Author + information
- Received December 16, 2004
- Revision received February 28, 2005
- Accepted March 10, 2005
- Published online August 16, 2005.
- Ricardo A. Costa, MD⁎,
- Gary S. Mintz, MD, FACC⁎,
- Stephane G. Carlier, MD, PhD⁎,⁎ (, )
- Alexandra J. Lansky, MD, FACC⁎,
- Issam Moussa, MD, FACC⁎,
- Kenichi Fujii, MD⁎,
- Hideo Takebayashi, MD⁎,
- Takenori Yasuda, MD⁎,
- Jose R. Costa Jr, MD⁎,
- Yoshihiro Tsuchiya, MD⁎,
- Lisette O. Jensen, MD, PhD†,
- Ecaterina Cristea, MD⁎,
- Roxana Mehran, MD, FACC⁎,
- George D. Dangas, MD, PhD, FACC⁎,
- Sriram Iyer, MD, FACC‡,
- Michael Collins, MD, FACC⁎,
- Edward M. Kreps, MD, FACC⁎,
- Antonio Colombo, MD, FACC§,
- Gregg W. Stone, MD, FACC⁎,
- Martin B. Leon, MD, FACC⁎ and
- Jeffrey W. Moses, MD, FACC⁎
- ↵⁎Reprint requests and correspondence:
Dr. Stephane Carlier, Intravascular Imaging and Physiology, The Cardiovascular Research Foundation, 55 East 59th Street, 5th floor, New York, New York 10022
Objectives We report intravascular ultrasound (IVUS) findings after crush-stenting of bifurcation lesions.
Background Preliminary results with the crush-stent technique are encouraging; however, isolated reports suggest that restenosis at the side branch (SB) ostium continues to be a problem.
Methods Forty patients with bifurcation lesions underwent crush-stenting with the sirolimus-eluting stent. Postintervention IVUS was performed in both branches in 25 lesions and only the main vessel (MV) in 15 lesions; IVUS analysis included five distinct locations: MV proximal stent, crush area, distal stent, SB ostium, and SB distal stent.
Results Overall, the MV minimum stent area was larger than the SB (6.7 ± 1.7 mm2vs. 4.4 ± 1.4 mm2, p < 0.0001, respectively). When only the MV was considered, the minimum stent area was found in the crush area (rather than the proximal or MV distal stent) in 56%. When both the MV and the SB were considered, the minimum stent area was found at the SB ostium in 68%. The MV minimum stent area measured <4 mm2in 8% of lesions and <5 mm2in 20%. For the SB, a minimum stent area <4 mm2was found in 44%, and a minimum stent area <5 mm2in 76%, typically at the ostium. “Incomplete crushing”—incomplete apposition of SB or MV stent struts against the MV wall proximal to the carina—was seen in >60% of non-left main lesions.
Conclusions In the majority of bifurcation lesions treated with the crush technique, the smallest minimum stent area appeared at the SB ostium. This may contribute to a higher restenosis rate at this location.
Percutaneous coronary intervention (PCI) of bifurcation lesions is complex and challenging (1). Regardless of the technique, restenosis rates after bare metal stenting were high (40% to 60%), especially at the ostium of the side branch (SB) (1–3) where lesions frequently present with negative remodeling after PCI and suboptimal angiographic results before PCI (1,4). With the advance of drug-eluting stents, the “crush” technique has been proposed to treat bifurcation lesions because of its predictability, high procedure success rate, and full coverage of the SB ostium (5,6). However, case reports suggest that restenosis at the SB ostium continues to be a problem (7). Intravascular ultrasound (IVUS) studies have shown that stent dimensions are important predictors of restenosis even with drug-eluting stents (8,9). However, there is little data on final stent dimensions in both branches after bifurcation lesion intervention regardless of the approach. For this reason, we report the IVUS findings after crush-stenting of bifurcation lesions. It was our hypothesis that, regardless of the angiographic result, imaging of both branches would reveal inadequate stent expansion in at least one branch.
Between April 2003 and May 2004, 40 patients with a single bifurcation lesion underwent successful stenting using the crush technique and sirolimus-eluting stents in both branches (Cypher, Cordis Corp., Miami Lakes, Florida) and postintervention IVUS analysis. This represents 33% of patients with bifurcation lesions treated with the crush technique during this time period. Intravascular ultrasound was performed in both branches in 25 lesions and just the main vessel (MV) in 15 lesions. In 14 of the 15 patients with just MV IVUS, there was no attempt to image the SB. We primarily report the findings in patients with IVUS of both branches. Written informed consent was obtained before all procedures. Clinical demographics were obtained by hospital chart review.
