Author + information
- Received January 15, 1999
- Revision received May 18, 1999
- Accepted June 28, 1999
- Published online November 1, 1999.
- Amadeo Betriu, MD∗,* (, )
- Monica Masotti, MD∗,
- Antoni Serra, MD∗,
- Joaquin Alonso, MD†,
- Francisco Fernández-Avilés, MD†,
- Federico Gimeno, MD†,
- Thierry Colman, MD‡,
- Javier Zueco, MD‡,
- Juan L Delcan, MD§,
- Eulogio Garcı́a, MD§ and
- José Calabuig, MD∥
- ↵*Reprint requests and correspondence:
Dr. Amadeo Betriu, Institut de Malalties Cardiovasculars, Hospital Clı́nic, Villarroel, 170, 08036 Barcelona, Spain
The purpose of this study was to test the hypothesis that stent implantation in de novo coronary artery lesions would result in lower restenosis rates and better long-term clinical outcomes than balloon angioplasty.
Placement of an intracoronary stent, as compared with balloon angioplasty, has proven to reduce the rate of restenosis. However, the long-term clinical benefit of stenting over angioplasty has not been assessed in large randomized trials.
We randomly assigned 452 patients with either stable (129 patients) or unstable (323 patients) angina pectoris to elective stent implantation (229 patients) or standard balloon angioplasty (223 patients). Coronary angiography was performed at baseline, immediately after the procedure and six months later. End points were the rate of restenosis at six months and a composite of death, myocardial infarction (MI) and target vessel revascularization over four years of follow-up.
Procedural success rate was achieved in 84% and 95% (balloon angioplasty vs. stent, respectively). The increase in the minimal luminal diameter was greater in the stent group both after the intervention (2.02 ± 0.6 mm vs. 1.43 ± 0.6 mm in the angioplasty group; p < 0.0001), and at six-month follow-up (1.98 ± 0.7 mm vs. 1.63 ± 0.7 mm; p < 0.001). The corresponding restenosis rates were 22% and 37%, respectively (p < 0.002). After four years, no differences in mortality (2.7% vs. 2.4%) and nonfatal MI (2.2% vs. 2.8%) were found between the stent and the angioplasty groups, respectively. However, the requirement for further revascularization procedures of the target lesions was significantly reduced in the stent group (12% vs. 25% in the angioplasty group; relative risk 0.49, 95% confidence interval 0.32 to 0.75, p = 0.0006); most of the repeat procedures (84%) were carried out within six months of entry into the study.
Patients who received an intracoronary stent showed a lower rate of restenosis than those treated with conventional balloon angioplasty. The benefit of stenting was maintained four years after implantation, as manifested by a significant reduction in the need for repeat revascularization.
Restenosis after successful coronary angioplasty, first recognized nearly two decades ago (1), continues to compromise the long-term efficacy of this intervention (2–9). A large number of clinical trials testing different pharmacological agents have failed to identify a compelling and clinically useful strategy that reduces the incidence of 30% to 40% of patients with recurrence, as assessed by angiography (10–15).
Although two pivotal randomized studies have shown that coronary artery stenting did reduce the risk of restenosis compared with balloon angioplasty (16,17), data regarding the effect of new interventional devices to prevent restenosis are still limited (18–22). In addition, the clinical efficacy of stenting relative to that of conventional angioplasty has not been assessed in the long run. Therefore, we conducted a randomized, multicenter trial (the START study) to test the hypothesis that stent implantation in de novo coronary artery lesions would result in lower restenosis rates and greater long-term clinical benefit than balloon angioplasty.
Participating sites and investigators
The participating hospitals and investigators were selected on the basis of recognized experience with both coronary angioplasty and stent implantation with the Palmaz-Schatz device (Johnson and Johnson Interventional Systems, Warren, New Jersey). The coordinating center and angiographic core laboratory were at the Hospital Clinic, University of Barcelona. All study procedures were performed by the investigators at the five Spanish participating centers listed in the Appendix. The study protocol was approved by the Institutional Review Board at each site.
