[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 18.207.136.184. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Figure 1.
Kaplan-Meier Cumulative Event Curves for Stent Groups
Kaplan-Meier Cumulative Event Curves for Stent Groups

Kaplan-Meier cumulative incidence curves at 5 years are given for participants in the randomized TWENTE (Real-World Endeavor Resolute vs Xience V Drug-Eluting Stent Study in Twente) trial for the zotarolimus-eluting stent (ZES) (Resolute; Medtronic Inc) and the everolimus-eluting stent (EES) (Xience V; Abbott Vascular) for target vessel failure (TVF), a composite of cardiac death, target vessel–related myocardial infarction (MI), or target vessel revascularization.

aP = .94 for comparison between groups, log-rank test.

bP = .18 for comparison between groups, log-rank test.

cP = .46 for comparison between groups, log-rank test.

dP = .28 for comparison between groups, log-rank test.

eP = .99 for comparison between groups, log-rank test.

fP = .87 for comparison between groups, log-rank test.

gP = .54 for comparison between groups, log-rank test.

hP = .15 for comparison between groups, log-rank test.

Figure 2.
Cumulative Incidence of Definite (DST) or Probable (PST) Stent Thrombosis in the Randomized Trial
Cumulative Incidence of Definite (DST) or Probable (PST) Stent Thrombosis in the Randomized Trial

Data include participants in the randomized TWENTE (Real-World Endeavor Resolute vs Xience V Drug-Eluting Stent Study in Twente) trial for the zotarolimus-eluting stent (ZES) (Resolute; Medtronic Inc) and the everolimus-eluting stent (EES) (Xience V; Abbott Vascular). Stent thrombosis was adjudicated by the Academic Research Consortium definition.

aP = .83 for comparison between groups, log rank test.

bIndicates dual antiplatelet therapy (acetylsalicylic acid plus P2Y12 receptor antagonist).

Figure 3.
Kaplan-Meier Cumulative Event Curves for All Comers
Kaplan-Meier Cumulative Event Curves for All Comers

Kaplan-Meier cumulative incidence curves at 5 years are given for the randomized TWENTE (Real-World Endeavor Resolute vs Xience V Drug-Eluting Stent Study in Twente) trial participants (n = 1391), nonenrolled eligible patients (n = 318), and both groups combined (n = 1709) for target vessel failure (TVF), a composite of cardiac death, target vessel–related myocardial infarction (MI), or target vessel revascularization. Blue dotted lines represent the cumulative event incidence for the pooled patient population of TWENTE trial participants and nonenrolled eligible patients.

aP = .02 for comparison between groups, log-rank test.

bP = .03 for comparison between groups, log-rank test.

cP = .80 for comparison between groups, log-rank test.

dP = .36 for comparison between groups, log-rank test.