The crush stent technique has been reported in detail elsewhere (5). In brief, both branches are wired and preferably predilated. The first stent is advanced into the SB, but not expanded, while a second stent is advanced into the MV to cover fully the bifurcation. The SB stent is retracted into the MV so that its proximal edge is 4 to 5 mm proximal to the carina and then expanded. The SB stent delivery balloon and guidewire are removed after a contrast injection ensures that no distal dissection is present and no additional SB stents are needed. The MV stent is then expanded. In 23 of the 25 patients, the SB was rewired and redilated with the kissing balloon technique regardless of the angiographic appearance. Aspirin 325 mg and a 300-mg loading dose of clopidogrel were administered preprocedure unless patients were already pretreated. Glycoprotein IIb/IIIa inhibitors were given per operator discretion, and all patients were prescribed aspirin (81 to 325 mg/day) indefinitely and clopidogrel (75 mg/day) for at least six months.
Cineangiograms were independently analyzed by the Cardiovascular Research Foundation's Angiographic Core Laboratory. Quantitative coronary angiography (QCA) was performed using the CMS-GFT algorithm (MEDIS, Leesburg, Virginia). For the MV, the minimum lumen diameter (MLD) and the mean reference diameter (RD) (average of 5-mm segments proximal and distal to the lesion) were used to calculate the diameter stenosis: [DS = (1 − MLD/RD) × 100]. For the SB, the MLD and the distal RD were used to calculate DS. The MLD at the SB ostium was also specifically measured.
IVUS imaging and analysis
All IVUS studies were performed postintervention after intracoronary administration of 100 to 200 μg nitroglycerin using a commercially available system (Boston Scientific, Natick, Massachusetts). The 40-MHz IVUS catheter was advanced >10 mm beyond the lesion, and an imaging run (using automated transducer pullback at 0.5 mm/s) was performed to a point >10 mm proximal to the lesion; IVUS imaging was recorded only during transducer pullback onto 0.5-inch high-resolution s-VHS videotape for off-line analysis.
Quantitative IVUS analysis was performed using computerized planimetry (TapeMeasure, INDEC Systems, Mountain View, California) by an independent experienced observer blinded to the procedure and clinical data. Measurements included stent and lumen cross-sectional areas (CSA). In the MV the proximal and distal references were analyzed; in the SB only the distal reference was analyzed. The reference segments were the least-diseased image slices (largest lumen with least plaque), but within the same segment. Stent CSA was assessed at five locations (Fig. 1):
MV proximal stent: minimum stent CSA (MSA) between the proximal MV stent edge and the crush area;
MV crush area: MSA within the crush segment;
MV distal stent: MSA between the crush area and the distal edge of the MV stent;
SB ostium: MSA <5 mm distal to the SB stent ostium;
SB distal stent: MSA >5 mm distal to the SB stent ostium.
Stent expansion was defined as MSA divided by distal reference lumen area.
Statistical analyses were performed using StatView 5.0 (SAS Institute, Cary, North Carolina). Data are presented as mean ± 1 SD or frequencies. For comparisons of categorical data, Fisher exact test was used. For comparisons of two continuous variables, a two-tailed unpaired Student ttest was used. Analysis of variance (ANOVA) was used for three subgroup comparison (proximal stent vs. crush area vs. distal stent). Correlation of the QCA and IVUS MLD with regression lines was performed using Pearson correlation coefficient. A p value <0.05 was considered significant.
Baseline clinical characteristics are shown in Table 1.Postintervention IVUS MLDs were smaller than QCA in the MV (2.55 ± 0.38 mm vs. 2.84 ± 0.40 mm, p = 0.001), but not in the SB (2.17 ± 0.34 mm vs. 2.11 ± 0.42 mm, p = 0.58). However, there was only a moderate correlation between IVUS and QCA MLDs at both locations: MV (r = 0.561, p < 0.001) and SB (r = 0.532, p = 0.006) (Fig. 2).
Non-left main artery (LM) bifurcation lesions with IVUS of both vessels
Table 2shows the procedural, QCA, and IVUS data. The SB ostium was predilated in 80%, and final kissing balloon inflations were performed in 90%. By QCA, SB lesions were shorter compared to MV lesions with smaller reference diameters and smaller pre- and postprocedure MLD.
