Provisional T-Drug-Eluting Stenting Technique for the Treatment of Bifurcation Lesions: Clinical, Myocardial Scintigraphy and (L

The treatment of bifurcation stenoses (BS) by percutaneous coronary intervention (PCI) is challenging due to their lower procedural success and higher restenosis rates compared with nonbifurcation coronary stenoses.^1,2 Although dedicated techniques and new stent designs have been developed for treating BS with bare-metal stents (BMS), the restenosis rates remained high, with up to 25.7% in a large randomized trial performed in the drug-eluting stent (DES) era.³ Indeed, in recent decades, several treatment strategies have been conceived for BS, including contemporary implantation of 2 side-by-side stents (e.g., the “crushing” or the “simultaneous kissing stent” techniques).^4,5 Although encouraging results were reported in terms of immediate technical results and short-term angiographic findings, the best technical solution is still debated.2 Moreover, all studies reported so far have been focusing mostly on true BS.¹ Other forms of bifurcation lesions, however, as classified by Lefevre et al6 may present similar procedural difficulties, but obviously true bifurcation lesions are associated with significantly higher rates of adverse events, in comparison to the other types of BS.^1,2,6
Finally, only clinical and short-term angiographic data have been analyzed to date, but no study has focused on the amount of myocardial ischemia such as that assessed by myocardial stress imaging, with the notable exception of a study employing fractional-flow reserve (FFR) measurements.⁷
The aim of our study was to assess the outcome of a default strategy of a provisional T-stenting approach for all kinds of BS using DES, focusing on mid-term clinical and myocardial scintigraphy endpoints, as well as on late (>9 months) exploratory coronary angiography follow up.