Table 1.  
Five-Year Clinical Outcome of TWENTE Randomized Trial Participants According to Assigned DES
Five-Year Clinical Outcome of TWENTE Randomized Trial Participants According to Assigned DES
Table 2.  
Five-Year Clinical Outcomes of the TWENTE Randomized Trial Participants vs Nonenrolled Eligible Patients
Five-Year Clinical Outcomes of the TWENTE Randomized Trial Participants vs Nonenrolled Eligible Patients
1.
Windecker  S, Serruys  PW, Wandel  S,  et al.  Biolimus-eluting stent with biodegradable polymer versus sirolimus-eluting stent with durable polymer for coronary revascularisation (LEADERS): a randomised non-inferiority trial.  Lancet. 2008;372(9644):1163-1173.PubMedGoogle ScholarCrossref
2.
Kedhi  E, Joesoef  KS, McFadden  E,  et al.  Second-generation everolimus-eluting and paclitaxel-eluting stents in real-life practice (COMPARE): a randomised trial.  Lancet. 2010;375(9710):201-209.PubMedGoogle ScholarCrossref
3.
Serruys  PW, Silber  S, Garg  S,  et al.  Comparison of zotarolimus-eluting and everolimus-eluting coronary stents.  N Engl J Med. 2010;363(2):136-146.PubMedGoogle ScholarCrossref
4.
Christiansen  EH, Jensen  LO, Thayssen  P,  et al; Scandinavian Organization for Randomized Trials with Clinical Outcome (SORT OUT) V Investigators.  Biolimus-eluting biodegradable polymer-coated stent vs durable polymer-coated sirolimus-eluting stent in unselected patients receiving percutaneous coronary intervention (SORT OUT V): a randomised non-inferiority trial [published correction appears in Lancet. 2013;382(9889):310].  Lancet. 2013;381(9867):661-669.PubMedGoogle ScholarCrossref
5.
Maeng  M, Tilsted  HH, Jensen  LO,  et al.  Differential clinical outcomes after 1 year versus 5 years in a randomised comparison of zotarolimus-eluting and sirolimus-eluting coronary stents (the SORT OUT III study): a multicentre, open-label, randomised superiority trial.  Lancet. 2014;383(9934):2047-2056.PubMedGoogle ScholarCrossref
6.
Bligaard  N, Thuesen  L, Saunamäki  K,  et al; SORT OUT II Investigators.  Similar five-year outcome with paclitaxel- and sirolimus-eluting coronary stents.  Scand Cardiovasc J. 2014;48(3):148-155.PubMedGoogle ScholarCrossref
7.
von Birgelen  C, Basalus  MW, Tandjung  K,  et al.  A randomized controlled trial in second-generation zotarolimus-eluting Resolute stents versus everolimus-eluting Xience V stents in real-world patients: the TWENTE trial.  J Am Coll Cardiol. 2012;59(15):1350-1361.PubMedGoogle ScholarCrossref
8.
von Birgelen  C, Sen  H, Lam  MK,  et al.  Third-generation zotarolimus-eluting and everolimus-eluting stents in all-comer patients requiring a percutaneous coronary intervention (DUTCH PEERS): a randomised, single-blind, multicentre, non-inferiority trial.  Lancet. 2014;383(9915):413-423.PubMedGoogle ScholarCrossref
9.
Hordijk-Trion  M, Lenzen  M, Wijns  W,  et al; EHS-CR Investigators.  Patients enrolled in coronary intervention trials are not representative of patients in clinical practice: results from the Euro Heart Survey on Coronary Revascularization.  Eur Heart J. 2006;27(6):671-678.PubMedGoogle ScholarCrossref
10.
Hutchinson-Jaffe  AB, Goodman  SG, Yan  RT,  et al; Canadian Acute Coronary Syndromes (ACS) Registry I and II Investigators and Canadian Global Registry of Acute Coronary Events (GRACE/GRACE 2) Investigators.  Comparison of baseline characteristics, management and outcome of patients with non–ST-segment elevation acute coronary syndrome in versus not in clinical trials.  Am J Cardiol. 2010;106(10):1389-1396.PubMedGoogle ScholarCrossref
11.
Rothwell  PM.  External validity of randomised controlled trials: “to whom do the results of this trial apply”?  Lancet. 2005;365(9453):82-93.PubMedGoogle ScholarCrossref
12.
de Boer  SPM, Lenzen  MJ, Oemrawsingh  RM,  et al.  Valuating the “all-comers” design: a comparison of participants in two “all-comers” PCI trials with non-participants.  Eur Heart J. 2011;32(17):2161-2167.PubMedGoogle ScholarCrossref
13.
Sen  H, Tandjung  K, Basalus  MW,  et al.  Comparison of eligible non-enrolled patients and the randomised TWENTE trial population treated with Resolute and Xience V drug-eluting stents.  EuroIntervention. 2012;8(6):664-671.PubMedGoogle ScholarCrossref
14.
Meredith  IT, Worthley  S, Whitbourn  R,  et al; RESOLUTE Investigators.  Clinical and angiographic results with the next-generation resolute stent system: a prospective, multicenter, first-in-human trial.  JACC Cardiovasc Interv. 2009;2(10):977-985.PubMedGoogle ScholarCrossref
15.
Iqbal  J, Serruys  PW, Silber  S,  et al.  Comparison of zotarolimus- and everolimus-eluting coronary stents: final 5-year report of the RESOLUTE All-Comers trial.  Circ Cardiovasc Interv. 2015;8(6):e002230.PubMedGoogle ScholarCrossref
16.
World Medical Association.  World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects.  JAMA. 2013;310(20):2191-2194.PubMedGoogle ScholarCrossref
17.
Tandjung  K, Sen  H, Lam  MK,  et al.  Clinical outcome following stringent discontinuation of dual antiplatelet therapy after 12 months in real-world patients treated with second-generation zotarolimus-eluting Resolute and everolimus-eluting Xience V stents: 2-year follow-up of the randomized TWENTE trial.  J Am Coll Cardiol. 2013;61(24):2406-2416.PubMedGoogle ScholarCrossref
18.
Löwik  MM, Lam  MK, Sen  H,  et al.  Safety of second-generation drug-eluting stents three years after randomised use in the TWENTE trial.  EuroIntervention. 2015;10(11):1276-1279.PubMedGoogle ScholarCrossref
19.
Cutlip  DE, Windecker  S, Mehran  R,  et al; Academic Research Consortium.  Clinical end points in coronary stent trials: a case for standardized definitions.  Circulation. 2007;115(17):2344-2351.PubMedGoogle ScholarCrossref
20.
Vranckx  P, Cutlip  DE, Mehran  R,  et al.  Myocardial infarction adjudication in contemporary all-comer stent trials: balancing sensitivity and specificity: addendum to the historical MI definitions used in stent studies.  EuroIntervention. 2010;5(7):871-874.PubMedGoogle ScholarCrossref
21.
Mehilli  J, Richardt  G, Valgimigli  M,  et al; ISAR-LEFT-MAIN 2 Study Investigators.  Zotarolimus- vs everolimus-eluting stents for unprotected left main coronary artery disease.  J Am Coll Cardiol. 2013;62(22):2075-2082.PubMedGoogle ScholarCrossref
22.
Park  KW, Kang  SH, Kang  HJ,  et al; HOST–ASSURE Investigators.  A randomized comparison of platinum chromium-based everolimus-eluting stents vs cobalt chromium-based zotarolimus-eluting stents in all-comers receiving percutaneous coronary intervention: HOST-ASSURE, a randomized, controlled, noninferiority trial.  J Am Coll Cardiol. 2014;63(25, pt A):2805-2816.PubMedGoogle ScholarCrossref
23.
Navarese  EP, Tandjung  K, Claessen  B,  et al.  Safety and efficacy outcomes of first and second generation durable polymer drug eluting stents and biodegradable polymer biolimus eluting stents in clinical practice: comprehensive network meta-analysis.  BMJ. 2013;347:f6530.PubMedGoogle ScholarCrossref
24.
Palmerini  T, Benedetto  U, Biondi-Zoccai  G,  et al.  Long-term safety of drug-eluting and bare-metal stents: evidence from a comprehensive network meta-analysis.  J Am Coll Cardiol. 2015;65(23):2496-2507.PubMedGoogle ScholarCrossref
25.
Piccolo  R, Stefanini  GG, Franzone  A,  et al.  Safety and efficacy of resolute zotarolimus-eluting stents compared with everolimus-eluting stents: a meta-analysis.  Circ Cardiovasc Interv. 2015;8(4):e002223.PubMedGoogle ScholarCrossref
26.
Gada  H, Kirtane  AJ, Newman  W,  et al.  5-year results of a randomized comparison of XIENCE V everolimus-eluting and TAXUS paclitaxel-eluting stents: final results from the SPIRIT III trial (clinical evaluation of the XIENCE V everolimus eluting coronary stent system in the treatment of patients with de novo native coronary artery lesions).  JACC Cardiovasc Interv. 2013;6(12):1263-1266.PubMedGoogle ScholarCrossref
27.
Smits  PC, Vlachojannis  GJ, McFadden  EP,  et al.  Final 5-year follow-up of a randomized controlled trial of everolimus- and paclitaxel eluting stents for coronary revascularization in daily practice: the COMPARE trial (A Trial of Everolimus-Eluting Stents and Paclitaxel-Eluting Stents for Coronary Revascularization in Daily Practice).  JACC Cardiovasc Interv. 2015;8(9):1157-1165.PubMedGoogle ScholarCrossref
28.
Jensen  LO, Thayssen  P, Christiansen  EH,  et al; SORT OUT IV Investigators.  Safety and efficacy outcomes of everolimus- vs sirolimus-eluting stents: the SORT OUT IV 5-year results.  J Am Coll Cardiol. 2016;67(7):751-762.PubMedGoogle ScholarCrossref
29.
Sabaté  M, Brugaletta  S, Cequier  A,  et al.  Clinical outcomes in patients with ST–segment elevation myocardial infarction treated with everolimus-eluting stents versus bare-metal stents (EXAMINATION): 5-year results of a randomised trial.  Lancet. 2016;387(10016):357-366.PubMedGoogle ScholarCrossref
30.
von Birgelen  C, Kok  MM, van der Heijden  LC,  et al.  Very thin strut biodegradable polymer everolimus-eluting and sirolimus-eluting stents versus durable polymer zotarolimus-eluting stents in allcomers with coronary artery disease (BIO-RESORT): a three-arm, randomised, non-inferiority trial.  Lancet. 2016;388(10060):2607-2617.PubMedGoogle ScholarCrossref
Original Investigation
March 2017