Similar to QCA, IVUS showed smaller SB distal reference measurements compared to the MV. Within the MV, there was no significant difference in stent CSA comparing the proximal stent, crush area, and distal stent (p ANOVA = 0.07). Within the SB there was no difference between the ostial and distal stent CSA (p = 0.6). However, the SB MSA was smaller than the MV (p < 0.0001); stent expansion was significantly less in the SB compared to the MV (p = 0.02); 55% of SB stents had an MSA <4 mm2(vs. 10% in the MV, p = 0.007), and 90% of SB stents had an MSA <5 mm2(vs. 20% in the MV, p = 0.006). Figure 3shows an example of SB stent underexpansion.
“Incomplete crushing”—defined as incomplete apposition of SB or MV stent struts against the MV wall proximal to the carina (Fig. 4)—was seen in 60%. Incomplete crushing was associated with SB underexpansion (77.1 ± 7.6% vs. 89.4 ± 13.1%, p = 0.04), but there were no other differences (Table 3).
The 15 patients with non-LM lesions and just MV IVUS were compared to patients with IVUS of both vessels. There was no difference in QCA MV and SB lesion length (p = 0.63 and p = 0.62), reference diameter (p = 0.11 and p = 0.23), and poststent MLD (p = 0.12 and p = 0.73). There was no difference in IVUS MV stent CSA (p = 0.31) and expansion (p = 0.32) or frequency of incomplete crush (64%).
LM (left anterior descending coronary artery/left circumflex artery) bifurcation lesions with IVUS of both vessels
Table 4shows the procedural, QCA, and IVUS data. The SB ostium (left circumflex artery) was predilated in four of five cases (80%), and final kissing balloon inflations were performed in all cases.
Intravascular ultrasound MV imaging showed a significant difference in stent CSA comparing the proximal stent, crush area, and distal stent (p ANOVA = 0.007); the proximal stent CSA was significantly larger than both the crush stent CSA (p = 0.004) and the distal stent CSA (p = 0.004). In the SB, however, the ostial and distal stent CSA were similar (p = 0.45) as was stent expansion (p = 0.31). Only one MV lesion and one SB lesion had an MSA <5 mm2with none <4.0 mm2. Incomplete crush was not found in these lesions.
Overall, mostly driven by non-LM lesions, the MV MSA (6.7 ± 1.7 mm2) was larger than the SB (4.4 ± 1.4 mm2, p < 0.0001). The results were similar among the different operators. When only the MV was considered, the MSA was found in the crush area (rather than the proximal or distal stent) in 56%. When both the MV and the SB were considered, the MSA was found at the SB ostium in 68%. The MV MSA measured <4 mm2in 8% of lesions and <5 mm2in 20% of lesions. For the SB an MSA <4 mm2was found in 44%, and an MSA <5 mm2in 76%.
Incomplete apposition of MV stent struts to the vessel wall (in addition to incomplete crush of the SB stent struts) was found in five patients: four in the proximal end of the stent and two in the distal end of the stent (one patient had incomplete apposition at both locations). The only patient with subacute stent thrombosis had incomplete crushing. There was no incomplete SB stent apposition. Incomplete crush was associated with smaller MV nominal balloon sizes (3.06 ± 1.07 mm vs. 3.33 ± 0.26 mm, p = 0.67) and smaller summed (MV + SB) nominal balloon sizes (5.50 ± 0.92 mm vs. 6.08 ± 0.37 mm, p = 0.17), but no difference in SB nominal balloon sizes or QCA MLD in the MV or SB.
Postprocedure analysis of bifurcation lesions treated with the crush-stent technique with IVUS imaging of both vessels showed: 1) the MSA was smaller in the SB than the MV; 2) the majority of SB lesions showed stent underexpansion with the smallest MSA typically found at the SB ostium; 3) stent underexpansion detected by IVUS was not suspected angiographically; and 4) incomplete stent apposition in the crush area was common. These findings have implications regarding the mechanism and frequency of restenosis after bifurcation stenting using either bare metal or drug-eluting stents.
Causes of in-stent restenosis
In-stent restenosis can be caused by chronic stent recoil (very rare), chronic stent underexpansion, and/or intimal hyperplasia (7,10). However, serial IVUS studies delineating these mechanisms did not include bifurcation lesions. Therefore, chronic stent recoil cannot be excluded at this location, especially because this location has unusual geometry and (potentially) greater external constrictive forces. Serial IVUS study of both branches at implantation and follow-up would be required to exclude chronic bifurcation stent recoil. Such a study has not been done. In fact, the current report is rare in that it includes 25 of bifurcation stented lesions in which both branches were imaged even at a single point in time, in this case, postimplantation.
