Author + information
- Received April 4, 2008
- Accepted May 18, 2008
- Published online August 1, 2008.
- Helen C. Routledge, MD, MRCP,
- Marie-Claude Morice, MD, FESC, FACC⁎ (, )
- Thierry Lefèvre, MD, FESC, FSCAI,
- Philippe Garot, MD, FESC,
- Federico De Marco, MD,
- Beatriz Vaquerizo, MD and
- Yves Louvard, MD, FSCAI
- ↵⁎Reprint requests and correspondence:
Dr. Marie-Claude Morice, ICPS Institut Cardiovasculaire, Paris Sud, Institut Hospitalier Jacques Cartier, 6 Avenue du Noyer Lambert, Massy, France.
Objectives Our goal was to determine whether the deployment of drug-eluting stents (DES) in bifurcation lesions, according to a uniform provisional side-branch T-stenting strategy (PTS), is a safe and effective treatment in the immediate and long term.
Background In comparison with simple stenoses, successful percutaneous intervention for coronary bifurcation lesions is limited by a higher incidence of procedural complications and need for repeat revascularization. The ideal strategy to overcome these limitations remains to be demonstrated while recent controversy surrounds the long-term safety of DES in bifurcations.
Methods Consecutive patients treated for bifurcation lesions using DES were studied in a prospective single-center registry. Between 2003 to 2005, 477 procedures were performed. The PTS strategy was used in 92%, with a side-branch stent in 28% and final kissing balloon inflation in 95%.
Results Angiographic success was achieved in 99% with 2.5% in-hospital major adverse cardiac events. The cumulative rate of major adverse cardiac events was 10.7% at 1 year and 13.6% at 2 years, including 6.9% and 8.9% target vessel revascularization. Deviation from the PTS strategy independently predicted 2-year mortality (odds ratio: 5.5 [95% confidence interval: 1.63 to 18.3], p < 0.01). The rate of definite or probable stent thrombosis at 2 years was 2.5% with half of all events occurring before hospital discharge.
Conclusions The PTS strategy for the treatment of bifurcation lesions is applicable to over 90% of patients in the real world. With DES, both safety and efficacy have been demonstrated in the long-term with <10% need for repeat revascularization in the first 2 years and a low incidence of late stent thrombosis.
While rapid advances have been made in the percutaneous treatment of obstructive coronary artery disease, progress in the treatment of those 15% to 20% lesions that involve the bifurcation between 2 branches has been limited by 3 main problems. Procedural success rates are lower, periprocedural myocardial infarction (MI) is more common, and restenosis rates and the need for repeat revascularization remain higher than in the treatment of simple coronary lesions (1–3). Although success rates in the main vessel (MV) improved with the introduction of coronary stenting, the incidence of side branch (SB) occlusion and restenosis remained high (4–7). With bare-metal stents (BMS), it became apparent that a stent deployed in the SB in fact offered no advantage over balloon angioplasty alone (5,7–9).
Encouragingly, revascularization rates seem to be reduced by the use of drug-eluting stents (DES) in bifurcation lesions (10), although this remains an ‘off-label’ indication, and according to initial studies the problem of SB restenosis persists (11). Recent trial data suggest that with DES a simple MV stenting strategy is at least equivalent in the short-term to more complex techniques (12). The longer-term efficacy of this strategy, however, still remains to be demonstrated, as does its applicability to the wider group of patients, lesions, and stents in a nontrial setting. Whether DES are the solution for bifurcation lesions and the optimal manner in which they should be deployed thus both remain in question. Finally, the current controversy concerning the risk of late DES thrombosis raises further questions regarding the safest treatment for patients with bifurcation coronary disease. At present, the longer-term follow-up of a large cohort of ‘real-world’ patients necessary to address this question is lacking.
Our center employs a uniform provisional SB T-stenting strategy (PTS) for treatment of bifurcation lesions, which was developed in the BMS era and is described in Figure 1. The main objective of this study was to determine whether the deployment of DES in bifurcation lesions, according to this strategy, could overcome the limitations described and prove a safe and effective treatment both in the immediate and long term.
Data were prospectively collected in a registry of consecutive patients undergoing percutaneous coronary intervention (PCI) for bifurcation lesions between 2003 and 2005. All patients in whom 1 or more DES was deployed in any bifurcation lesion other than the left main were included. No patient was excluded on the basis of clinical presentation.