Methods

Patients. From January 2004 to July 2005, we treated 53 consecutive patients, 46 males and 7 females, with BS. Patients with acute myocardial infarction (AMI) complicated by cardiogenic shock, multivessel disease requiring surgical revascularization and patients with diffuse lesions at side branches (SB), or lesions not involving the ostium of, but the mid tract of the SB were excluded from the study. We thus enrolled all patients whose lesions had significant atherosclerotic involvement at or near (≤5 mm distance) a bifurcation branch, with ≥2.0 reference diameter by visual estimate. To further improve characterization of enrolled BS, the location of plaque burden was classified according to Lefevre et al: Type 1 lesions were defined as true BS involving the main branch (MB), proximally and distally, and the ostium of the SB; Type 2 lesions involved only the MB at the bifurcation site and not the ostium of the SB; Type 3 lesions were located in the MB proximally to the bifurcation; Type 4 lesions involved the ostium of each branch of the bifurcation with no proximal disease, whereas in Type 4A and 4B lesions, the ostium of, respectively, the main (4A) or the side branch (4B) was involved.⁶ In addition, for analysis purposes, we separated lesions into two comprehensive groups according to the specific bifurcation type: (a) including Type 1 and 4 lesions (i.e., stenoses involving the ostia of both branches); and (b) other bifurcation lesions.
Procedure. All patients were pretreated with aspirin and a 300 mg loading dose of clopidogrel before the index procedure. During the procedure, patients received intravenous unfractioned heparin (100 IU/kg) to maintain an activated clotting time between 250 and 300 seconds. Administration of glycoprotein IIb/IIIa inhibitors was left to the operator’s discretion. Written, informed consent was obtained from all patients.
A stepwise approach was followed for each kind of lesion: predilatation of the lesion or initial kissing balloon was employed at the operator’s discretion. Next, a DES — either the sirolimus-eluting Cypher™ stent (Cordis Corp., Miami, Florida), or the paclitaxel-eluting Taxus^® stent (Boston Scientific Corp., Natick, Massachusetts) — was placed in the MB, with final kissing balloon performed where >50% stenosis was seen on the SB. Provisional stenting (using another DES) was performed in the SB (T-stenting technique) in cases of persistent >50% diameter stenosis with distal TIMI flow <3 or subocclusive dissection. Finally, after SB stenting, a final kissing balloon was recommended in all patients.
Procedural success was defined as a <30% final diameter stenosis in all the stented lesions and as <50% final diameter stenosis in the SB (in cases of balloon-only dilatation), and the absence of major clinical complications (death, emergency bypass surgery, AMI, or repeat revascularization of the target lesion during the hospital stay). Hemostasis of the artery access site was achieved using the Angio-Seal™ (St. Jude Medical, Inc., St. Paul, Minnesota) for patients who underwent femoral artery puncture, and by simple compression in radial artery puncture cases. Serial determinations of troponin I and creatine kinase were performed every 6 hours during the first 24 hours after the procedure. All patients were kept after discharge on maintenance aspirin therapy, and clopidogrel was administered for >9 months following DES implantation.
In-hospital and follow-up outcome assessment. Clinical follow up was prospectively conducted during the hospital stay and after discharge with office visits for >9 months. Six-month exercise stress/rest single-photon emission computed tomography (20 mCi 99Tc sestamibi) myocardial scintigraphy (MIBI) was performed in patients with exertional angina or in asymptomatic patients, while coronary angiography was performed in cases of unstable angina or documented ischemia by MIBI, or after >9 months in asymptomatic patients.
Reference diameter of the artery, percentage diameter stenosis and minimal luminal diameter were measured offline by quantitative coronary analysis before, after PTCA and at follow up. Specifically, measurements were obtained in 2 segments: the MB (i.e., in the stent and in the margins 5 mm proximal and distal to the main vessel stent edges) and the SB (i.e., the stented or balloon-treated segments and the 5 mm distal to those).
Myocardial scintigraphy. Stress/rest single-photon emission computed tomography (using 740 MBq of 99mTc sestamibi) was performed 6 months after the index procedure and in washout from anti-ischemic drugs. Short-, horizontal long-, and vertical long-axis slices were obtained and interpreted by 2 investigators blinded to the other data, with divergences resolved by consensus. Only vertical and horizontal long-axis slices were considered to define the apex. The left ventricle was divided into 13 segments. Segments tributary of right and circumflex coronary arteries constituted the posterior region; the others (including an isolated apical abnormality) constituted the anterior region. Sestamibi uptake was evaluated with the assistance of bull’s eye images. Segmental uptake was classified using a 4-point scale (normal: 0; equivocal or minimally reduced: 1; moderately reduced: 2; severely reduced: 3). Studies were considered to be positive in the presence of reversible Grade 2–3 defect in >1 segment of the same region.⁸
Endpoints. As relevant endpoints, we assessed the rate of major adverse cardiovascular events (MACE), defined as the composite of cardiac death, nonfatal AMI, target vessel revascularization (TVR) and target lesion revascularization (TLR), as well as their individual components. Myocardial infarction was adjudicated as Q-wave (due to the development of >2 new pathologic Q-waves) or non-Q (an increase in creatine kinase level to >3 times the upper limit of normal range in the absence of Q-waves). TLR was defined as repeat percutaneous or surgical revascularization, driven by symptoms or results of laboratory tests, in the target lesion. In addition, we appraised the 9-month rate of clinical restenosis, defined as either TLR and/or binary angiographic restenosis at clinically- or scintigraphically-mandated angiography. Finally, we formally adjudicated the rate of stent thrombosis defined as any of the following: angiographic demonstration of stent closure or intrastent thrombus, unexplained sudden death, or myocardial infarction in the myocardial region of the target vessel occurring after stent implantation and without concomitant documentation of a patent stent. Stent thrombosis was further distinguished during the intraprocedural, subacute (<30 days) and late (>30 days) periods.
Statistical analysis. Continuous variables are presented as mean ± standard deviation, and were compared with the unpaired Student’s t-test. Categorical variables are presented as n/N (%), with pertinent 95% confidence intervals, and were compared with the Chi-square or Fisher’s exact tests when appropriate. Statistical significance was set at the two-tailed 0.05 level. Computations were performed using SPSS 11.0 software (SPSS, Inc., Chicago, Illinois).