Five-Year Outcome After Implantation of Zotarolimus- and Everolimus-Eluting Stents in Randomized Trial Participants and Nonenrolled Eligible Patients: A Secondary Analysis of a Randomized Clinical Trial

Author Affiliations
  • 1Department of Cardiology, Thoraxcentrum Twente, Medisch Spectrum Twente, Enschede, the Netherlands
  • 2Health Technology and Services Research, MIRA–Institute for Biomedical Technology and Technical Medicine, University of Twente, Enschede, the Netherlands
  • 3Department of Cardiology, Ziekenhuisgroep Twente, Almelo and Hengelo, the Netherlands
  • 4Department of Epidemiology, Medisch Spectrum Twente, Enschede, the Netherlands
  • 5Department of Research Methodology, Measurement and Data Analysis, University of Twente, Enschede, the Netherlands
JAMA Cardiol. 2017;2(3):268-276. doi:10.1001/jamacardio.2016.5190
Key Points

Questions  Are newer-generation drug-eluting stents safe and efficacious at the 5-year follow-up, and do outcomes of eligible nonenrolled patients treated with the same stents support the randomized clinical trial findings?

Findings  In this secondary analysis of the TWENTE (Real-World Endeavor Resolute vs Xience V Drug-Eluting Stent Study in Twente) randomized clinical trial, which examined 81% of all the eligible patients, the rate of the main composite clinical end point of target vessel failure was similar for zotarolimus- and everolimus-eluting stents. In all eligible patients, the event rate was only slightly higher than in trial participants only.

Meaning  At long-term follow-up of this randomized clinical trial, newer-generation stents were safe and efficacious, and the outcomes of nonenrolled eligible patients support the trial’s validity.

Abstract

Importance  Long-term follow-up after a clinical trial of 2 often-used, newer-generation drug-eluting stents (DESs) in a broad patient population is of interest. Comprehensive long-term outcome of eligible nonenrolled patients has never been reported.

Objective  To assess 5-year safety and efficacy of 2 newer-generation DESs in randomized participants with non–ST-elevation acute coronary syndromes or stable angina and to evaluate long-term outcomes of nonenrolled eligible patients treated with the same DESs.

Design, Setting, and Participants  The TWENTE (Real-World Endeavor Resolute vs Xience V Drug-Eluting Stent Study in Twente) trial is an investigator-initiated, patient-blinded, randomized, comparative DES trial that enrolled patients from June 18, 2008, to August 26, 2010. Most patients had non–ST-elevation acute coronary syndromes and complex lesions. Of all 1709 eligible patients, 1391 (81.4%) were treated in the TWENTE trial with zotarolimus-eluting (ZES, n = 697) or everolimus-eluting (EES, n = 694) cobalt-chromium stents. The remaining 318 eligible patients (18.6%) were not enrolled but underwent nonrandomized treatment with the same DESs. Data were analyzed from August 26, 2015, to October 11, 2016. Event rates (percentages) were derived from log-rank analysis and may differ from straightforward calculation (nominator/denominator). The 5-year follow-up of the TWENTE participants was prespecified in the trial protocol; that of the nonenrolled participants was ad hoc.

Main Outcomes and Measures  Target vessel failure (TVF), a composite of cardiac death, target vessel–related myocardial infarction, or target vessel revascularization.

Results  Of 1709 eligible participants, 1233 (72.1%) were men, 476 (27.9%) were women, and mean (SD) age was 64.6 (10.6) years. Among the 1370 of 1391 TWENTE trial participants (98.5% follow-up), TVF was similar between those in the ZES (16.1%) and EES (18.1%) groups (P = .36). Stent thrombosis rates were low: definite (7 of 697 [1.0%] vs 4 of 694 [0.6%]; P = .37) and occurred after more than 1 year in 3 (0.4%) with ZES vs 4 (0.6%) with EES (P = .69). The 318 nonenrolled eligible patients (308 patients [96.9%] of whom were followed up) were older and had more advanced disease than trial participants. Their TVF rate was higher than that of trial participants (71 of 318 [23.3%] vs 233 of 1391 [17.1%]; P = .02), which partly reflects a difference in cardiac mortality (23 of 318 [7.7%] vs 60 of 1391 [4.5%]; P = .03). Similar 5-year rates were found for myocardial infarction (91 of 1391 [6.7%] vs 22 of 318 [7.2%]; P = .80) and target vessel revascularization (129 of 1391 [9.7%] vs 34 of 318 [11.4%]; P = .36) between trial participants and nonenrolled eligible patients. In all eligible patients (ie, trial participants plus nonenrolled eligible patients), the TVF rate was only slightly higher than in trial participants only (18.3% vs 17.1%).

Conclusions and Relevance  Long-term outcome data from nonenrolled eligible patients support the validity of the TWENTE trial findings and present, with the trial, a strong case for the long-term safety and efficacy of the newer-generation DESs used.