The uniform PTS strategy (Fig. 1) begins with wiring of both branches in all cases. Pre-dilation of the MV is performed if needed, and a DES deployed in the MV. Guidewires are exchanged and kissing balloon inflation performed; a stent is then deployed in the SB for significant residual stenosis (>50%) or if the flow is Thrombolysis In Myocardial Infarction (TIMI) flow grade <3, and this is followed by a final kissing balloon inflation. Final decisions regarding technical strategy in each case were made by a consensus of opinion among the interventional team. Either sirolimus-eluting (Cypher, Cordis Corp., Miami Lakes, Florida) or paclitaxel-eluting (Taxus, Boston Scientific Corp., Natick, Massachusetts) stents were deployed at the operators' discretion or according to availability.
Before elective intervention, patients were pre-treated with clopidogrel, or, in the case of urgent procedures, a loading dose of 300 mg was given before catheterization. Intravenous aspirin 250 to 500 mg was given at the start of any angioplasty and intravenous unfractionated heparin to achieve and maintain an activated clotting time of 300 to 350 s. Aspirin 75 mg daily was continued indefinitely with clopidogrel (75 mg or 150 mg for patients >80 kg), for a minimum of 2 months after deployment of a Cypher and 6 months after a Taxus stent. Glycoprotein IIb/IIIa inhibitors were used in fewer than 5% cases at the discretion of the operator.
Indications for intervention were divided into acute ST-segment elevation myocardial infarction (STEMI), non-STEMI or unstable angina, post-MI, stable angina, and silent ischemia. Post-MI included all patients in whom the bifurcation was a non-target lesion documented during primary PCI or a target lesion treated within a month of STEMI.
A bifurcation lesion was defined as any lesion treated involving or adjacent to a SB >2 mm in diameter. Lesions were classified as types 1 to 4 according to the Lefèvre classification (13) and according to the internal angle between the MV and the SB (in the working view), with a Y-shaped lesion having an angle <70° and a T-shaped lesion >70°.
Final treatment strategy was divided in 4 categories: PTS (13), systematic T-stenting with the SB stented first (14), culotte (15), and simultaneous kissing stents (16). The crush technique was not used (11).
Follow-up, end points
A 12-lead electrocardiogram was performed 1 h post-procedure and repeated at least once before discharge. Creatine kinase (CK) was measured at 18 to 24 h post-procedure in all cases and repeated as appropriate if new symptoms occurred. The 6-, 12-, and 24-month clinical follow-up was obtained by clinic visit or telephone interview. Control angiography was clinically driven (in presence of symptoms and/or of stress-induced ischemia). Revascularization was performed in the presence of a significant (>50%) diameter luminal narrowing. Cumulative end points were analyzed at hospital discharge, 6, 12, and 24 months. Details of all deaths were obtained by consulting both the clinical records and attending physician. Major adverse cardiac events (MACE) were defined as any cardiac death, early reintervention, Q- or non–Q-wave MI or target vessel revascularization (TVR). The definition of peri- or post-procedural MI was based on a rise in CK to twice the upper limit of normal. The 23 patients presenting with STEMI or non-STEMI and an elevated baseline CK were exempt from this part of the analysis. During follow-up, a clinical syndrome accompanied by the same rise in enzymes was used to define MI and, if new Q waves were documented, Q-wave MI.
Target lesion revascularization (TLR) was defined as reintervention, either percutaneous or surgical, for in-stent or in-segment restenosis. TVR was any intervention to treat or bypass a lesion in any part of the treated vessel or its branches. Stent thrombosis was defined as definite, probable, or possible and classified according to timing by the Academic Research Consortium criteria (17).
Discrete variables are expressed as percentages, and comparisons between 2 groups were made with chi-square or Fisher exact test. Continuous variables are described as mean ± SD with intergroup comparisons being made with unpaired Student t test. A p value <0.05 was considered statistically significant. Patient and procedural variables that appeared to be associated with the occurrence of adverse events during follow-up, with a p value <0.1 in univariate analysis, were entered into multivariate models using binary logistic regression. In order to further decipher modifiable factors that might influence outcome when treating all patients with bifurcation lesions, procedural variables alone were entered into a second multivariate analysis. All data were analyzed with the use of SPSS 14 software (SPSS Inc., Chicago, Illinois).
In total, between January 2003 and December 2005, 477 consecutive bifurcation procedures were undertaken using DES.
The characteristics of the patients are summarized in Table 1. Almost one-third were patients with diabetes, and over two-thirds had multivessel disease. Left ventricular ejection fraction was >55% in 90% of patients where assessed by left ventriculography (63% cases), although this was less likely to be recorded in acute MI.