Results

Baseline and procedural data. The study diagram is shown in Figure 1, while baseline demographic and clinical data are shown in Table 1. In particular, 2 patients (3.8%) had 3-vessel, 16 patients (30.2%) had two-vessel disease and 35 patients (66.0%) had single-vessel disease. Table 2 summarizes baseline angiographic characteristics of patients according to the Lefevre classification. Procedural characteristics are provided in Table 3. Specifically, platelet glycoprotein IIb/IIIa inhibitors were administered in 12 of the patients (22.6%). Subanalyses according to the bifurcation group, and showing that 27 (50.9%) had been treated for Type 1 or 4 lesions (Group A) and another 26 (49.1%) subjects for other type of bifurcation lesions (Group B), are reported in Table 4. Finally, SB stenting was performed in 4 (7.5%) patients, while MB stenting with balloon-only dilatation of the SB was done in the remaining 49 (92.5%).
In-hospital and 1-month outcomes. Primary success was obtained in all patients. No patient developed AMI according to the immediate postprocedural serial enzyme determinations. Five patients (9.4%), however, showed nondiagnostic (peak creatine kincase <3 times the upper limit of normal) increases in markers of myocardial damage after the procedure. No case of in-hospital stent thrombosis or access site complications occurred (Table 5). On the other hand, 1 patient suffered stent thrombosis in the main branch 5 days after PCI due to unsupervised withdrawal of clopidogrel treatment. This patient was successfully treated with repeat urgent PCI, developing only non-Q-wave myocardial infarction.
Clinical, scintigraphy and coronary angiography follow up. Clinical follow up was available for 52 patients (98.1%) at a mean of 14 ± 3 months. Clinical events from 1 month to the end of follow up were observed in 5 patients (9.4% [95% confidence interval 0.1–17.4%]). One patient had unstable angina 5 months after PCI, with coronary angiography showing progression of disease at the proximal left anterior descending (LAD) artery, which was successfully treated by PCI in the absence of restenosis at the index mid LAD-diagonal bifurcation lesion. Another 3 patients had recurrent exertional angina, and during exercise scintigraphy showed reversible perfusion defects in the territory of the treated BS in 2 cases, and in a remote site in the third patient. These 3 patients underwent coronary angiography within 7 days from MIBI. The first patient showed no restenosis at the index LAD-diagonal BS, but did show critical progression of atherosclerotic disease at the proximal LAD. The other patient showed proximal edge restenosis in the MB stent of the index left circumflex-obtuse marginal (LCX-OM) BS (and in-stent restenosis in a LAD stenosis treated with a BMS). The remaining patient showed persistently good results BS, but progression of disease due to a critical right coronary artery (RCA) stenosis. These 3 patients underwent successful coronary PCI. Finally, 1 patient with a recent AMI who, at the index admission, had a Type 3 LAD-diagonal BS and critical stenoses of the OM and the posterior descending artery (PDA) and had undergone PCI of the BS, showed signs and symptoms 6 months later of heart failure due adverse left ventricular remodeling, including worsening of initially mild mitral regurgitation that became moderate-to-severe. A MIBI stress test showed a reversible defect in the territory of the BS, and coronary angiography demonstrated good results in the MB stent, but 90% critical restenosis at the ostium of the SB. The patient underwent elective surgical intervention including mitral valve repair and coronary artery bypass grafting to the diagonal, OM and PDA (Table 6).
The other 47 patients remained asymptomatic in the absence of any reversible defect in the territory of treated BS at the 6-month MIBI study. Nonetheless, 26 out of the 47 patients (55.3%) without a control angiogram driven by ischemic signs or symptoms underwent follow-up elective coronary angiography >9 months after PCI (mean 11 ± 3), with only 1 of them having had provisional stenting of the SB. Elective coronary angiography, was performed in these 26 patients, revealing no restenosis in the MB stent and 3 significant restenoses in the SB (all not stented). Interestingly, none of these side branch restenoses resulted in perfusion-reversible defects or clinical manifestations in these patients, and were thus left untreated without apparent subsequent clinical sequelae.
Thus, among the 31 patients who underwent follow-up coronary angiography (1 because of recurrent angina, 4 because of recurrent angina and/or positive stress test, and 26 according to the study protocol as still asymptomatic), MB restenosis occurred in only 1 patient (3.2% [0.6–9.4%]), while SB restenosis was observed in 4 patients (12.9% [5.1–24.9%]). Thus, a total of 5 patients (i.e., a 16.1% rate [8.9–23.3%]) met the endpoint criterion of clinical restenosis.

