Randomized Comparison of Elective Stent Implantation and Coronary Balloon Angioplasty Guided by Online Quantitative Angiography and Intracoronary Doppler
Background—The purpose of this study was to compare long-term outcomes of coronary stenting in all lesions (elective stenting) or only in lesions with inadequate morphological and functional results after balloon angioplasty (guided PTCA).
Methods and Results—Treatment of multivessel disease, with any lesion length and vessel size, was allowed provided that all lesions were suitable for stent implantation. Patients were randomized to elective stent implantation (n=370) or guided PTCA (n=365). An optimal PTCA result (residual diameter stenosis ≤35%, coronary flow reserve measured with a Doppler guidewire >2.0, absence of threatening dissections) was achieved in 166 lesions (43%). The remaining 218 lesions underwent stent implantation (provisional stenting). Final residual diameter stenosis was lower in the elective and provisional stent groups (9.3% and 10.2%) than in the optimal PTCA group (24.8%, P<0.00001). On an intention-to-treat analysis, the probability of ≥1 major adverse cardiac event at 12 months was 17.8% in the elective stenting group and 18.9% in the guided PTCA group (20.1% for optimal PTCA and 18.0% for the provisional stenting subgroup, P=NS). The incidence of repeat target lesion revascularization at 1 year was 14.9% in the elective stent group and 15.6% in the guided PTCA group (17.6% for optimal PTCA and 14.1% for the provisional stenting subgroup, P=NS).
Conclusions—When balloon angioplasty is guided by online quantitative angiography and Doppler-derived coronary flow reserve, with provisional stenting reserved for suboptimal results, early and late clinical outcomes are comparable to those achieved by elective stenting of all patients.
For almost 20 years after its introduction by Andreas Grüntzig in 1977, percutaneous transluminal balloon angioplasty (PTCA) had been the most widely applied percutaneous intervention for treatment of coronary artery disease. In the past few years, however, coronary stenting has largely replaced balloon angioplasty, because this technique has a higher immediate success rate and a lower restenosis rate.1 2 3
Intracoronary ultrasound has demonstrated that many cases of recurrent symptoms after “successful” PTCA are the result of an angiographic overestimation of an insufficient initial lumen enlargement.4 5 Intracoronary ultrasound, however, requires the separate insertion of a new catheter and is not always able to determine whether the complex neolumen created by the dilatation is sufficient to normalize coronary blood flow.
Intracoronary Doppler ultrasound can assess the functional severity of coronary stenoses and can easily be integrated into a standard interventional procedure with the use of Doppler transducers mounted on the tip of an angioplasty guidewire.6 7 A recent large prospective study has shown that distal coronary flow velocity reserve (CFR) after PTCA is predictive of short- and long-term recurrence of symptoms and need for repeat revascularization.8
The aim of this randomized multicenter study was the comparison of the long-term clinical outcome of 2 treatment strategies: stent implantation applied to all lesions (elective) or limited to those lesions with inadequate angiographic and physiological results after PTCA as assessed by online quantitative coronary angiography (QCA) and Doppler CFR (guided PTCA with provisional stenting).
Because the goal of the study was to evaluate the impact of this strategy on the everyday patient population of a catheterization laboratory, the only inclusion criterion was the suitability of all the lesions under treatment for stent implantation. Exclusion criteria were limited to chronic total occlusions, graft and ostial stenoses, second restenosis after PTCA, stent restenosis, elective planned Rotablator or directional atherectomy, recent (<24 hours) myocardial infarction, or previous Q-wave myocardial infarction with akinesia or diskinesia in the territory supplied by the target vessel. The study protocol was approved by the Institutional Review Board of all the 55 participating centers (see Appendix), and all patients gave written informed consent.
Angioplasty Procedure and Doppler Measurements
The preintervention angiogram was repeated after the injection of nitroglycerin 0.1 to 0.3 mg or isosorbide dinitrate 1 to 3 mg in all patients considered eligible for the study. If lesion characteristics suitable for stenting were confirmed, the randomization envelope was opened, and the patient was assigned to the strategy of elective stenting or Doppler-guided balloon angioplasty (Figure⇓). In the PTCA group, the use of a balloon diameter matching the reference vessel diameter and inflated to high pressure (10 to 12 atm) was recommended. The PTCA result was accepted only if all of the following 3 criteria were met: online QCA measurement of residual diameter stenosis in the worst view <35%; absence of National Heart, Blood, and Lung Institute (NHLBI) classification type C through F dissections or of any dissection inducing flow reduction or thought to be at risk of occlusion; and CFR distal to the stenosis >2.0. If ≥1 of these criteria were not met, the investigator could choose between the use of bigger balloons and higher inflation pressures, or if these attempts were not feasible or were unsuccessful, crossing over to stent implantation (provisional stenting).