In the bare-metal stent era, stenting both branches of a bifurcation increased the risk of restenosis compared to only stenting the MV (1–3). However, imaging of the SB was not performed consistently after stent implantation. Therefore, as shown in the current analysis, the impact of SB stent underexpansion may have been underappreciated, even in patients with good angiographic results of both branches. Thus, angiography appears to have a limited ability to detect underexpanded bifurcation stents.
Ostial SB stent underexpansion may be the dominant mechanism of restenosis with drug-eluting stents because even a small amount of intimal hyperplasia superimposed on significant stent underexpansion can result in restenosis. A few reports of bifurcation lesions treated with two drug-eluting stents still indicate a relatively high incidence of restenosis, and the site of restenosis is systematically at the SB ostium regardless of the technique—T-stenting, modified T-, Y-, or V-stenting (7,11,12). It is not clear whether bifurcation lesions are associated with more intimal hyperplasia compared to nonbifurcations.
In the IVUS substudy of the Sirolimus-Eluting Stent Versus Standard Stents in Patients With Stenosis in Native Coronary Artery (SIRIUS) trial, MSA >5 mm2for the total cohort and MSA >4.5 mm2for vessels <2.8 mm by QCA were thresholds that predicted an “adequate” IVUS lumen at follow-up (9). The positive predictive value of the IVUS stent dimensions was 90%. This was not surprising; once an effective drug (in this case sirolimus) inhibited most of the intimal hyperplasia, the main cause of in-stent restenosis became stent underexpansion. In our group of non-LM bifurcation lesions, MSA <5.0 mm2was found in 80% of lesions (mostly the SB stent ostium), and MSA <4.5 mm2was found in 69% of stents with an mean reference diameter <2.8 mm (again, mostly the SB stent ostium). Thus, our bifurcation lesions frequently had an MSA below the threshold associated with restenosis.
The crush technique was developed to overcome two potential problems with bifurcation stenting: stent underexpansion and incomplete coverage of the SB ostium. In the current study, the MSA at the SB was 3.9 mm2in non-LM lesions and 6.1 mm2in LM lesions. In addition, the MSA was found at the SB ostium in 13 of 20 non-LM lesions and 4 of 5 LM lesions. Regardless of predilation of the SB ostium (80% of cases) and final kissing balloon inflations (all but two cases), IVUS imaging confirmed that the crush technique is still associated with SB stent underexpansion, especially at the SB ostium. Furthermore, the current study found incomplete crushing (i.e., incomplete apposition) of the stents against the MV wall proximal to the carina in the majority of cases. This has been reported by others (13) and may affect drug delivery, thereby contributing to restenosis.
Only 25 of 40 patients had final IVUS in both branches. However, QCA showed no difference between patients with IVUS of both branches versus IVUS of just the main branch. This study includes only patients with a successful procedure (Thrombolysis In Myocardial Infarction flow grade 3 and final DS <30% in both branches and no complications); however, this only highlights the current findings. Predilation plus final kissing balloon inflations were performed in 76% of patients; therefore, it was not possible to assess the impact of these procedural steps.
Conclusions and clinical implications
In the majority of bifurcation lesions treated with the crush technique, the MSA appeared at the SB ostium where the stent CSA is frequently below the cutoff that has been associated with a low rate of restenosis. As a result, stent underexpansion may contribute to the higher reported restenosis rate at this location. In the bare metal stent era, the goal was to optimize the MV intervention and achieve an adequate result in the SB without compromising the MV. In the drug-eluting stent era, it may alsobe necessary to optimize the result in the SB. The fact that the stent-crush technique (even completed with kissing balloon inflation) leaves small ostial SB lumen diameters and frequent poor MV wall apposition (incomplete crushing) fits well with the reported incidence of high SB ostial recurrence. This should be factored into strategies for bifurcation lesions.
- Abbreviations and Acronyms
- cross-sectional area
- diameter stenosis
- intravascular ultrasound
- left main artery
- minimum lumen diameter
- minimum stent area
- main vessel
- percutaneous coronary intervention
- quantitative coronary angiography
- side branch
- Received December 16, 2004.
- Revision received February 28, 2005.
- Accepted March 10, 2005.
- American College of Cardiology Foundation
- Reimers B.,
- Colombo A.,
- Tobis J.
- Yamashita T.,
- Nishida T.,
- Adamian M.G.,
- et al.
- de Feyter P.J.,
- Kay P.,
- Disco C.,
- Serruys P.W.
- Sonoda S.,
- Morino Y.,
- Ako J.,
- et al.
- Colombo A.,
- Moses J.W.,
- Morice M.C.,
- et al.