The lesions treated are described in Table 2. Two-thirds were located at the left anterior descending-diagonal bifurcation and the majority were type 1 (Medina 1,1,1). Over 86% lesions had >50% SB narrowing. In 5%, the lesion treated was the result of BMS restenosis.
The details of procedures and angiographic results are summarized in Table 3. A PTS strategy was used for 92% of lesions, and a second stent was deployed in the SB in 28% of cases. Final kissing balloon angioplasty was achieved in over 95% of all procedures. Neither patient or lesion characteristics differed significantly between the eventual 1- and 2-stent groups other than a lower proportion of posterior descending-posterolateral artery bifurcation lesions in the 2-stent group (3% vs. 10%, p < 0.02).
In-hospital and follow-up data are summarized in Table 4. Angiographic success, defined as residual MV stenosis <30% and SB stenosis <50% by visual assessment and TIMI flow grade 3 in both branches, was achieved in 98.5%. Recovery to hospital discharge was uncomplicated in 97.5%. Five patients (1%) died in-hospital.
One death was the result of rupture of the diagonal vessel and cardiac tamponade during the procedure. One patient, who initially presented in cardiogenic shock after an out-of-hospital cardiac arrest, died of probable myocardial rupture on day 5 after an initial period of stability. One patient suffered a small vertebral territory cerebral infarct post-procedure; although antiplatelet therapy was not interrupted, the patient subsequently died of progressive cardiac failure on day 8. In 2 cases, death followed reintervention for subacute stent thrombosis at 6 and 7 days, respectively. In total, 1.3% of procedures were followed by emergency reintervention, and in all cases this was the result of acute (3 cases) or subacute (3 cases) thrombotic occlusion. In all but the 2 described in the above text, reintervention was successful and the patients were discharged without further complication. Two patients had CK rise without any associated symptoms or complication bringing the total in-hospital MACE rate to 2.5%.
Complete follow-up data were available for 99% of all patients at 12 months. Results are summarized in Table 4. Total mortality was 3.4% including in-hospital deaths. Five of the 9 additional cardiac deaths in year 1 occurred in patients with symptoms of heart failure documented before the initial procedure, and 4 of these deaths were attributable to probable stent thrombosis. The rate of death was lower (1.9%) in patients with a stable clinical presentation. Two deaths occurred after repeat coronary intervention. There were no Q-wave MIs, but 1 non–Q-wave MI in a patient who died. TLR was undertaken in 18 patients (3.9%) and revascularization in another segment of 1 of the 2 vessels involved in the bifurcation in a further 14 (3.0%), so there was a total TVR of 6.9%. In 3 of these cases, TVR was by bypass surgery. Overall cumulative MACE at 12 months was 10.7%.
2-year data were available in 97% of cases and are summarized in Table 4. Total mortality was 5.4% (total cardiac mortality 3.7%), with 3 additional cardiac deaths occurring in the second year (2 sudden deaths and 1 death after a Q-wave MI in a nontreated territory). Of those surviving, 92% were entirely free of angina. In total, 8.9% of patients had undergone TVR over the 2-year period, including 1 more patient undergoing bypass surgery in the second year. Freedom from MACE at 2 years was 86.3% (Fig. 2).
Predictors of outcome at 2 years
The need for reintervention was not determined by any procedural variable, with no significant relation to the use of 1 or 2 stents or to the type of stent deployed. In multivariate analysis, the only predictors of TVR were type 4b (Medina 0,0,1) bifurcations (odds ratio: 2.98 [95% confidence interval (CI): 1.04 to 8.55, p < 0.05] and MV angiographic success, a negative predictor of TVR (odds ratio: 0.07 [95% CI: 0.05 to 0.11], p < 0.0005).
The risk of death at 2 years was associated with advanced age, diabetes, multivessel disease, and an initial presentation with acute coronary syndrome (Table 5), but not with the vessel involved, the lesion type, or bifurcation angle. Concerning procedural factors, however, the non-use of a PTS strategy was associated with mortality both in-hospital (60% of all deaths vs. 7.2% of all survivors, p < 0.005) and at 2 years (28% vs. 6.4%, p = 0.001) as was the deployment of stents in both branches (44% vs. 26%, p < 0.05). By multivariate analysis (Table 6), the non-use of a PTS strategy was the only independent predictor of death at 2 years (odds ratio: 5.5 [95% CI: 1.6 to 18.3], p < 0.01).