Patient selection, recruitment and randomization
Eligible patients were those with angina or objective evidence of myocardial ischemia and coronary artery lesions deemed suitable for either balloon angioplasty or stenting with the Palmaz-Schatz device. The angiographic criteria for inclusion were the presence of diseased coronary arteries that had not undergone previous dilation, that had stenoses of at least 70% on visual inspection and a lesion length of 15 mm or less, and that were suitable for dilation with a 3-mm balloon or larger. Patients with multivessel coronary artery disease were eligible and, according to the judgment of the angiographer, more than one lesion per patient could be randomized. Specific angiographic exclusion criteria related to the characteristics of the lesion (involvement of the ostium of a side branch measuring ≥2.5 mm in diameter), the vessel (total occlusion, size <3 mm, heavy calcification or severe tortuosity), or stenosis of the main left coronary artery exceeding 25%.
Patients with acute myocardial infarction (MI) (within one week of the procedure) or cardiogenic shock, but not those with unstable angina, were excluded. Patients presenting with unstable angina were included because they account for the majority of procedures performed in our laboratories. Patients with any condition precluding a full anticoagulation regimen, patients with life-threatening conditions making the follow-up angiography unlikely and those unable to give informed consent were also excluded.
Randomization, stratified according to center, was carried out by means of sealed envelopes in the catheterization laboratory after the initial set of angiograms was obtained. The actual treatment assignments were crosschecked against the computer-generated randomization sequence.
Angiographic protocol and interventional procedures
Pre- and postprocedural angiograms were obtained after the administration of intracoronary nitroglycerin (200 μg) in a minimum of two orthogonal views best documenting the target lesion. The obliquity and angulation of each of these views were carefully recorded for later duplication at the six-month follow-up angiogram. Images of distal empty catheters were also obtained in each view, for calibration purposes.
The cineangiograms were forwarded to the core laboratory for independent, blinded assessment of the initial and follow-up quantitative coronary angiograms. In order to avoid any influence of the technique used or the procedural outcome on the analysis, the angiograms obtained at the end the procedure were analyzed first, and those corresponding to follow-up were analyzed last. Orthogonal views of the target lesions (including those demonstrating the more severe stenosis) were selected from identical radiographic projections from paired initial and follow-up optimal projections. Because of inadequate orthogonal views, 12% of the quantitative angiographic results were based on single-plane data. End-diastolic cine frames from the selected views were digitized with a cine-video converter and analyzed with a computer-assisted edge-detection algorithm (23).
Because of the potential for subacute thrombosis, a rigorous anticoagulation regimen was planned. Before stent implantation, all patients received aspirin (125 mg p.o., q.d.), dipyridamole (100 mg p.o., t.i.d.), a calcium channel antagonist and dextran 40 (100 ml per hour for 3 h). Intraoperatively, 10,000 U of heparin was given at the beginning of the procedure and 5,000 U immediately before stent placement. Dextran-40 at a rate of 50 ml per hour and intravenous nitroglycerin were administered throughout the procedure. After stent placement, a heparin infusion (15 U/kg/h) was started and continued until the therapeutic range on warfarin was titrated to international normalized ratio >2.5. Aspirin, the calcium channel antagonist, dipyridamole and warfarin were continued for two to three months, after which time the latter two were discontinued.
This therapeutic regimen was strictly followed for this first 100 stent procedures. Afterwards, and according to the recommendations of the Data and Safety Monitoring Board, oral ticlopidine (250 mg b.i.d, for six weeks) was started, whereas dextran, dipyridamol and warfarin were suppressed and the heparin infusion was maintained for a maximum of 24 h.