In all patients, angiograms suitable for QCA analysis were repeated in the same projections after treatment. In a subgroup of patients, a final Doppler assessment was performed in the treated artery as well after elective and provisional stent implantation and, whenever present, in an angiographically normal artery (<30% diameter stenosis).
Coronary Flow Velocity Reserve
The tip of a 0.014-in Doppler guidewire (FloWire, Endosonics) was advanced ≥2 cm distal to the stenosis, and the Doppler signal was optimized by gentle rotation, advancement, or withdrawal until a stable flow velocity envelope was obtained. After the basal velocity had been recorded, 12 or 24 μg of adenosine was injected as a bolus via the guiding catheter into the right or left coronary artery, respectively. An automatic detection system was used to search for the peak hyperemic velocity and to calculate and display CFR. Duplicate CFR measurements were performed, and the average of the 2 measurements was used. Flow obstruction due to the guiding catheter being engaged in the coronary ostium was excluded by monitoring of the pressure waveform during hyperemia (pressure damping).
Patients were discharged according to local clinical practice. Two ECGs (after the procedure and before discharge) and ≥1 measurement of total serum creatine kinase level between 6 and 18 hours after the procedure were performed to rule out the presence of Q-wave or non–Q-wave myocardial infarction. The concentration of myocardial isoenzymes was also assessed in case of an increase of the total creatine kinase activity. The protocol required that all patients were treated with aspirin (100 to 325 mg OD) and ticlopidine (250 mg BID) for 3 days before the procedure. Aspirin was maintained in all patients indefinitely. Ticlopidine was continued for 1 month only in the patients who received a stent. Intravenous inhibitors of the IIb/IIIa platelet glycoprotein receptors became available in most participating centers during the study and were used at the investigators’ discretion (4% of patients).
The numbers of guiding catheters, guidewires, angioplasty balloons and stents, the type and amount of contrast medium, and procedure duration (from femoral puncture to last angiogram) were recorded.
The 6-month control included patient history, physical examination, ECG, and noninvasive stress testing. In case of recurrence of angina or development of ischemia during the stress test, a repeat angiogram was recommended, and in case of restenosis, repeat percutaneous revascularization or bypass surgery was performed, as appropriate.
Patients or families were contacted at 1-year follow-up to document occurrence of death, myocardial infarction, repeat revascularization, or recurrent symptoms.
Definition of End Points
The primary end point of the study was the development of ≥1 lesion-related major adverse cardiac events at 12 months, defined as death, myocardial infarction, or repeat target lesion revascularization. Death was considered to be cardiac and related to the treated lesions unless a different cause could be demonstrated. Myocardial infarction was defined as the development of new pathological Q waves in the territory of distribution of the treated artery or an increase in total creatine kinase levels to twice the normal activity for each participating hospital with a concomitant increase in the creatine kinase MB component. Repeat revascularization of the target lesion was defined as a new percutaneous treatment or bypass graft operation because of occlusion or restenosis at the site of 1 of the initially treated lesions or within 5 mm at each extremity.
All events were reviewed by an independent Critical Events Committee (see Appendix), and the final event adjudication was based on specific queries to the investigators and review of source documents.
Clinical and procedural data were entered by the investigators in a dedicated case record form, and prints of the online QCA analysis and of the Doppler measurements were included.
The Doppler prints were examined by a Doppler Core Laboratory (see Appendix), and in case of insufficient quality, the video recordings of the Doppler examination were reevaluated. Only data approved by this laboratory were considered valid for the final analysis.
All procedural angiograms were analyzed independently in a core laboratory using a quantitative angiographic system (CMS MEDIS version 4.0) validated in in vivo experimental models as previously described.9 The guiding catheter was used as scaling device. The analysis was performed in the same working projection as used by the investigators for online analysis both before treatment and after PTCA/stenting.
All analyses were performed on an intention-to-treat basis.
