographic ICA MLD by QCA immediately post-procedure correlated negatively with the restenotic area at follow-up (r = −0.31; p = 0.03), but the distal ICA reference did not (r = −0.03; p = 0.85) (Fig. 4).No significant difference (p = NS) in percentage restenotic area was detected among women, symptomatic patients, diabetics, or in arteries with angiographic calcification, contralateral occlusion, or previous CEA (data not shown). There was also no significant correlation (p = NS) between age, maximal balloon pressure, or post-dilation balloon diameter with percentage restenotic area.
IVUS predictors of restenosis
The univariate correlations are shown in Table 3.Both immediate post-procedure minimum ICA SA (r = −0.37; p < 0.01) and ICA stent expansion (r = −0.44; p = 0.001) correlated negatively with percentage restenotic area at 6-month follow-up. Superficial calcium subtending two or more quadrants by IVUS was associated with higher percentage restenotic area (53.4 ± 37.4% vs. 34.5 ± 38.1%; p < 0.05) but was no longer significant after adjustment for either post-procedure minimum ICA SA (p = 0.87) or stent expansion (p = 0.29) by multivariate ANCOVA.
The absolute amount of ICA NIH correlated with percentage restenotic area at follow-up (r = 0.40; p < 0.01). The correlation was strongest when the amount of ICA NIH was expressed as a percentage of the stent area (r = 0.73; p < 0.001). Conversely, there was no relation between percentage ICA late stent enlargement and percentage restenotic area (r = −0.04; p = 0.78).
The present study uses serial IVUS imaging to provide a detailed understanding of carotid artery remodeling and restenosis after self-expanding stent deployment. The major findings were the following: 1) marked late enlargement of the stent at six-month follow-up, with an average 49% increase in stent area at the site of the ICA lesion; 2) 37% average loss of luminal area at the site of the ICA lesion from neointimal proliferation at six months; 3) greater late stent enlargement correlated closely with more neointimal formation; as a result, the amount of NIH determined the restenotic area and the degree of late stent enlargement did not; and 4) post-procedural ICA stent dimensions including stent expansion, MLD, and minimal SA were strong predictors of restenosis, measured by either QCA or IVUS.
Self-expanding stent enlargement, chronic injury, and neointima proliferation
Serial IVUS studies have found that self-expanding stents deployed in coronary arteries continue to enlarge and that the stent area increases between 25% and 33% by six months (10–12). However, when compared to balloon-expandable stents, they also induce up to 30% to 40% more NIH, and the overall late lumen loss between the two stent designs is similar (11,12). In the carotid territory, our findings mirror the findings of these coronary studies; the carotid stent enlarged along its entire length, but it came at the expense of exaggerated neointimal proliferation, translating into a small net lumen loss at six months. There are no other CAS IVUS studies, but Willfort-Ehringer et al. (14), using serial external duplex ultrasound, reported that self-expanding carotid stents enlarged by between 17% and 40%. However, neointimal proliferation was such that they also found negative arterial remodeling prevailed.
Why do self-expanding carotid stents induce such marked neointimal proliferation? The most likely explanation is that the stent (median diameter 8 mm in the present study) is oversized compared with the internal carotid artery and that this chronic expansile force causes ongoing injury to the deep wall, promoting greater neointima (18–20). The strong correlation between the amount of late stent enlargement and neointima in our study supports this. von Birgelen et al. (21) found a similar correlation in the coronary circulation after oversized self-expanding stents were deployed in coronary arteries. Other factors important in coronary stent design, such as strut thickness or the type of metallic alloy, may also affect the amount of carotid neointima (22,23). However, the present study was not large enough to assess for variable neointimal proliferation between different stent designs.