At four years, clinical information was obtained directly by interview in 131 patients (30%), by telephone in 262 (60%) and from the refering physician in the remaining 43 (10%). Exercise testing was performed immediately before the six-month follow-up coronary angiography, unless early restudy had been clinically indicated. The following parameters were assessed: mortality, cause of death, infarction, target vessel revascularization (by percutaneous transluminal coronary angioplasty [PTCA] or coronary artery bypass grafting [CABG]) and cardiac status. A new MI was diagnosed on the basis of new pathologic Q waves according to the Minnesota code (24)or an increase in serum CK/CK-MB levels at least twice the upper limit of normal. Electrocardiograms (ECGs) were neither blinded nor centrally reviewed, and there was no uniform protocol for sampling CK/CK-MB after procedures. To compare the subjective cardiac status after the two treatment strategies, the patients’ symptoms suggestive of heart failure or angina pectoris were classified according to the New York Heart Association (NYHA) or the Canadian Cardiovascular Society (CCS), respectively. Yearly follow-up data were checked for plausibility, completeness, correctness and consistency with the six-month data.
The primary end point of the study was angiographic restenosis defined as more than 50% reduction in luminal diameter six months after a successful intervention. Other angiographic variables assessed included the success rate, the absolute minimal luminal diameter of the target lesion after the procedure and at follow-up and the caliber of the reference vessel segment.
Angiographic success was defined as a reduction in stenosis to 50% or less after the procedure and at follow-up. Procedural success was defined as evidence of angiographic success in the absence of death, acute MI or need of either coronary artery bypass surgery or repeat angioplasty during hospitalization.
The secondary end point of the study was a composite of cardiac death, acute MI and target vessel revascularization during the four years of follow-up. Other events assessed were abrupt vessel closure occurring outside the catheterization room and hemorrhagic complications (cerebrovascular accident, bleeding requiring transfusion or the need for surgical vascular repair).
Sample size calculation, data management and statistical analysis
The sample size was calculated assuming, first, an initial success rate of 90% for each of the two techniques and, second, a 90% compliance with the six-month follow-up angiography. Ascribing a 35% restenosis rate for angioplasty and a 20% for stent, a total of 564 patients would need to be entered assuming a power of 90% and a significance level of 0.05.
All the data were prospectively recorded by the research coordinating center and investigators at each site in a case report form, forwarded to the coordinating center, and verified by range and consistency checks and double data entry, with queries sent back to the sites about any missing or inconsistent data.
Continuous data were summarized as medians with 25th and 75th percentiles, unless otherwise stated. Selected baseline clinical and angiographic outcomes were compared between treatment groups by the chi-square test or Fisher’s exact test in the case of discrete variables and by the Student’s ttest in the case of continuous variables. Relative risk (RR) with 95% confidence intervals (CIs) was used to compare treatments with regard to major clinical outcomes. Kaplan-Meier survival curves were used to characterize the timing of the clinical end points and its components during the four-year follow-up period, and the treatments were compared by the log-rank statistic. All tests of significance were two-tailed, and the groups were compared by the intention-to-treat principle.
A protocol-specified interim analysis was performed when enrollment reached 210 patients, with the data were reviewed by an independent Data and Safety Monitoring Board.
Baseline clinical angiographic characteristics of the patients
Between July 1992 and December 1995, 452 patients were enrolled: 223 were assigned to undergo PTCA, and 229 coronary stenting. Table 1summarizes the baseline clinical and angiographic characteristics of the patient population.
Mean age was 59 years (25th and 75th percentiles: 52 and 66 years) and most patients were male (85% and 87%, angioplasty vs. stent, respectively). The prevalence of risk factors (diabetes mellitus, hypertension, hypercholesterolemia and smoking habits) was similar in the two study groups. The clinical indication for coronary artery revascularization was unstable angina in 69% of patients allocated to angioplasty and 74% in those allocated to stent. The remaining patients entered the study because of stable angina, mainly class II or III of the CCS classification.