Cases of definite, probable, and possible stent thrombosis are illustrated in Figure 3. At 2 years, the incidence of definite and probable stent thrombosis was 2.5% with one-half of all these events occurring before hospital discharge. The incidence of late stent thrombosis (2 probable cases) was 0.43%. There were no definite or probable cases of very late stent thrombosis but 4 unexplained or sudden deaths between 12 and 24 months, falling into the category of possible very late stent thrombosis. Despite small numbers, diabetes showed a significant association, with 61% of all such events occurring in diabetic patients (p < 0.05 vs. nondiabetic patients).
The main findings from this large, real-world registry are that the provisional SB T-stenting strategy described is a safe and feasible technique, applicable to over 90% of all bifurcation lesions encountered. Combining this technique with DES, we have demonstrated long-term efficacy in over 450 patients with a low rate of reintervention at 2 years.
The primary challenge in bifurcation PCI, to attain procedural success, remains an important determinant of outcome, but as illustrated here this can be overcome in the vast majority of cases by using this PTS strategy. The strategy was associated with an overall rate of in-hospital adverse events of 2.5%. This compares favorably with historical data, where events after bifurcation treatment have occurred at a rate of between 7% and 13% (3,8). A rate of post-procedural non–Q-wave MI of <0.5% in the present study suggests a very low incidence of SB loss. This is unrelated to the use of DES (18) and instead supports the technical aspects of the procedure. We suggest that the systematic wiring of both branches (19) and the use of a final kissing inflation (20–22), irrespective of SB stenting, are essential steps although the relative importance of each cannot be determined from this registry where both were used as default.
Deviation from PTS strategy
In keeping with data from our BMS registry (9), deviation from a PTS strategy was an independent predictor of mortality. This finding, in the present study where all operators conformed where possible to a uniform strategy, reflects lesion characteristics felt by the team to be unfavorable for PTS. A systematic SB stenting approach is more likely to be chosen in cases with long complex SB disease. Interestingly, as for BMS (9), type 4 lesions and longer SB stents were associated with a worse outcome. Although SB lesion length was not included in this analysis, intravascular ultrasound data suggest that significant SB disease is a predictor of MACE (23). The ideal treatment strategy for this small proportion of patients (8%) with lesions deemed unsuitable for PTS revascularization warrants further investigation.
Sustained efficacy of DES in bifurcations
The medium-term efficacy of DES in PCI for bifurcation disease (10) has already been shown to be superior to that of BMS, where 6-month event rates were as high as 27% to 50% (7,8). In this study, the low rate of TVR at 6.9% (TLR 3.9%) by 1 year is sustained at 8.9% at 2 years (TLR 5.2%). Whether this improvement in longer-term efficacy is solely attributable to the DES or again reflects the stenting strategy employed necessitates a comparison with other techniques. According to a registry of 241 bifurcations treated with DES and a systematic 2-stent ‘crush’ technique, 9-month TLR was 9.7% (24), while in 205 patients treated for bifurcation disease using PTS, TLR rates at 24 ± 5 months were 8% (25). Interstudy comparisons of this nature are far from conclusive, but randomized studies comparing the many bifurcation techniques described are, so far, few in number. Two small studies (154 patients in total) have suggested equivalent or lower revascularization rates with sirolimus-eluting stents comparing a provisional to a systematic SB stenting strategy (10,26). Subsequently, the multicenter Nordic study (12) randomized 207 patients with bifurcation lesions to a very simple (1 stent in the MB, treatment of SB only in cases of TIMI flow grade 0 to 1) and 206 to a systematic 2-stent strategy. There were no significant differences in TVR at 6 months; however, the 2-stent strategy was associated with a higher rate of biomarkers of myocardial injury (18% vs. 8%, p < 0.01). In contrast to our data, these periprocedural non–Q-wave MI, which may represent SB occlusion, were not included in the calculation of cumulative MACE, and their significance may not yet be apparent at 6-month follow-up (27,28). The PTS strategy used in our series lies between simple and complex with the aim of avoiding SB occlusion and obtaining a good angiographic result in both branches using 1 stent only when possible. Notably, the use of an SB stent itself in our series did not influence long-term TVR suggesting that the difference in revascularization rates between 1 and 2 DES in bifurcations, at least when T-stenting is used provisionally, may be less important than was the case in the BMS era (7,9).
Long-term safety of DES in bifurcations
It is important to determine whether the reduction in need for revascularization when using DES in bifurcations is offset by a reduction in long-term safety. The results presented here do not suggest that this is the case. Survival free of death or MI in the present study at 12 months was 95.5% (92.7 at 2 years), while in our similar BMS registry concerning over 1,000 patients, the equivalent figure at 9 months was 94.5% (9).