Trial Registration  clinicaltrials.gov Identifier: NCT01066650

Introduction

Randomized clinical trials that compare novel drug-eluting stents (DESs) in large, greatly unrestricted patient populations are indispensable because they permit a reliable evaluation of the efficacy and safety of these devices.1-8 Availability of long-term follow-up from such trials in all comers is a prerequisite for trustworthy judgment of the long-term effects of percutaneous coronary interventions (PCIs) using these devices.5 However, almost all randomized clinical trials (RCTs) in broad patient and all-comer populations involve some selection.9-11 Information on the clinical characteristics and short-term mortality of patients who did not participate in an RCT is infrequently reported.4,12,13

The zotarolimus-eluting stent (ZES) (Resolute; Medtronic Inc) and the everolimus-eluting stent (EES) (Xience V; Abbott Vascular) are newer-generation DESs that were developed to increase device biocompatibility and improve long-term outcomes.2,3,14 The randomized TWENTE (Real-World Endeavor Resolute vs Xience V Drug-Eluting Stent Study in Twente) trial demonstrated noninferiority of the ZES vs EES in 1391 patients, representing 81.4% of all 1709 eligible patients of a broad population.7 The Nonenrolled TWENTE registry compared the clinical characteristics and outcomes of the 1391 trial participants vs 318 nonenrolled eligible patients who were treated with the same DESs.13 After 12 months, the rates of adverse events were fairly similar for all trial participants (treatment arms pooled) vs the nonenrolled eligible patients.13

Five-year outcome data from randomized comparisons of the ZES and EES are of clinical interest but are so far only available from the RESOLUTE AC (A Randomized Comparison of a Zotarolimus-Eluting Stent With an Everolimus-Eluting Stent for Percutaneous Coronary Intervention) trial.15 In addition, long-term outcome data from eligible nonenrolled patients treated with the same stents as the randomized patients have not yet been reported. In the present study, we report the 5-year clinical outcomes of patients enrolled in the randomized TWENTE trial and eligible nonenrolled patients.

Methods
Study Design and Patient Populations

The TWENTE trial is an investigator-initiated, patient-blinded, randomized, comparative DES trial with limited exclusion criteria.7 During the RCT, nonenrolled eligible patients were treated at the operator’s discretion with one of the European Conformity–certified DESs that were also examined in the randomized trial (ie, ZES or EES), using the same routine clinical and procedural strategies. The RCT and registry complied with the Declaration of Helsinki16 for investigation in human beings and were approved by the medical ethical committee of Twente and the institutional review board. All participants in the RCT provided written informed consent. For the Nonenrolled TWENTE registry, patients were not required to change their behavior or take action other than following their regular treatment; therefore, according to Dutch law and as approved by the medical ethical committee of Twente, written informed consent from patients in this registry was not required.

Patients were enrolled in the TWENTE trial from June 18, 2008, to August 26, 2010, at Thoraxcentrum Twente, Enschede, the Netherlands. Patients were randomized in a 1:1 ratio for treatment with the cobalt-chromium ZES or EES. No limit for lesion length, reference vessel size, number of target lesions, or number of vessels to be treated was applied. The main exclusion criterion was an ST-segment elevation myocardial infarction (MI) before PCI (<48 hours).7 Outcomes of the TWENTE trial participants were reported until the 3-year follow-up.17,18 Operators generally avoided the use of different stents in a single patient. Procedural details and the 12-month clinical course of the nonenrolled eligible patients have been published elsewhere.13 The outcomes of all eligible patients were followed up as participants of the TWENTE trial or within the Nonenrolled TWENTE registry. Stent-level analyses were only performed in participants of the TWENTE trial.

Definition of Clinical End Points

The same end point definitions were applied in the Nonenrolled TWENTE registry as used for the randomized TWENTE trial.7,13 Clinical end points, including stent thrombosis, were defined according to the Academic Research Consortium, including the addendum on the definition of MI.19,20 The prespecified main end point of the RCT—target vessel failure (TVF)—was composed (in hierarchical order) of cardiac death, target vessel–related MI, or clinically indicated target vessel revascularization.9 Death was considered cardiac unless an unequivocal noncardiac cause could be established. Myocardial infarction was defined by any creatine kinase concentration of more than double the upper reference limit with elevated confirmatory cardiac biomarkers.20 A target vessel–related MI was related to the target vessel or could not be related to another vessel; further MI classification was based on laboratory, electrocardiographic, angiographic, and/or clinical data.7,20 Revascularization procedures were considered clinically indicated if the visually assessed angiographic percentage diameter of stenosis was at least 50% in the presence of ischemic signs or symptoms or if the diameter stenosis was at least 70% irrespective of ischemic signs or symptoms.20

Prespecified secondary end points included the individual components of TVF, all-cause mortality, and stent thrombosis. Further composite end points included target lesion failure (cardiac death, target vessel–related MI, or clinically indicated target lesion revascularization), major adverse cardiac events (all-cause death, any MI, emergent coronary bypass surgery, or clinically indicated target lesion revascularization), and a more global patient-oriented composite end point (all-cause death, any MI, or any coronary revascularization).

Acquisition and Analysis of Clinical Follow-up Data

We obtained clinical follow-up data at visits to outpatient clinics, by medical questionnaire, and/or by telephone follow-up (with staff blinded to the assigned DES). The contract research organization CardioResearch Enschede BV, Enschede, the Netherlands, coordinated the trial and data management. The independent contract research organizations Cardialysis, Rotterdam, the Netherlands, and Diagram, Zwolle, the Netherlands, which were blinded to the assigned DES, performed the adjudication of adverse clinical events for all patients. Angiographic analysts performed a visual assessment of target lesion characteristics offline.

Assessment of Reasons for Nonenrollment and Potential Selection

To identify the explicitly stated and implicit reasons for nonenrollment of eligible patients, 5 members of the research group reviewed all medical files available. If the reason for nonenrollment was not explicitly stated, the presumed reason for nonenrollment was retrospectively established by consensus of the committee members. Reasons for nonenrollment were classified into 1 of the following categories: (1) explicit refusal of the informed patient; (2) inability of the operator to obtain informed consent (eg, owing to severe anxiety or partial deafness); (3) logistic aspects (eg, no randomization envelopes available, forgotten to randomize, or time pressure); (4) omission of informing the patient before treatment on the ward; and (5) unknown. In the absence of (conceivable) reasons, the committee also searched for indications of potential selection, which was suspected if vessels with particularly high risk were treated (eg, left main stem, degenerated vein grafts, or all 3 coronary arteries) or if operators indicated an increased procedural risk, high technical complexity of the procedure, or serious comorbidities.