The number of diseased vessels (vessels with ≥70% diameter stenosis) was similar in both subsets. The left anterior descending coronary artery was the target vessel in about half of the patients in each group, and the most common morphologies of the attempted lesion, according to American Heart Association/American College of Cardiology classification (25), were type B (82% vs. 89%, angioplasty vs. stent). Finally, the two groups were also well balanced with regard to stenosis severity, as assessed by quantitative coronary angiography, lesion length and left ventricular ejection fraction.
Immediate and six-month follow-up angiographic results
Procedural success (defined as a residual stenosis <50% in the absence of death, acute MI, need for emergency CABG or bail-out stenting) was achieved in 84% of balloon angioplasty patients and 95% of the stenting group (RR 0.30, 95% CI, 0.16 to 0.57; p < 0.001). The rate of vessel closure was not significantly different (4% in the PTCA group and 2.6% in the stent group).
A total of 397 patients had six-month angiograms: 198 in the stent group and 199 in the PTCA group. The luminal dimensions before and after the procedure and at six-month follow-up are given in Table 2. Mean reference and minimal luminal diameters before the procedure were similar in the two groups. The increase in the minimal luminal diameter as a result of the intervention was greater in the stent group (2.02 ± 0.6 vs. 1.43 ± 0.6 in the PTCA group; p < 0.0001). In spite of a greater six-month late loss in the stent cohort (0.87 ± 0.8 vs. 0.64 ± 0.7, p < 0.005), the net gain was higher in this subset (1.19 ± 0.8 vs. 0.84 ± 0.7; p < 0.0001).
The cumulative frequency distributions of the minimal luminal diameters are shown in Figure 1. The preprocedural curves were nearly superimposed, indicating a very similar distribution of minimal luminal diameters among the stent and angioplasty groups. The postprocedure curve of stent lay to the right of the corresponding angioplasty curve as a consequence of the larger minimal luminal diameters obtained after stent implantation. At six-month follow-up, both groups had reduced values, but a significant difference in diameter persisted in favor of the stent group.
Using the binary definition, the rate of restenosis appeared to be 37% in the angioplasty group and 22% in the stent group (RR 0.60; 95% CI 0.43 to 0.82; p = 0.0013).
In-hospital and follow-up clinical outcomes (Table 3)
The in-hospital mortality rates in the PTCA and stenting groups were 1.3% and 0.9%, respectively, and those of acute nonfatal MI were 1.8% and 1.3% (angioplasty vs. stent). Necessity for emergency CABG was identical in the two groups (1.3%). Twenty-five patients of the PTCA group (11%) crossed over to stenting; in 22 of them, the reason for stenting was threatened or abrupt closure secondary to dissection.
Bleeding complications requiring either blood transfusion or surgical repair were more commonly observed among patients allocated to the coronary stent group than among those treated with balloon angioplasty (4.4% vs. 1.3%, respectively). The difference, however, subsided when the analysis was confined to the stent subset treated with the association of ticlopidine and aspirin (1.3%).
Clinical data were available for 436 patients (96%) eligible for six-month follow-up. There were no significant differences between the two groups with regard to death and nonfatal MI, but fewer patients in the stent group underwent a second revascularization of the target lesion: 9.3% versus 19% in the PTCA group (RR 0.49; 95% CI 0.30 to 0.81; p < 0.01 ), representing a 51% reduction in favor of the stent implantation. The event-free survival was 86% in the stent group, compared with 78% in the angioplasty group (p = 0.09). The population of patients who remained free of angina, either spontaneously or as a result of repeat procedures, was higher in the stent group (81% vs. 73%; RR 0.70, 95% CI 0.49 to 0.99; p = 0.04).
At four years (52 ± 17 months), clinical information was available for 436 patients. The number of new events after six months was low in both groups: death, 0.5% versus 0.9%; nonfatal MI 0% versus 0.4% and repeat revascularization 4.3% versus 1.3% (angioplasty vs. stent). Of interest, the per-protocol six-month angiography did not result in any excess of revascularization procedures (3%, between six and seven months).