The current predominant safety concern after PCI with DES is the risk of stent thrombosis, and bifurcation treatment has been shown to be a predictor of such events (29–31). Moreover, in the first study to demonstrate DES efficacy in bifurcation lesions (10), the rate of definite and probable stent thrombosis was 4.7% at 6-month follow-up, with all events occurring in patients treated with 2 stents (6.3% vs. 0%). A more recent study showed, using the ‘crush’ technique, a risk of stent thrombosis at 9 months of 4.3% (15). In the present study, the rate of stent thrombosis (definite and probable) at 2 years was 2.5%, and half of all these events occurred before hospital discharge with each patients undergoing immediate reintervention. The rate of definite, probable, and possible stent thrombosis of 3.6% is likely to be an overestimate as it assumes that all cardiac or unexplained deaths in a population of patients with cardiovascular disease over 2 years are the result of stent thrombosis. Here it is worth considering recent results of the ARTS II (Arterial Revascularization Therapies Study part II) post-hoc analysis in which the treatment of bifurcation lesions with sirolimus-eluting stents, using 1 stent in the majority of cases, had no adverse influence on 1-year outcomes of multivessel angioplasty (32).
In this registry, although the need for revascularization was low and over 90% patients were asymptomatic at 2 years, without systematic angiographic follow-up the rate of clinically silent SB restenosis may be underestimated. Comparisons with some historical data are limited by the fact that a few studies have used troponin in some patients as a marker of periprocedural infarction (3). This is a nonrandomized single-center study; thus, neither can conclusions be drawn regarding differences between stents nor can the reproducibility of the PTS technique in other centers guaranteed. A further limitation is that the importance of the final kissing inflation can only be hypothesized. Some investigators have demonstrated that, at least when using suitably adapted devices, good results can be obtained from balloon dilation of the SB alone after MB stenting (33). We remain limited to assumptions made from bench testing models, which demonstrate that a kissing inflation is required in PTS to ensure adequate stent apposition proximal to and at the level of bifurcation (21). Finally, despite being a large series of bifurcation angioplasty, with the small number of deaths, the ability to determine univariate predictors of mortality remains limited.
Using this provisional SB T-stenting strategy for the treatment of bifurcation lesions, the rate of angiographic success is high and the procedure is safe with a low incidence of periprocedural MI and few early complications. This strategy is applicable to over 90% nonselected lesions in the real world. Deployment of DES in bifurcations is effective in the long term, with a rate of reintervention in single figures up to 2 years.
Although the risk of stent thrombosis remains higher in bifurcations than in simple lesions, the additional risk of late DES thrombosis is low with this approach, and this risk does not outweigh the benefit observed. A small subset of patients with lesions unsuitable for PTS have worse immediate- and long-term outcomes, and further work is required to determine the optimal approach to their treatment.
- Abbreviations and Acronyms
- bare-metal stent(s)
- confidence interval
- creatine kinase
- drug-eluting stent(s)
- major adverse cardiac events
- myocardial infarction
- main vessel
- percutaneous coronary intervention
- provisional T-stenting
- side branch
- ST-segment elevation myocardial infarction
- target lesion revascularization
- target vessel revascularization
- Received April 4, 2008.
- Accepted May 18, 2008.
- American College of Cardiology Foundation
- Ellis S.G.,
- Vandormael M.G.,
- Cowley M.J.,
- et al.
- Al Suwaidi J.,
- Berger P.B.,
- Rihal C.S.,
- et al.
- Yamashita T.,
- Nishida T.,
- Adamian M.G.,
- et al.
- Colombo A.,
- Moses J.W.,
- Morice M.C.,
- et al.
- Ge L.,
- Airoldi F.,
- Iakovou I.,
- et al.
- Steigen T.K.,
- Maeng M.,
- Wiseth R.,
- et al.
- Cutlip D.E.,
- Windecker S.,
- Mehran R.,
- et al.,
- Academic Research Consortium
- Hoye A.,
- Iakovou I.,
- Ge L.,
- et al.
- Pan M.,
- Suarez de Lezo J.,
- Medina A.,
- et al.
- Ellis S.G.,
- Chew D.,
- Chan A.,
- et al.
- Bhatt D.L.,
- Topol E.J.
- Kuchulakanti P.K.,
- Chu W.W.,
- Torguson R.,
- et al.
- Machecourt J.,
- Danchin N.,
- Lablanche J.M.,
- et al.
- Tsuchida K.,
- Colombo A.,
- Lefèvre T.,
- et al.