Discussion

The findings of this study reporting on the clinical, scintigraphic and angiographic results of provisional T-DES implantation in bifurcation coronary lesions suggest that provisional T-stenting provides satisfactory procedural and mid-term results in unselected patients with BS, and that single-stent implantation may be successful in most subjects, as in our cohort provisional SB stenting was necessary only in the few cases of disease involving both ostia (Lefevre Type 1 and 4 lesions). Our data also further support the superior results of DES in patients with BS, as shown by the comparison to historical controls with BMS.^9,10 Moreover, this work encourages the systematic use of scintigraphy as a noninvasive but effective means for follow up on patients after bifurcation stenting. Indeed, MIBI may provide detailed information on the anatomic and functional long-term results after PCI for BS, thus nicely complementing findings from the clinical assessment, and potentially even overwhelming the results of coronary angiography.¹¹ Finally, it is worth noting that isolated side branch restenosis did not strictly imply an abnormal stress MIBI, complementing what has already and more rigorously been established by periprocedural fractional-flow reserve assessment.⁷
Bifurcation stenting in the drug-eluting stent era. Interplay between devices and techniques. Percutaneous treatment of BS is still a challenging procedure, given the lack of a practical and effective ad hoc device.^1,2,12 Interventionalists therefore have to deploy the most effective device using the most effective technique. This interplay between device and technique is one of the most important reasons for the reputation of BS as one of the still unmet challenges of interventional cardiology.¹ However, while some certainty exists that DES are the best devices among those currently available, doubt persists about the best DES deployment technique among the several already described in the literature.^1,2
DES have revolutionized current clinical practice by virtually eliminating restenosis in simple lesions and significantly reducing its rate in the higher-risk ones, and this holds true even in the setting of BS.^3,9,13,14 Indeed, comparisons to historical controls with BMS have largely established the superiority of DES versus the former devices, with rates of binary restenosis in the MB and SB dropping from 11–22% to 5–10% and from 11–31% to 5–14%, respectively.^{3,4,9,14–17} While promising, such results are however still less than optimal, as also testified by the pivotal BIFURCATION study, which included 85 patients randomized to either stenting of both branches or stenting of the MB with provisional stenting of the SB. The 6-month results of this study showed a total restenosis rate of 25.7%, which was not significantly different between the double-stenting (28.0%) and the provisional SB-stenting (18.7%) groups. These results were, however, limited by the high crossover rate, with more lesions treated with 2 stents (N = 63) than with stent/balloon (N = 22). In addition, Colombo et al reported a total of 3 cases of stent thrombosis (3.5%), with a 3-month course of double antiplatelet therapy.³
Several techniques for BS stenting have been described and informally tested in recent years; these include T, standard T, modified T, culotte, crush — with its variations such as the reverse crush and step crush — V-stenting and simultaneous kissing stent, Y and skirt stenting.^{1,2,4,5,18–20} Thorough appraisal and comparison of these strategies is not yet possible, as most of the studies reporting such techniques were limited by nonrandomized design, lack of control groups and/or usage of BMS. To date, only 2 randomized clinical trials have been reported,^3,14 with a third trial ongoing.² Aside from the above-mentioned BIFURCATION study, whose results were confounded by the high rate of crossover between study groups, the trial by Pan et al randomized 91 patients to DES for the MB with provisional SB stenting versus systematic stenting for both branches. This study reported low restenosis rates with the two strategies (2% versus 10% for the MB, and 5% versus 15% for the SB, respectively), and no differences in clinical outcomes. Interestingly, the crossover to the double-stent strategy in this work was very low (2.1%), indicating that a conservative approach might be feasible and safe in such types of BS.¹⁴
These are the premises for our choice of using the provisional T-stenting technique as the default approach for BS, even if the potential advantages of newer strategies, such as simultaneous kissing stent or crush stenting, should not be dismissed. In fact, a recent study compared 57 patients treated with a single DES versus 117 patients treated with 2 DES (with crush stenting used in 67% of the latter cases), and reported similar rates of adverse events and restenosis (4.8% versus 9.6% in the MB, and 4.8% and 13.5% in the SB, respectively).⁹ Nonetheless, on the basis of our results, a provisional DES T-stenting technique appears effective in most cases of BS, and only a few of the Type 1 and 4 BS may need to be treated with SB stenting for suboptimal results of balloon-only inflation.