Ordinal and nominal variables were analyzed by χ2, odds ratio, and correspondence analysis technique. The comparative study between groups, for the presence of major adverse cardiac events as a nominal response variable, was performed with univariate and multivariate linear logistic models.
All statistical tests are 2-tailed and assume as significant level probabilistic values <0.05.
Patient and Lesion Characteristics
Between September 1996 and March 1998, 750 patients were enrolled in the study in 28 centers in the United States and 27 centers in Europe, Canada, Japan, Australia, and South Korea. The inclusion criteria were not met in 7 patients (0.9%), and no follow-up was obtained in 8 patients (1.1%). Tables 1⇓ and 2⇓ report clinical data, quantitative and qualitative lesion characteristics, and basal CFR measurements of the remaining 735 patients, who were randomly assigned to elective stent implantation (n=370) or to balloon angioplasty guided by QCA and Doppler (n=365). The high incidence of previous myocardial infarction, unstable angina, and multivessel disease and the complex lesion characteristics reflect the broad inclusion criteria and the “real-world” nature of the study population.
The Figure⇑ summarizes the procedural results on a per-lesion basis: of the 384 lesions randomized to guided PTCA, the procedural end points could be met only in a minority of lesions (166, 43%). A Doppler CFR ≤2.0 was the reason to switch to provisional stenting in 62% of the cases with suboptimal PTCA results (135 of 218 lesions), with 44% of them (59 lesions) associated with persistent >30% diameter stenoses and/or threatening dissections. Table 3⇓ reports procedural variables and angiographic and CFR results in the 3 groups of optimal PTCA, provisional stenting, and elective stenting.
Table 4⇓ shows all 4 major adverse events, divided according to the time of occurrence at follow-up. No significant differences were for any of the study end points. In particular, the optimal PTCA group had a low incidence of 1-month and 12-month revascularization (1.9% and 17.6%), suggesting that a normal CFR and low residual stenosis after PTCA almost eliminate the risk of abrupt closure and achieve a long-term clinical outcome similar to that with stent implantation.
The incidence of target lesion revascularization was compared in various subgroups of the 2 randomization arms, divided according to clinical and angiographic characteristics, and the similar outcomes of the 2 treatment strategies were confirmed in all subgroups. In small vessels (reference lumen diameter ≤2.75 mm), target lesion revascularization was lower after guided PTCA (16.8%) than after elective stenting (25.6%), but the difference was not statistically significant (P=0.13).
Resource Utilization and Cost Analysis
Similar procedural times were observed in the elective stenting group (104±53 minutes) and in the guided PTCA group (106±50 minutes, P=NS). The only significant difference in resource consumption was a lower number of balloons and stents in the guided PTCA group than in the elective stenting group (0.705 stents per lesion in the guided PTCA group and 1.243 in the elective stenting group, P<0.005).
Differences From Previous Studies
The results of this study are at variance with the conclusions of previous randomized trials comparing balloon angioplasty and stent implantation, showing improved immediate and long-term outcome with stent implantation. Fundamental differences in enrollment criteria and trial design may explain this discrepancy. In the Benestent and STRESS trials,1 2 3 only stable patients with good left ventricular function and no previous myocardial infarction and lesions well suitable for implantation of a single 15- or 18-mm Palmaz-Schatz stent were selected. When used in the “real world” of angioplasty, as in DESTINI, with a majority of lesions >10 mm long and type B2 and C American Heart Association/American College of Cardiology classification and 47.4% of lesions located in vessels with a reference diameter ≤3.0 mm, the stent loses some of its glitter.
The second and more important difference is the liberal use in DESTINI of stent implantation, which was performed in the PTCA arm whenever the angiographic result or CFR measurement was suboptimal. In Benestent II,3 the largest and most recent of the randomized PTCA versus stent trials, crossover to stents was permitted for patients with a severe flow-limiting dissection that did not respond to repeat prolonged balloon dilatation. The difference between the 13% (55/413) bail-out stenting of Benestent II and the 57% (218/384) provisional stenting of DESTINI also explains the better final angiographic results of the PTCA arm of DESTINI despite the more complex lesion characteristics (final minimal lumen diameter of 2.29±0.52 mm in DESTINI versus 2.13±0.39 mm in the Benestent II study).