Statistical Analysis

Data were analyzed from August 26, 2015, to October 11, 2016. The 5-year follow-up of the TWENTE participants was prespecified in the trial protocol; that of the nonenrolled participants was ad hoc. Analyses were based on intention to treat using SPSS software (version 22.0; SPSS Inc). Categorical variables were assessed with the use of χ2 or Fisher exact tests, whereas continuous variables were assessed with the Wilcoxon rank sum test or 1-sample t test. The time to the main end point and its components was assessed according to the Kaplan-Meier method, and the log-rank test was applied to compare groups. Therefore, percentages of outcome factors may differ slightly from the results of a straightforward calculation of nominator divided by denominator. We calculated hazard ratios using Cox proportional hazards regression analysis. Logistic regression was used to test for interaction between subgroups and stent type with respect to the main end point. Unless otherwise specified, P values and CIs were 2 sided. P < .05 was considered significant.

Results
Clinical Outcome of the TWENTE Trial

A total of 1709 eligible participants (1233 men [72.1%] and 476 women [27.9%]; mean [SD] age, 64.6 [10.6] years) were included in the analysis. Of 1391 TWENTE trial participants (1009 men [72.5%]; 382 women [27.5%]), 5-year follow-up was obtained from 1370 (98.5%), including 683 of 697 patients (98.0%) randomized to ZES and 687 of 694 (99.0%) randomized to EES (eFigure 1 in the Supplement). Between stent groups, no difference was found in baseline patient, lesion, and procedural characteristics (eTable 1 in the Supplement). At the 5-year follow-up, 61 of the 1228 trial participants (5.0%) were receiving dual antiplatelet therapy without any between-stent difference.

The main end point of TVF was met by 110 of 697 patients (16.1%) in the ZES group vs 123 of 694 (18.1%) in the EES group (hazard ratio, 0.89; 95% CI, 0.69-1.15; P = .36) (Table 1 and Figure 1). No significant difference was seen in the individual components of TVF, including cardiac death (25 of 697 [3.7%] vs 35 of 694 [5.2%]; P = .18), target vessel–related MI (46 of 697 [6.8%] vs 45 of 694 [6.6%]; P = .94), and clinically driven target vessel revascularization (60 of 697 [8.9%] vs 69 of 694 [10.5%], respectively; P = .41). In addition, the rates of other composite end points were similar for both DESs (Table 1). An exploratory subgroup analysis of the main end point of TVF showed consistent results across subgroups, with the only exception being single-vessel treatment, favoring ZES (67 of 523 [12.8%] vs 92 of 532 [17.3%], respectively; P = .05; P = .03 for interaction) (eFigure 2 in the Supplement).

For ZES and EES, the rates of definite (7 of 697 [1.0%] vs 4 of 694 [0.6%]; P = .37) and definite or probable (13 of 697 [1.9%] vs 14 of 694 [2.1%], respectively; P = .83) stent thrombosis were low and similar (Figure 2). Very late definite stent thrombosis (>1 year) occurred in 3 patients in the ZES group (0.4%) vs 4 patients in the EES group (0.6%) (P = .69). In eTable 2 in the Supplement, annual cumulative event rates for both stent groups are presented.

Nonenrolled Eligible Patient Population

Five-year follow-up data were available for 308 of 318 eligible patients (96.9%) (224 men [70.4%]; 94 women [29.6%]) who had not been enrolled in the randomized trial (eFigure 1 in the Supplement). Reasons for nonenrollment included (1) explicit refusal to participate (33 of 318 [10.4%]); (2) inability to obtain informed consent (57 [17.9%]); (3) logistic aspects (40 [12.6%]); (4) omission of informing the patient before treatment (113 [35.5%]); and (5) unknown reasons (75 [23.6%]). Indications of potential selection were found in 61 of 318 (19.2%) nonenrolled eligible patients, representing 3.6% of all 1709 eligible patients. Nonenrolled eligible patients were significantly older than randomized participants (66.0 [10.9] vs 64.2 [10.8] years; P = .01), significantly more often had a history of MI (137 [43.1%] vs 450 [32.4%]; P < .001), PCI (92 [28.9%] vs 288 [20.7%]; P = .001), and bypass surgery (54 [17.0%] vs 148 [10.6%]; P = .002), and significantly more often had chronic renal failure (21 [6.6%] vs 38 [2.7%]; P = .001) and a left ventricular ejection fraction of less than 30% (13 of 199 [6.5%] vs 32 of 1051 [3.0%]; P = .02). In addition, they were treated more often for complex target lesions (355 of 466 [76.2%] vs 1484 of 2116 [70.1%]; P = .009) and in-stent restenosis (37 of 466 [7.9%] vs 75 of 2116 [3.5%]; P < .001) (eTable 3 in the Supplement). At the 5-year follow-up, 19 of 264 patients (7.2%) were receiving dual antiplatelet therapy (P = .15 vs RCT participants).

TWENTE Trial Participants vs Nonenrolled Eligible Patients

The TWENTE trial participants and nonenrolled eligible patients differed in TVF (233 of 1391 [17.1%] vs 71 of 318 [23.3%]; P = .02), which was partly attributable to a difference in cardiac death (60 of 1391 [4.5%] vs 23 of 318 [7.7%]; P = .03). The Kaplan-Meier curves showed that, until the 4-year follow-up, the cardiac death rates were similar, after which the slope of the curve increased for nonenrolled patients (Figure 3). Both patient groups had quite similar 5-year rates of target vessel–related MI (91 of 1391 [6.7%] vs 22 of 318 [7.2%]; P = .80) and target vessel revascularization (129 of 1391 [9.7%] vs 34 of 318 [11.4%]; P = .36). The 5-year definite-or-probable stent thrombosis rate was low in trial participants and nonenrolled eligible patients (27 of 1391 [2.0%] vs 3 of 318 [1.0%]; P = .23). Further outcome data are presented in Table 2, and landmark analyses for TVF and its components are presented in eFigure 3 in the Supplement.

Table 2 also shows the event rates of all eligible patients (ie, a pooled population of trial participants and nonenrolled eligible patients). The 5-year TVF rate was 304 of 1709 (18.3%) in all eligible patients and 233 of 1391 (17.1%) in randomized trial participants only (stent arms pooled); this finding is also visualized in eFigure 4 in the Supplement.