The mean follow-up period was 52.4 ± 17 months. The difference in long-term follow-up is displayed in the cumulative frequency distribution curves for the clinical end points in both treatment groups in Figure 2. A primary clinical end point occurred in 63 (29.9%) of the patients randomly assigned to PTCA and in 38 (16.9%) of those assigned to stent implantation (RR 0.57, 95% CI 0.40 to 0.81; p = 0.00136). No significant differences were found in the occurrence of death (2.7% vs. 2.4%, stent vs. PTCA) or MI (2.2% vs. 2.8%), but the requirement for a further percutaneous intervention or CABG of the target lesion was significantly reduced in the stent group (12% vs. 24.6% in the PTCA group, RR 0.49, 95% CI 0.32 to 0.75; p = 0.00062). Cumulative rate of angina at four years was 23% in the stent group and 26% in the PTCA group. Most of the patients who develop angina during the follow-up did so within the first six months of the index procedure: 35 of 43 (81%) in the stent group and 45 of 54 (83%) in the PTCA group.
Balloon angioplasty has been applied to increasingly complex clinical and anatomic settings with high success rates. Despite intensive investigation, however, the incidence of restenosis has remained unacceptably high. The clinical consequences and the economic impact of restenosis have prompted the use of adjunctive pharmacological interventions and the development of alternative percutaneous revascularization techniques, including the implantation of coronary artery endoprostheses. The different pharmacological trials aimed at preventing coronary artery restenosis have been disappointing, with a single exception: the antioxidant anticholesterolemic agent probucol (11,14); however, the need for four weeks of pretreatment limits the usefulness of the drug, particularly in patients presenting with unstable angina. The theoretical advantages of stenting over balloon angioplasty vis à vis of restenosis, mainly its ability to serve as scaffold had not been proven in 1992, when the START study was designed (26). At that time large nonrandomized representative series using the Palmaz-Schatz device showed a restenosis rate of 20% (27–30), which compared favorably with that of balloon angioplasty; the potential of subacute coronary occlusion (up to 7%), however, was a cause of concern (27,31). With this background, therefore, we felt, as others did (16,17), that a clinical randomized trial was needed. In distinction to variance with the other trials, however, we considered that extension of follow-up to several years would allow us to assess more accurately the clinical benefit of the stent procedure.
Nature, characteristics and limitations of the trial
The present investigation confirms that, compared with angioplasty, treatment of new focal lesions in large coronary artery vessels by stent placement reduces the rate of restenosis at six months. This effect translates into clinical benefit, as manifested by the higher proportion of patients that remained event-free at the end of the four years of follow-up. Because the trial was underpowered to detect differences in either death, MI or repeat intervention, when the latter was removed from the composite end point, the differences in event-free survival between stent and PTCA were no longer significant. The number of events after six months was low, and thus, no differences could be demonstrated between the two groups if the events occurring within the first six months after the index intervention are removed from the analysis.
The larger minimal luminal diameter (and lesser residual stenosis) in stented vessels accounts for the more favorable angiographic outcome, as illustrated by the cumulative-frequency distribution curves shown in Figure 1. It is important to note that although the greater initial gain was partially offset by a more pronounced deterioration during the follow-up period (indicating that a greater gain is followed by a greater loss), the final result (the “net gain”) was better in the stent-treated patients.
As in any randomized trial, “real life” could not be adequately reproduced due to the selection bias imposed by stringent entry criteria. In fact, because patients were enrolled on the basis of their suitability for either procedure, only lesions in relatively large nontortuous vessels were randomized. In a prospective survey involving 300 consecutive procedures, we found that only 20% of the patients with coronary artery disease that undergo conventional balloon angioplasty daily in our laboratories would have qualified as potential candidates for this trial. As a consequence, one should not extrapolate the results of our study to patients with different anatomic substrata or to other types of stents.