A limitation of the technique of provisional stenting used in this study is the fact that the majority of lesions in the group randomized to guided PTCA actually required stent placement. Because this limitation hampers the practical advantages of using this strategy, one should consider how balloon dilatation can be made more effective in the achievement of the predetermined angiographic and physiological criteria. Quantitative angiography is a simple, rapid, and relatively inexpensive method to compare the lumen dimensions achieved after PTCA at the lesion site and in the reference segment, but it leads to systematic undersizing of the dilatation balloon because of the atherosclerotic involvement of the reference segment. Balloons matching the media-to-media diameter have been shown to be safe and effective in improving the results of balloon dilatation,10 with ≥80% of lesions treated with ultrasound-guided PTCA alone in the Tübingen experience.11
The importance of an objective assessment of the angiographic results with online QCA and of the confirmation with CFR measurements of the restoration of flow conduction is shown by the opposite outcomes (significant difference in favor of stent implantation) of a retrospective comparison of PTCA and stent in complex lesions12 and of the recent randomized trial OPUS (OPtimal angioplasty Versus primary Stenting) using visual assessment for estimation of PTCA results.13 In the OPUS trial, crossover to stenting occurred in 37% of lesions, but the selection of inadequate PTCA results was clearly suboptimal, because a better clinical outcome was already observed in the stent group 2 months after treatment, too soon to represent a difference in restenosis. The strict inclusion criteria, different from those of DESTINI (single stent in >3.0-mm vessels), explain the extremely low rate of target lesion revascularization in the stent group (4%). A previous smaller-scale single-center trial of provisional stenting has shown a similar outcome of elective stenting and provisional stenting, but this study used a selection criterion that was not practical (development of late recoil after PTCA) and a coil stent with a restenosis rate higher than that of slotted tubular stents.14
The detection of a restored vessel conductance with CFR has limitations, because a low CFR may persist because of impaired microvascular vasodilatation15 and because CFR is affected by the acute changes of hemodynamic conditions and baseline velocity.16 17 The comparison with an untreated vessel (relative flow reserve) may overcome these limitations,18 but this method is cumbersome and requires the presence of a normal artery so that even the highly motivated DESTINI investigators measured CFR in a normal vessel only in a minority of cases. Fractional flow reserve, a physiological index of stenosis severity measurable with dedicated pressure guidewires, has been shown to be independent of the hemodynamic conditions at the time of measurement and not affected by changes of the microvascular tone.19 20 A previous observational study has shown that a myocardial fractional flow reserve >0.90 and a residual diameter stenosis <35% after PTCA are associated with a very low incidence of target lesion revascularization at 1 year (11%)21 and suggests that the improved specificity of this index may reduce the need for unnecessary stent implantations induced by the above-mentioned limitations of CFR.
In this study, follow-up angiography was performed only if clinically indicated to avoid distortion of the main end point (major adverse cardiac events, including target lesion revascularization) due to the “oculostenotic reflex.” A previous small-scale observational study with a design similar to that of DESTINI has recently shown a similar restenosis rate in the optimal PTCA and elective stenting groups.22
Because the clinical outcomes of the 2 treatment arms were equivalent, the operator should decide whether PTCA or stent should be used on the basis of cost and technical complexity. It is our belief that in optimal lesions for stent implantation, similar to the lesions included in the 3 main randomized PTCA versus stent trials (“Benestent”-like lesions),1 2 3 the operator’s preference will always be for elective stent implantation, with no need for a careful examination of morphological and physiological end points after PTCA.
The rapidly decreasing cost of stents and the use of direct stenting, without predilatation,23 may render stent implantation even more appealing in the future. The technique of aggressive guided PTCA used in this trial can find an application in small vessels and for complex lesions in which stent implantation requires the use of multiple stents (long lesions, bifurcations), is technically more demanding, and may induce a malignant type of restenosis with diffuse hyperplastic reaction extended to the stent ends, which is difficult to treat and at high risk of new recurrences.24 25 In case of intracoronary brachytherapy, aggressive PTCA may also be preferable, because stenting is associated with higher late subacute thrombosis.26
When optimal angiographic and physiological end points are met after PTCA, the early and late clinical outcomes are similar to the outcome observed after elective stent implantation.