Predictors of restenosis and clinical implications
The final post-procedural stent dimensions measured by either QCA or IVUS are strong predictors of coronary restenosis (8,9). Because the carotid artery is a much larger caliber vessel than the coronary, it has been thought that optimal stent expansion may not be necessary. However, this study has demonstrated that an underexpanded carotid stent with a small final lumen post-procedure has a higher risk of restenosis. In coronary arteries, the mean size of the proximal and distal reference lumen is a predictor of restenosis (8,9). However, we found the distal internal carotid reference lumen was not a determinant of restenosis.
We have previously reported the safety and feasibility of IVUS-guided CAS and found it useful to assess severe superficial calcification (a predictor of stroke in this study), stent apposition, and plaque prolapse (2). However, its routine use to assess post-procedural stent dimensions will remain limited because of a potential small increase in thromboembolic risk and increased procedure time. In the absence of IVUS, on-line QCA could be used as a surrogate to identify those patients with small final stent dimensions at risk of restenosis.
Should efforts be made to make the final post-procedure carotid lumen as large as possible to reduce the risk of restenosis? That is a double-edged sword, because high post-dilation pressures may increase the risk of embolization and perforation and induce greater neointimal formation (11,18,24,25). Progressive carotid stent enlargement is such that if neointima was minimized, restenosis and the need for aggressive post-dilation would be virtually eliminated. The role of self-expanding carotid stents eluting agents, such as sirolimus or paclitaxel, that inhibit neointima (26,27) may merit further investigation.
This study is observational with a modest sample size, although the data was collected prospectively and adjudicated independently. Significant clinical and procedural predictors of restenosis may have been detected with a larger number of patients. Patients were high risk for CEA with advanced vascular disease, and one-quarter were restenotic after previous surgery, so the findings may not be generalized to all patients with carotid disease. The mean follow-up was only six months, and further longer-term vessel remodeling and stent enlargement may occur. If measurement of the EEM had been feasible, remodeling of the entire carotid wall could have been assessed. Volumetric IVUS assessment may have better accuracy (28), although two-dimensional measurements were made at multiple sites.
This serial IVUS study shows that although self-expanding carotid stents generate considerable neointimal hyperplasia, the process is balanced by marked late stent enlargement. Small post-procedural internal carotid stent dimensions were associated with a higher risk of restenosis.
↵1 Dr. Lessio is deceased.
Supported in part by a grant from Boston Scientific Corporation.
- Abbreviations and Acronyms
- carotid artery stenting
- common carotid artery
- carotid endarterectomy
- external elastic membrane
- internal carotid artery
- intravascular ultrasound
- luminal area
- minimal luminal diameter
- neointimal hyperplasia area
- quantitative carotid angiography
- stent area
- Received October 19, 2005.
- Revision received January 2, 2006.
- Accepted January 16, 2006.
- American College of Cardiology Foundation
- Yadav J.S.,
- Roubin G.S.,
- Iyer S.,
- et al.
- Roubin G.S.,
- New G.,
- Iyer S.S.,
- et al.
- Hoffmann R.,
- Mintz G.S.,
- Mehran R.,
- et al.
- Kasaoka S.,
- Tobis J.M.,
- Akiyama T.,
- et al.
- Kobayashi Y.,
- Honda Y.,
- Christie L.G.,
- et al.
- Tanabe K.,
- Serruys P.W.,
- Degertekin M.,
- et al.
- Lessio S.,
- Clark D.J.,
- Schainfeld R.,
- et al.
- Mintz G.S.,
- Nissen S.E.,
- Anderson W.D.,
- et al.
- Sarembock I.J.,
- LaVeau P.J.,
- Sigal S.L.,
- et al.
- Kastrati A.,
- Mehilli J.,
- Dirschinger J.,
- et al.
- Al-Mubarak N.,
- Roubin G.S.,
- Vitek J.J.,
- Iyer S.S.,
- New G.,
- Leon M.B.
- Topol E.J.,
- Yadav J.S.
- Rosenfield K.,
- Losordo D.W.,
- Ramaswamy K.,
- et al.