The number of crossovers also deserves comment. Despite that switching to the alternative treatment was discouraged, 11% of our patients assigned to undergo balloon angioplasty received stents. Twenty-three of the 25 patients who were switched to stent placement had balloon-induced dissections and threatening vessel closure, thus being prone to serious clinical events if stent placement had not been available. The trial, therefore, could be viewed as a comparison between two treatment strategies: elective stenting and elective balloon angioplasty plus stenting in case of unsatisfactory results (mostly, as a bailout procedure). However, the bias introduced by stent placement in a number of patients randomized to balloon angioplasty does not change the primary end point of the study, as it may only contribute to narrow a difference that remains significant.
Finally, a comment is mandatory with respect to the associated drug therapies. First, it seems very unlikely that the different pharmacological regimens of the two study groups would have had any influence on the observed differences in the angiographic outcomes, because all clinical studies performed have consistently ruled out any beneficial effect of the anticoagulant therapy on restenosis (32,33). Concerning the change from the intensive systemic anticoagulation therapy to a pure antithrombotic regimen in the stent patients, it is fair to say that although not envisaged in the initial protocol, the decision was taken following the recommendations of the independent Data and Safety Monitoring Board. Pertinent to this, in a post hoc analysis, we failed to show any significant differences in the rate of restenosis of the two stent subpopulations; they differed only in the length of hospital stay and the rate of bleeding complications, which were both lower in patients receiving the combination of aspirin and ticlopidin.
Comparison with previous trials
The relevance of the present investigation could be better appreciated when analyzed in the light of the two previous randomized trials.
As to the clinical features of the study population, it is noteworthy that more than 70% of the patients in our study presented with unstable angina, a diagnosis that was made is less than half of the patients enrolled in the STRESS trial and in no patient in the BENESTENT trial. Unstable angina has been recognized as a risk factor for coronary restenosis after conventional balloon angioplasty and, besides differences in measurement methodologies, it may well explain the higher restenosis rates observed in the conventional angioplasty subsets of the two former trials as compared with the latter trial (42% and 37%, vs. 32%).
With reference to the procedural results, it seems relevant to point out again the higher rate of crossovers in our series (11%), as compared with the two studies of reference (6.9% and 5.1% in the STRESS and BENESTENT trials, respectively). These wide differences could be due to several factors: the will to achieve the best possible result, something that has recently arisen deep-seated concern (34), different approaches in front of a coronary dissection, or, most likely, differences in arterial substratum or the clinical condition of the patient. Here, again, unstable angina may play a major role, as suggested by the correspondence between the crossover gradients and the prevalence of unstable angina in the three studies.
Last, but not least, the long duration of this trial has allowed us to incorporate into the study the more relevant technical advances occurring while it was in progress. The first of these advances refers to the exclusive administration of antiplatelet drugs to prevent acute/subacute stent thrombosis. The second consists of the use of high pressures for stent deployment, to optimize stent-to-vessel wall apposition. Besides, the clinical follow-up data showed that the benefit of elective coronary artery stenting is maintained four years after the procedure and results in a significant reduction in the requirement for further revascularization of the target lesion.
In summary, because of the clinical characteristics of our patients and differences in both procedural outcomes and adjunctive drug therapies, the information provided by the START trial is to be considered complementary to that of the two previous randomized trials.
The results of the present study, in keeping with previous work, reinforce the potential of elective stenting in the treatment of suitable coronary artery lesions. In fact, judicious use of coronary stents appears justified in selected patients with new focal lesions in large target vessels, particularly when the risk for restenosis is high and/or dissection after angioplasty is likely. Despite the reduction in hospital stay, stent placement remains a rather expensive procedure, mainly because of the cost of the device and the adjunctive balloon catheter (35). Although the initial high cost may be offset, at least in part, by a decrease in the need for repeat procedures later on, we have not collected prospective data in costs to assess this trade-off.