DESTINI Study Group
C. Di Mario, A. Colombo, Milan, Italy; J. Moses, New York, NY.
Critical Events Committee
G. Roubin, New York, NY; C. Vassanelli, Novara, Italy; J. Fajadet, Toulouse, France.
QCA Core Laboratory and Data Coordinating and Analysis Center
L. Di Francesco, M. Repetto, A. Rombolotti, M. Recchia, MCR-Milan Cardiovascular Research, Milan, Italy.
J. Moses, E. Lawrence, New York, NY.
Responsible Investigators, Participating Centers, Number of Patients Enrolled
T. Anderson, J. Knutson, Foothills Hospital, Calgary, Canada (50); J. Moses, I. Moussa, Lenox Hill Hospital, New York, NY (50); R. Bonan, J.C. Tardif, Montreal Heart Institute, Montreal, Canada (50); T. Muramatsu, Kawasaki Central Hospital, Kawasaki-shi Kanagawa, Japan (50); A. Colombo, C. Di Mario, EMO Centro Cuore Columbus, Milan, Italy (43); A. Jain, West Virginia University/Ruby Memorial Hospital, Morgantown, W Va (36); J. Suarez de Lezo, M. Pan, Hospital Reina Sophia, Cordoba, Spain (33); S.Y. Cho, Y.S. Jang, Yonsei University Hospital, Seoul, South Korea (30); M. Kern, R. Bach, St Louis University, St Louis, Mo (28); I. Meredith, Monash Medical Center, Clayton, Australia (28); S. Kazziha, St John’s Hospital, Detroit, Mich (25); B. Weiner, University of Massachusetts Medical Center, Worcester, Mass (24); V. Aharonian, Kaiser Foundation Hospital, Los Angeles, Calif (23); S.J. Park, S.W. Park, Asan Medical Center, Seoul, Korea (21); H. Mudra, Innenstadt Muenchen, Muenchen, Germany (21); A. Frey, Herz-Zentrum, Bad Krozingen, Germany (20); M.L. Simard, White Memorial Hospital, Los Angeles, Calif (20); T. Fischell, Borgess Medical Center, Kalamazoo, Mich (19); E. Verna, S. Repetto, Ospedale di Circolo, Varese, Italy (15); B.W. Choi, Ajou University School of Medicine, Suwon, Korea (13); E. Caracciolo, Veterans Hospital/John Cochran, St Louis, Mo (11); M. Mosseri, Hadassah-Hebrew University Medical Center, Jerusalem, Israel (10); H. Madyoon, L. Chroushore, St Joseph’s, Stockton, Calif (9); H. Nonogi, National Cardiovascular Center, Osaka, Japan (9); M. Leon, A. Pichard, Washington Hospital Center, Washington, DC (7); M. Ayres, V. Rhule, Fort Sanders Regional Medical Center, Knoxville, Tenn (7); T. Akasaka, Kobe General Hospital, Chuo-ku Kobe, Japan (7); T. Kondo, Komaki Hospital, Komaki City, Japan (7); S. Brener, Cleveland Clinic Hospital, Cleveland, Ohio (6); A.H. Gershlick, Glenfield General Hospital, Leicester, England (6); T. Suzuki, National Toyohashi Higashi Hospital, Toyohashi, Japan (6); F. Crea, A. Maseri, Policlinico Universitario Gemelli, Roma (6); I. Penn, Vancouver Hospital, Vancouver, Canada (5); M. Nobuyoshi Kokura Memorial Hospital, Kitakyushu City, Japan (5); C. Seiler, B. Meier, Inselspital Bern, Bern, Switzerland (4); W. Kussmaul, J.D. Joye, Allegheny University Hospital, Philadelphia, Pa (4); D. Senior, N. Xenopoulos, Jewish Hospital, Louisville, Ky (4); A. Anwar, Baylor University Medical Center, Dallas, Tex (3); R. Stewart, University of Wisconsin, Madison, Wis (3); S. Werns, University of Michigan, Ann Arbor, Mich (3); R. White, Baptist Hospital, Oklahoma City, Okla (2); R. Bowerman, University Community Hospital, Tampa, Fla (2); P. Overlie, Methodist Hospital, Lubbock, Tex (2); K. Parr, Methodist Hospital, Indianapolis, Ind (2); M. Tobias, Akron General Medical Center, Akron, Ohio (2); S. Bailey, University of Texas Health Science Center, San Antonio, Tex (2); A. Gaglione, Villa Bianca, Bari, Italy (2); N. Kapilowicz, R. Beyar, Rambam Hospital, Haifa, Israel (2); T. Yamaguchi, Ohashi Hospital, Ohashi Meguro-ku, Japan (2); S. Lieberman, East Texas Medical Center, Tyler, Tex (2); M. Bertrand, J.M. Lablanche, Centre Hospitalier Regional et Universitaire, Lille, France (1); J.M. Ruggio, Pacific Cardiovascular Associates Medical Group, Fountain Valley, Calif (1); J. Margolis, Miami Heart Institute, Miami Beach, Fla (1); V. Sethi, Hackensack Medical Center, Hackensack, NJ (1); J. Lawson, Stony Brook, NY (1)
This study was supported by an unrestricted grant from Cardiometrics/Endosonics, Rancho Cordova, Calif, and by grants from Cordis–Johnson & Johnson, Warren, NJ, and from Boston Scientific, Natwick, Mass. The authors acknowledge the technical assistance of Lucia Di Francesco, PhD (Trial Coordinator, non-US centers) and Emily Lawrence, MSc (Trial Coordinator, US centers). The assistance of Andrew Ford, Regina Jones, and Adam Savakus in the organization and conduction of the study is gratefully acknowledged.