Discussion

Randomized studies generally do not randomize every eligible patient and are therefore susceptible to selection.9-11 The TWENTE trial enrolled a large proportion of all eligible patients (81.4%), whereas excellent multicenter DES trials9-11 previously enrolled 40% of eligible patients or did not report such details. Minor selection cannot be excluded and may—from a clinician’s perspective—sometimes appear reasonable in patients with end-stage coronary heart disease or excessive comorbidities. The present analysis is special in that it reports long-term outcome data from the randomized trial and a registry of nonenrolled eligible patients, which together provide unique, complementary insights.

In the present study, TWENTE trial participants treated with ZES vs EES showed 5-year TVF rates that were relatively low (16.1% vs 18.1%) in both treatment arms, as were the rates of the individual components of TVF. Our findings are in line with those of other RCTs that compared similar ZESs and EESs,3,8,21,22 network meta-analyses that—among other DESs—constituted the 2 DESs assessed in the TWENTE trial,23,24 and a dedicated meta-analysis.25

Comprehensive information was available also on the long-term outcome of nonenrolled eligible patients, who were treated in a nonrandomized fashion with the same DESs. The adverse events of these 2 patient populations were adjudicated by the same independent clinical event committee, which further increases the usefulness of these data. Moreover, 5-year follow-up rates of trial participants and nonenrolled eligible patients were high (98.5% and 96.9%, respectively). In nonenrolled eligible patients, who were older and more often had a history of coronary revascularizations and MI than RCT participants, the TVF rate was higher (23.3% vs 17.1%). Until the 4-year follow-up, the rate of cardiac death remained similar in both patient groups. During the fifth year of follow-up, cardiac mortality increased more noticeably in the nonenrolled patients. The 5-year cardiac death rate (7.7% vs 4.5%) reflected the more advanced age and disease stage of the nonenrolled patients and contributed to the aforementioned difference in TVF. Our data suggest that if all eligible patients had been randomized, clinical event rates might still have been favorable and, in general, only slightly higher than those actually obtained for the randomized population. Hence, the long-term outcome data from the nonenrolled eligible patients support the external validity and findings of the TWENTE trial.

Previous Studies Addressing Nonenrolled Patients

Patients enrolled in coronary intervention trials are often not fully representative of patients in clinical practice.9-11 Nevertheless, data on the clinical outcome of eligible nonenrolled patients or of all nonparticipants (ie, nonenrolled eligible plus per-protocol excluded patients) are scarce. High-quality clinical trials occasionally report the clinical characteristics of the nonparticipating patients to provide insight into the degree (or the absence) of selection.4 A single high-volume PCI center reported baseline characteristics and 12-month all-cause mortality of 579 patients who participated in 2 randomized all-comer trials12 and compared these data with the mortality of 663 nonparticipants. Baseline characteristics differed significantly between trial participants and nonparticipants, with the latter being older and having more heart failure and unstable clinical syndromes, and finally a significantly higher 1-year all-cause mortality (6.9% vs 3.1%).12 Because the nonparticipants included patients with exclusion criteria (eg, cardiogenic shock) and the outcome was focused on 12-month all-cause mortality,12 a meaningful comparison with the 5-year outcome of the Nonenrolled TWENTE registry cannot be made. Nevertheless, in our study, 5-year all-cause mortality also tended to be higher in nonenrolled eligible patients than in trial participants (43 [13.8%] vs 142 [10.3%]).

Previous Randomized Studies With 5-Year Follow-up Comparing Newer-Generation DESs

Newer-generation durable polymer DESs have previously been assessed in randomized trials with long-term follow-up. In patients with low to moderate procedural risk in the SPIRIT III (Clinical Evaluation of the Xience V Everolimus Eluting Coronary Stent System) trial, the 5-year rate of TVF was lower in patients using the EES than in those using early-generation paclitaxel-eluting stents (19.3% vs 24.5%).26 At the 5-year follow-up of the COMPARE (Trial of Everolimus-Eluting Stents and Paclitaxel-Eluting Stents for Coronary Revascularization in Daily Practice) trial, all comers treated with the EES (Xience V) showed significantly lower rates of various composite end points than patients treated with paclitaxel-eluting stents, including TVF (12.6% vs 17.8%) and definite or probable stent thrombosis (3.1% vs 5.9%).27 In the 5-year results of the SORT OUT IV (Danish Organization of Randomized Trials with Clinical Outcome) trial, the composite end point of major adverse cardiac events and definite stent thrombosis was significantly lower in all-comer patients treated with the EES (Xience V) than in patients treated with early-generation sirolimus-eluting stents (14.0% vs 17.4% and 0.4% vs 2.0%, respectively).28 Moreover, in the EXAMINATION (Clinical Evaluation of the Xience V Stent in Acute Myocardial Infarction) trial, the use of EES in patients with acute ST-segment elevation MI resulted in a lower 5-year mortality than treatment with bare-metal stents (8.7% vs 11.8%).29 The randomized RESOLUTE AC trial has reported 5-year outcome data, showing similar efficacy and safety of ZES (Resolute) and EES in an all-comer population (rate of TVF, 20.0% vs 19.1%).15 The outcome of the randomized TWENTE trial supports these findings in general. The exploratory subgroup analysis for the main end point of TVF also showed consistent results, with single-vessel treatment being the only exception (favoring ZES), which may most likely reflect a play of chance.

Limitations

Scientific evidence from multicenter trials is generally considered higher ranking than that from single-center trials. Our conclusions do not apply to patients with acute ST-segment elevation MI undergoing primary PCI because such patients were not studied. Patients with acute ST-segment elevation MI are of considerable interest for the assessment of novel stents,30 whereas their enrollment may be challenging. The subgroup analysis was not prespecified but assessed the same subgroups as the RESOLUTE AC trial.3 Results of comparisons between nonenrolled eligible patients and the trial population are hypothesis generating.

Conclusions

At the 5-year follow-up, the second-generation DESs used in the RCT showed favorable and similar long-term results in a broad patient population, most of whom were treated for non–ST-segment elevation acute coronary syndromes and complex coronary lesions. Outcome data from nonenrolled eligible patients support the validity of the randomized TWENTE trial and present, with the results of the randomized trial, a strong case for the long-term safety and efficacy of both devices. Moreover, these data underline the importance of aiming at high study enrollment.

Back to top
Article Information

Corresponding Author: Clemens von Birgelen, MD, PhD, Department of Cardiology, Thoraxcentrum Twente, Medisch Spectrum Twente, Koningsplein 1, 7512 KZ Enschede, the Netherlands (c.vonbirgelen@mst.nl).