Steering Committee: A. Betriu (Study Chairman), M. Masotti, A. Serra (Barcelona); J. L. Delcan, Garcı́a E (Madrid); J. Calabuig (Pamplona); T. Colman, J. Zueco (Santander); J. Alonso, F. Fernández-Avilés (Valladolid).
Data and Safety Monitoring Committee: C. Crexells, F. Navarro-López , G. Sanz.
Coordinating Center: Hospital Clı́nic (Barcelona). A. Serra (Director), T. Martorell, N. Puig.
Angiographic Core Laboratory: Institut de Malalties Cardiovasculars (Barcelona). M. Masotti, L. Almirall.
Participating Centers and Investigators: (values in parentheses denote the number of patients enrolled). Hospital Clı́nic, Barcelona (167): A. Betriu, M. Masotti, A. Serra; Hospital Clı́nico Universitario, Valladolid (129): J. Alonso, F. Fernández-Avilés, F. Gimeno. Hospital Universitario Marqués de Valdecilla, Santander (43): T. Colman, J. Zueco; Hospital General Universitario Gregorio Marañón, Madrid (43): J. L. Delcan, E. Garcı́a; Clı́nica Universitaria, Pamplona (9): J. Calabuig.
☆ This work was supported in part by a grant from Johnson and Johnson International Systems (Warren, New Jersey).
- coronary artery bypass grafting
- Canadian Cardiovascular Society
- confidence interval
- myocardial infarction
- percutaneous transluminal coronary angioplasty
- relative risk
- Received January 15, 1999.
- Revision received May 18, 1999.
- Accepted June 28, 1999.
- American College of Cardiology
- Mata L.A,
- Bosch X,
- David P.R,
- Rapold H.J,
- Corcos T,
- Bourassa M.G
- Califf R.M,
- Fortin D.F,
- Frid D.J,
- et al.
- Beatt K.J,
- Serruys P.W,
- Hugenholtz P.G
- Rensing B.J,
- Hermans W.R.M,
- Deckers J.W,
- de Feyter P.J,
- Tijssen J.G.P,
- Serruys P.W
- Ernst S.M.P.G,
- van der Feltz T.A,
- Bal E.T,
- et al.
- Serruys P.W,
- Luijten H.E,
- Beatt K.J,
- et al.
- REDUCE Trial Group,
- Karsch K.R,
- Preisack M.B,
- Baildon R,
- et al.
- Yokoi H,
- Daida H,
- Kuwabara Y,
- et al.
- Bauters C,
- Meurice T,
- Hamon M,
- McFadden E,
- Lablanche J.M,
- Bertrand M.E
- Lablanche J.M,
- McFadden E.P,
- Meneveau N,
- et al.
- Bittl J.A,
- Sanborn T.A
- BOAT Investigators,
- Baim D.S,
- Cutlip D.E,
- Sharma S.K,
- et al.
- Simontom C.A,
- Leon M.B,
- Baim D.S,
- et al.
- Mancini G.J.B,
- Simon S.B,
- McGillem M.J,
- LeFree M.T,
- Friedman H.Z,
- Vogel R.A
- Blackburn H,
- Keys A,
- Simonson E,
- Rantaharju P,
- Punsar S
- Ryan T.J,
- Faxon D.P,
- Gunnar R.M,
- et al.
- Betriu A,
- Serra A,
- Masotti M,
- et al.
- Shatz R.A,
- Goldberg S,
- Leon M,
- et al.
- Kimura T,
- Nosaka H,
- Yokoi H,
- Iwabuchi M,
- Nobuyoshi M
- Haude M,
- Erbel R,
- Straub U,
- et al.
- Thornton M.A,
- Grüntzig A.R,
- Hollman J,
- et al.
- Urban P,
- Buller N,
- Kox K,
- et al.