A complete list of DESTINI Investigators can be found in the Appendix.
- Received May 15, 2000.
- Revision received August 10, 2000.
- Accepted August 10, 2000.
- Copyright © 2000 by American Heart Association
Nissen SE, Di Mario C, Tuzcu ME. Intravascular ultrasound, angioscopy, Doppler and pressure measurements. In: Topol EJ, ed. Textbook of Cardiovascular Medicine. Philadelphia, Pa: Lippincott Raven; 1998:2119–2153.
Doucette JW, Douglas Corl P, Payne HP, et al. Validation of a Doppler guide wire for intravascular measurement of coronary artery flow velocity. Circulation.. 1992;85:1899–1911.
Carlier SG, Di Mario C, Kern MJ, et al. Intracoronary Doppler and pressure monitoring. In: Topol EJ, ed. Textbook of Interventional Cardiology. 2nd ed. Philadelphia, Pa: Saunders; 1999:748–781.
Serruys PW, Di Mario C, Piek J, et al. Prognostic value of intracoronary flow velocity and diameter stenosis in assessing the short and long-term outcome of coronary balloon angioplasty. The DEBATE Study (Doppler End-points Balloon Angioplasty Trial Europe) Circulation.. 1997;96:3369–3377.
Keane D, Haase J, Slager CJ, et al. Comparative validation of quantitative coronary angiography systems: results and implications from a multicenter trial using a standardized approach. Circulation.. 1995;91:2174–2183.
Stone GW, Hodgson JMcB, St Goar FG, et al, for the CLinical Outcome with Ultrasound Trial (CLOUT) Investigators. Improved procedural results of coronary angioplasty with intravascular ultrasound guided balloon sizing. Circulation.. 1997;95:2044–2052.
Haase KH, Athanasiadis A, Mahrholdt H, et al. Acute and 1 year follow-up results after vessel size adapted PTCA using intracoronary ultrasound. Eur Heart J.. 1998;19:263–272.
Wilson RF, Johnson MR, Marcus ML, et al. The effect of coronary angioplasty on blood flow reserve. Circulation.. 1988;77:873–885.
Baumgart D, Haude M, Goerge G, et al. Improved assessment of coronary stenosis severity using the relative flow velocity reserve. Circulation.. 1998;98:40–46.
De Bruyne, Bartunek J, Sys SU, et al. Simultaneous coronary pressure and flow measurements in humans: feasibility, reproducibility, and hemodynamic dependence of coronary flow velocity reserve, hyperemic flow versus pressure slope index, and fractional flow reserve. Circulation.. 1996;94:1843–1849.
Bech GJ, Pijls NH, De Bruyne B, et al. Usefulness of fractional flow reserve to predict clinical outcome after balloon angioplasty. Circulation.. 1999;99:883–888.
Dupoy P, Pelle G, Garot P, et al. Physiologically guided angioplasty in support to a provisional stenting strategy: immediate and six-month outcome. Cathet Cardiovasc Diagn.. 2000;49:369–375.
Mehran R, Dangas G, Abizaid A, et al. Angiographic patterns of in-stent restenosis: classification and implications for long-term outcome. Circulation.. 1999;100:1872–1878.
Costa MA, Sabaté M, van der Giessen WJ, et al. Late coronary occlusion after intracoronary brachytherapy. Circulation.. 1999;100:789–792