While some techniques may appear quicker and easier, final kissing balloon is beneficial in most cases and should be performed whenever residual stenosis is present in the MB or SB, especially when 2 stents have been deployed.⁴ Safety is another important issue in PCI of BS. While no adequately controlled comparisons are available, it is plausible that a strategy of electively implanting 2 stents instead of 1 might be associated with a higher rate of stent thrombosis in comparison to a provisional T-stenting technique, thus the importance of final kissing balloon and prolonged (>6 months) double antiplatelet treatment.⁴
Finally, very recent randomized evidence appears to be in agreement with our findings.²¹ In particular, in a study enrolling 413 patients randomly allocated to single stenting (with bailout SB stenting) versus default double stenting, no significant differences in MACE were found at 6-month follow up.²¹
Means to assess the results of bifurcation stenting: Coronary angiography, intravascular ultrasound, fractional-flow reserve and myocardial scintigraphy. Outcomes after PCI for BS have been appraised in several ways. Clinical outcomes, while by definition the most relevant endpoints for patients and physicians, have been thoroughly assessed only in large multicenter studies, with the common conclusions that BS are higher-risk lesions in comparison to nonbifurcation lesions, but in the absence of further mechanistic or technical details. Indeed, issues of limited statistical power and heterogeneity in techniques and devices limit most previous and past studies of bifurcation lesions targeted to clinical outcomes only.
Angiographic analyses at baseline and follow up have contributed the most to the appraisal of early and late outcomes of PCI for BS, and indeed the vast majority of studies dedicated to date to BS have been based mainly on angiographic assessment.^3,4,9,14 Intravascular ultrasound (IVUS) may provide even more precise assessment of restenotic processes, but both IVUS and angiographic analysis lack a strict correlation to clinically relevant ischemia. Moreover, prediction of adverse events cannot easily be made using IVUS or angiographic data, and a simple morphologic analysis is usually insufficient to accurately assess the functional significance of BS.
To overcome these problems, an in vivo functional assay of the clinical significance of a coronary stenosis has been introduced into clinical practice. Indeed, an intriguing study was recently reported by Koo et al on the FFR appraisal of MB and SB sites after PCI for BS.⁷ Specifically, FFR was obtained in jailed SB after stent placement in the MB. The main findings of the study were that no diameter stenosis <75% was functionally significant, and that wide variations in FFR were observed, even in lesions with >75% stenosis.
On the other hand, more recent imaging techniques (positron emission tomography, magnetic resonance imaging and computed tomography) may be used to demonstrate reduced flow, even in relatively small SBs, which may not cause acute coronary syndromes but may nonetheless play a role in determining stable angina.
Unmet challenges in bifurcation stenting. While current devices and techniques have been developed and refined in such ways that PCI of bifurcation coronary disease may appear less problematic today than what appeared a decade ago, unmet challenges are still evident. Most studies have been conducted in patients without long or diffuse disease, where treatment strategies are obviously quite different and the risk of suboptimal results increased. Unprotected left main lesions, while still in a way a type of BS, are also a world of their own. In this case, in fact, restenosis may translate into sudden death or pump failure, and it is uncertain whether a double-stent strategy is more or less effective than a single-stent strategy. Promising results have been preliminarily reported with drug-eluting-dedicated stents for bifurcation lesions (e.g., the CAPPELLA, DEVAXX, PETAL, or TRYTON stents), but dedicated clinical trials have not yet been completed and are eagerly awaited. Study limitations. Drawbacks of the present study are several-fold and include the lack of a control group and randomized allocation to provisional T-stenting technique. In addition, sample size is relatively small (N = 53), and only few (N = 4) patients were treated with 2 stents. Another major caveat of our work is the incomplete angiographic follow up, which may have led to biased estimates in the rate of BS restenosis, even if nonsystematic angiography usually translates into an overestimation of restenosis rates. Thus, our findings should maintain their overall validity and robustness. It is well recognized that only adequately-sized randomized clinical trials can provide definitive answers to clinical questions, and we thus eagerly await the completion of the randomized Coronary bifurcations: Application of the Crushing Technique Using Sirolimus-eluting stents (CACTUS) trial comparing a provisional SB strategy with the crush technique in over 200 patients.²