Accepted for Publication: November 6, 2016.

Published Online: January 18, 2017. doi:10.1001/jamacardio.2016.5190

Author Contributions: Drs von Birgelen and van der Heijden contributed equally to this manuscript. Drs von Birgelen and van der Heijden had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: von Birgelen, van der Heijden, Basalus, Kok, Sen, van Houwelingen, Stoel, Tandjung, Löwik.

Acquisition, analysis, or interpretation of data: von Birgelen, van der Heijden, Basalus, Kok, Louwerenburg, van Houwelingen, de Man, Linssen, Doggen, van der Palen, Löwik.

Drafting of the manuscript: von Birgelen, van der Heijden.

Critical revision of the manuscript for important intellectual content: van der Heijden, Basalus, Kok, Sen, Louwerenburg, van Houwelingen, Stoel, de Man, Linssen, Tandjung, Doggen, van der Palen, Löwik.

Statistical analysis: van der Heijden, Kok, Doggen, van der Palen.

Obtained funding: von Birgelen.

Administrative, technical, or material support: van der Heijden, Sen, Louwerenburg, van Houwelingen, Stoel, Linssen, Löwik.

Study supervision: von Birgelen, van der Heijden, Basalus, Kok, van Houwelingen, de Man, Doggen.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr von Birgelen reports serving as an unpaid consultant to Boston Scientific and Medtronic; receiving a speaker’s honorarium from AstraZeneca and Biotronik; and serving as the principal investigator of the TWENTE trials. No other disclosures were reported.

Funding/Support: This study was supported by educational and/or research grants from AstraZeneca, Biotronik, Boston Scientific, and Medtronic (research department of Thoraxcentrum Twente); by equal, unrestricted research grants from Abbott Vascular and Medtronic (TWENTE trial to its 2-year follow-up); and by a research grant from Medtronic (TWENTE trial from the 3- to 5-year follow-up).

Role of the Funder/Sponsor: The study sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Renate Wiggers-van der Leest, CardioResearch Enschede, collected the TWENTE trial follow-up data with great thoroughness and dedication, for which she received no special compensation. We gratefully acknowledge the invaluable support of the study by the staff of the research department, catheterization laboratory, wards, and administrative department of Thoraxcentrum Twente. We thank our colleagues in the cooperating hospitals and the general physicians and pharmacists in the Twente region for their support and for treating the patients so well.