Conclusions

This study reporting on provisional T-drug-eluting stent implantation in bifurcation coronary lesions suggests the safety and efficacy of this stenting technique in the DES era, and also supports the beneficial impact of default follow-up myocardial scintigraphy, with coronary angiography reserved for selected patients and lesions.

References

1. Colombo A. Bifurcation lesions. Italian Heart J 2005;6:475–488.
   
2. Iakovou I, Ge L, Colombo A. Contemporary stent treatment of coronary bifurcations. J Am Coll Cardiol 2005;46:1146–1155.
   
3. Colombo A, Moses JW, Morice MC, et al. Randomized study to evaluate sirolimus-eluting stents implanted at coronary bifurcation lesions. Circulation 2004;109:1244–1249.
   
4. Ge L, Airoldi F, Iakovou I, et al. Clinical and angiographic outcome after implantation of drug-eluting stents in bifurcation lesions with the crush stent technique. Importance of final kissing balloon post-dilation. J Am Coll Cardiol 2005;46:613–620.
   
5. Sharma SK, Choudhury A, Lee J, et al. Simultaneous kissing stents (SKS) technique for treating bifurcation lesions in medium-to-large size coronary arteries. Am J Cardiol 2004;94:913–917.
   
6. Lefevre T, Louvard Y, Morice MC, et al. Stenting of bifurcation lesions: Classification, treatments, and results. Catheter Cardiovasc Interv 2000;49:274–283.
   
7. Koo B-K, Kang H-J, Youn T-J, et al. Physiologic assessment of jailed side branch lesions using fractional flow reserve. J Am Coll Cardiol 2005;46:633–637.
   
8. Vigna C, Stanislao M, De Rito V, et al. Dipyridamole stress echocardiography vs dipyridamole sestamibi scintigraphy for diagnosing coronary artery disease in left bundle branch block. Chest 2001;120:1534–1539.
   
9. Ge L, Tsagalou E, Iakovou I, et al. In-hospital and nine-month outcome of treatment of coronary bifurcational lesions with sirolimus-eluting stent. Am J Cardiol 2005;95:757–760.
   
10. Hoye A, van Mieghem CA, Ong AT, et al. Percutaneous therapy of bifurcation lesions with drug-eluting stent implantation: The Culotte technique revisited. Int J Cardiovasc Intervent 2005;7:36–40.
11. Galassi AR, Foti R, Azzarelli S, et al. Usefulness of of exercise tomographic myocardial perfusion imaging for detection of restenosis after coronary stent implantation. Am J Cardiol 2000;85:1362–1364.
12. Lefevre T, Ormiston J, Guagliumi G, et al. The FRONTIER Stent Registry. Safety and feasibility of a novel dedicated stent for the treatment of bifurcation coronary artery lesions. J Am Coll Cardiol 2005;46:592–598.
13. Biondi-Zoccai GGL, Agostoni P, Abbate A, et al. Adjusted indirect comparison of intracoronary drug-eluting stents: Evidence from a meta-analysis of randomized bare-metal stent-controlled trials. Int J Cardiol 2005;100:119–123.
14. Pan M, de Lezo JS, Medina A, et al. Rapamycin-eluting stents for the treatment of bifurcated coronary lesions: A randomized comparison of a simple versus complex strategy. Am Heart J 2004;148:857–864.
    
15. Anzuini A, Briguori C, Rosanio S, et al. Immediate and long-term clinical and angiographic results from Wiktor stent treatment for true bifurcation narrowings. Am J Cardiol 2001;88:1246–1250.
16. Yamashita T, Nishida T, Adamian MG, et al. Bifurcation lesions: Two stents versus one stent. Immediate and followup results. J Am Coll Cardiol 2000;35:1145–1151.
17. Daemen J, Lemos PA, Serruys PW. Multi-lesion culotte and crush bifurcation stenting with sirolimus-eluting stents: Long-term angiographic outcome. J Invasive Cardiol 2003;15:653–656.
18. Schampaert E, Fort S, Adelman AG, Schwartz L. The Vstent: A novel technique for coronary bifurcation stenting. Cathet Cardiovasc Diagn 1996;39:320–326.
19. Fort S, Lazzam C, Schwartz L. Coronary “Y” stenting: A technique for angioplasty of bifurcation stenoses. Can J Cardiol 1996;12:678–682.
20. Chevalier B, Glatt B, Royer T, Guyon P. Placement of coronary stents in bifurcation lesions by the “culotte” technique. Am J Cardiol 1998;82:943–949.
21. Steigen TK, Maeng M, Wiseth R, et al. Randomized study on simple versus complex stenting of coronary artery bifurcation lesions. The Nordic Bifurcation Study. Circulation 2006;114:1955–1961.