References
1.
Windecker  S, Serruys  PW, Wandel  S,  et al.  Biolimus-eluting stent with biodegradable polymer versus sirolimus-eluting stent with durable polymer for coronary revascularisation (LEADERS): a randomised non-inferiority trial.  Lancet. 2008;372(9644):1163-1173.PubMedGoogle ScholarCrossref
2.
Kedhi  E, Joesoef  KS, McFadden  E,  et al.  Second-generation everolimus-eluting and paclitaxel-eluting stents in real-life practice (COMPARE): a randomised trial.  Lancet. 2010;375(9710):201-209.PubMedGoogle ScholarCrossref
3.
Serruys  PW, Silber  S, Garg  S,  et al.  Comparison of zotarolimus-eluting and everolimus-eluting coronary stents.  N Engl J Med. 2010;363(2):136-146.PubMedGoogle ScholarCrossref
4.
Christiansen  EH, Jensen  LO, Thayssen  P,  et al; Scandinavian Organization for Randomized Trials with Clinical Outcome (SORT OUT) V Investigators.  Biolimus-eluting biodegradable polymer-coated stent vs durable polymer-coated sirolimus-eluting stent in unselected patients receiving percutaneous coronary intervention (SORT OUT V): a randomised non-inferiority trial [published correction appears in Lancet. 2013;382(9889):310].  Lancet. 2013;381(9867):661-669.PubMedGoogle ScholarCrossref
5.
Maeng  M, Tilsted  HH, Jensen  LO,  et al.  Differential clinical outcomes after 1 year versus 5 years in a randomised comparison of zotarolimus-eluting and sirolimus-eluting coronary stents (the SORT OUT III study): a multicentre, open-label, randomised superiority trial.  Lancet. 2014;383(9934):2047-2056.PubMedGoogle ScholarCrossref
6.
Bligaard  N, Thuesen  L, Saunamäki  K,  et al; SORT OUT II Investigators.  Similar five-year outcome with paclitaxel- and sirolimus-eluting coronary stents.  Scand Cardiovasc J. 2014;48(3):148-155.PubMedGoogle ScholarCrossref
7.
von Birgelen  C, Basalus  MW, Tandjung  K,  et al.  A randomized controlled trial in second-generation zotarolimus-eluting Resolute stents versus everolimus-eluting Xience V stents in real-world patients: the TWENTE trial.  J Am Coll Cardiol. 2012;59(15):1350-1361.PubMedGoogle ScholarCrossref
8.
von Birgelen  C, Sen  H, Lam  MK,  et al.  Third-generation zotarolimus-eluting and everolimus-eluting stents in all-comer patients requiring a percutaneous coronary intervention (DUTCH PEERS): a randomised, single-blind, multicentre, non-inferiority trial.  Lancet. 2014;383(9915):413-423.PubMedGoogle ScholarCrossref
9.
Hordijk-Trion  M, Lenzen  M, Wijns  W,  et al; EHS-CR Investigators.  Patients enrolled in coronary intervention trials are not representative of patients in clinical practice: results from the Euro Heart Survey on Coronary Revascularization.  Eur Heart J. 2006;27(6):671-678.PubMedGoogle ScholarCrossref
10.
Hutchinson-Jaffe  AB, Goodman  SG, Yan  RT,  et al; Canadian Acute Coronary Syndromes (ACS) Registry I and II Investigators and Canadian Global Registry of Acute Coronary Events (GRACE/GRACE 2) Investigators.  Comparison of baseline characteristics, management and outcome of patients with non–ST-segment elevation acute coronary syndrome in versus not in clinical trials.  Am J Cardiol. 2010;106(10):1389-1396.PubMedGoogle ScholarCrossref
11.
Rothwell  PM.  External validity of randomised controlled trials: “to whom do the results of this trial apply”?  Lancet. 2005;365(9453):82-93.PubMedGoogle ScholarCrossref
12.
de Boer  SPM, Lenzen  MJ, Oemrawsingh  RM,  et al.  Valuating the “all-comers” design: a comparison of participants in two “all-comers” PCI trials with non-participants.  Eur Heart J. 2011;32(17):2161-2167.PubMedGoogle ScholarCrossref
13.
Sen  H, Tandjung  K, Basalus  MW,  et al.  Comparison of eligible non-enrolled patients and the randomised TWENTE trial population treated with Resolute and Xience V drug-eluting stents.  EuroIntervention. 2012;8(6):664-671.PubMedGoogle ScholarCrossref
14.
Meredith  IT, Worthley  S, Whitbourn  R,  et al; RESOLUTE Investigators.  Clinical and angiographic results with the next-generation resolute stent system: a prospective, multicenter, first-in-human trial.  JACC Cardiovasc Interv. 2009;2(10):977-985.PubMedGoogle ScholarCrossref
15.
Iqbal  J, Serruys  PW, Silber  S,  et al.  Comparison of zotarolimus- and everolimus-eluting coronary stents: final 5-year report of the RESOLUTE All-Comers trial.  Circ Cardiovasc Interv. 2015;8(6):e002230.PubMedGoogle ScholarCrossref
16.
World Medical Association.  World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects.  JAMA. 2013;310(20):2191-2194.PubMedGoogle ScholarCrossref
17.
Tandjung  K, Sen  H, Lam  MK,  et al.  Clinical outcome following stringent discontinuation of dual antiplatelet therapy after 12 months in real-world patients treated with second-generation zotarolimus-eluting Resolute and everolimus-eluting Xience V stents: 2-year follow-up of the randomized TWENTE trial.  J Am Coll Cardiol. 2013;61(24):2406-2416.PubMedGoogle ScholarCrossref
18.
Löwik  MM, Lam  MK, Sen  H,  et al.  Safety of second-generation drug-eluting stents three years after randomised use in the TWENTE trial.  EuroIntervention. 2015;10(11):1276-1279.PubMedGoogle ScholarCrossref
19.
Cutlip  DE, Windecker  S, Mehran  R,  et al; Academic Research Consortium.  Clinical end points in coronary stent trials: a case for standardized definitions.  Circulation. 2007;115(17):2344-2351.PubMedGoogle ScholarCrossref
20.
Vranckx  P, Cutlip  DE, Mehran  R,  et al.  Myocardial infarction adjudication in contemporary all-comer stent trials: balancing sensitivity and specificity: addendum to the historical MI definitions used in stent studies.  EuroIntervention. 2010;5(7):871-874.PubMedGoogle ScholarCrossref
21.
Mehilli  J, Richardt  G, Valgimigli  M,  et al; ISAR-LEFT-MAIN 2 Study Investigators.  Zotarolimus- vs everolimus-eluting stents for unprotected left main coronary artery disease.  J Am Coll Cardiol. 2013;62(22):2075-2082.PubMedGoogle ScholarCrossref
22.
Park  KW, Kang  SH, Kang  HJ,  et al; HOST–ASSURE Investigators.  A randomized comparison of platinum chromium-based everolimus-eluting stents vs cobalt chromium-based zotarolimus-eluting stents in all-comers receiving percutaneous coronary intervention: HOST-ASSURE, a randomized, controlled, noninferiority trial.  J Am Coll Cardiol. 2014;63(25, pt A):2805-2816.PubMedGoogle ScholarCrossref
23.
Navarese  EP, Tandjung  K, Claessen  B,  et al.  Safety and efficacy outcomes of first and second generation durable polymer drug eluting stents and biodegradable polymer biolimus eluting stents in clinical practice: comprehensive network meta-analysis.  BMJ. 2013;347:f6530.PubMedGoogle ScholarCrossref
24.
Palmerini  T, Benedetto  U, Biondi-Zoccai  G,  et al.  Long-term safety of drug-eluting and bare-metal stents: evidence from a comprehensive network meta-analysis.  J Am Coll Cardiol. 2015;65(23):2496-2507.PubMedGoogle ScholarCrossref
25.
Piccolo  R, Stefanini  GG, Franzone  A,  et al.  Safety and efficacy of resolute zotarolimus-eluting stents compared with everolimus-eluting stents: a meta-analysis.  Circ Cardiovasc Interv. 2015;8(4):e002223.PubMedGoogle ScholarCrossref
26.
Gada  H, Kirtane  AJ, Newman  W,  et al.  5-year results of a randomized comparison of XIENCE V everolimus-eluting and TAXUS paclitaxel-eluting stents: final results from the SPIRIT III trial (clinical evaluation of the XIENCE V everolimus eluting coronary stent system in the treatment of patients with de novo native coronary artery lesions).  JACC Cardiovasc Interv. 2013;6(12):1263-1266.PubMedGoogle ScholarCrossref
27.
Smits  PC, Vlachojannis  GJ, McFadden  EP,  et al.  Final 5-year follow-up of a randomized controlled trial of everolimus- and paclitaxel eluting stents for coronary revascularization in daily practice: the COMPARE trial (A Trial of Everolimus-Eluting Stents and Paclitaxel-Eluting Stents for Coronary Revascularization in Daily Practice).  JACC Cardiovasc Interv. 2015;8(9):1157-1165.PubMedGoogle ScholarCrossref
28.
Jensen  LO, Thayssen  P, Christiansen  EH,  et al; SORT OUT IV Investigators.  Safety and efficacy outcomes of everolimus- vs sirolimus-eluting stents: the SORT OUT IV 5-year results.  J Am Coll Cardiol. 2016;67(7):751-762.PubMedGoogle ScholarCrossref
29.
Sabaté  M, Brugaletta  S, Cequier  A,  et al.  Clinical outcomes in patients with ST–segment elevation myocardial infarction treated with everolimus-eluting stents versus bare-metal stents (EXAMINATION): 5-year results of a randomised trial.  Lancet. 2016;387(10016):357-366.PubMedGoogle ScholarCrossref
30.
von Birgelen  C, Kok  MM, van der Heijden  LC,  et al.  Very thin strut biodegradable polymer everolimus-eluting and sirolimus-eluting stents versus durable polymer zotarolimus-eluting stents in allcomers with coronary artery disease (BIO-RESORT): a three-arm, randomised, non-inferiority trial.  Lancet. 2016;388(10060):2607-2617.PubMedGoogle ScholarCrossref
×