Author + information
- Received September 27, 2004
- Revision received December 9, 2004
- Accepted December 20, 2004
- Published online August 16, 2005.
- Patrick W. Serruys, MD, PhD, FACC⁎,⁎ (, )
- Andrew T.L. Ong, MBBS, FRACP⁎,
- Lex A. van Herwerden, MD, PhD⁎,
- J. Eduardo Sousa, MD, PhD, FACC†,
- Adib Jatene, MD‡,
- Johannes J.R.M. Bonnier, MD, PhD§,
- Jacques P.M.A. Schönberger, MD, PhD§,
- Nigel Buller, MBBS, FRCP∥,
- Robert Bonser, MBChB, FRCP, FRCS∥,
- Clemens Disco, MSc¶,
- Bianca Backx, PhD¶,
- Paul G. Hugenholtz, MD, FACC¶,
- Brian G. Firth, MD, PhD, FACC# and
- Felix Unger, MD, FACC**
- ↵⁎Reprint requests and correspondence:
Prof. Patrick W. Serruys, Head of the Interventional Cardiology Department, Ba 583, Thoraxcenter, Erasmus Medical Center Rotterdam, Dr. Molewaterplein 40, 3015 GD, Rotterdam, the Netherlands
Objectives The long-term (five-year) comparative results of treatment of multivessel coronary artery disease with stenting or coronary artery bypass grafting (CABG) is at present unknown.
Background The Arterial Revascularization Therapies Study (ARTS) was designed to compare CABG and stenting in patients with multivessel disease.
Methods A total of 1,205 patients with the potential for equivalent revascularization were randomly assigned to CABG (n = 605) or stent implantation (n = 600). The primary clinical end point was freedom from major adverse cardiac and cerebrovascular events (MACCE) at one year; MACCE at five-year follow-up constituted the final secondary end point.
Results At five years, there were 48 and 46 deaths in the stent and CABG groups, respectively (8.0% vs. 7.6%; p = 0.83; relative risk [RR], 1.05; 95% confidence interval [CI], 0.71 to 1.55). Among 208 diabetic patients, mortality was 13.4% in the stent group and 8.3% in the CABG group (p = 0.27; RR, 1.61; 95% CI, 0.71 to 3.63). Overall freedom from death, stroke, or myocardial infarction was not significantly different between groups (18.2% in the stent group vs. 14.9% in the surgical group; p = 0.14; RR, 1.22; 95% CI, 0.95 to 1.58). The incidence of repeat revascularization was significantly higher in the stent group (30.3%) than in the CABG group (8.8%; p < 0.001; RR, 3.46;95% CI, 2.61 to 4.60). The composite event-free survival rate was 58.3% in the stent group and 78.2% in the CABG group (p < 0.0001; RR, 1.91;95% CI, 1.60 to 2.28).
Conclusions At five years there was no difference in mortality between stenting and surgery for multivessel disease. Furthermore, the incidence of stroke or myocardial infarction was not significantly different between the two groups. However, overall MACCE was higher in the stent group, driven by the increased need for repeat revascularization.
A meta-analysis including nine trials of multivessel coronary artery disease treated by percutaneous balloon angioplasty alone or coronary artery bypass grafting (CABG) showed a statistically significant benefit in terms of survival in favor of surgery at five and eight years (1). However, these survival data were from early studies that did not use stents in the initial revascularization procedure. The Stent or Surgery Trial (SoS), which involved the use of stents, reported similar findings after a median follow-up of two years (2). However, the Argentine Randomized Trial: Coronary Angioplasty with Stenting versus Coronary Bypass Surgery With Multivessel Disease (ERACI-II) suggested that the trend in favor of CABG for survival at 2.5 years was no longer present in the stent era (3).
There are currently no data available on the comparative survival after multivessel stenting or CABG beyond three years. The present study reports on the five-year survival and event-free survival of the patients enrolled in the Arterial Revascularization Therapies Study (ARTS) trial (4).
The study protocol, summarized here, has been previously published (4,5).
Between April 1997 and June 1998, 1,205 patients from 67 participating centers were randomized to either stent implantation (n = 600) or CABG (n = 605). The study population included 208 diabetic patients. The indications for revascularization included silent ischemia, stable or unstable angina pectoris, and the presence of at least two de novo lesions located in different major epicardial coronary arteries, potentially amenable to stent implantation. For each patient, entry into the study required agreement from both surgeon and interventional cardiologist that an equivalent degree of revascularization could potentially be obtained using either approach.
Specific exclusion criteria from the randomized trial may be summarized as follows: left ventricular ejection fraction <30%, left main stenosis, history of a cerebrovascular accident, transmural myocardial infarction within the preceding week, and severe hepatic or renal disease and need for concomitant major surgery. All patients gave written informed consent.
Five-year clinical follow-up
The study protocol required all patients to have follow-up clinic visits with an electrocardiogram at one, two, and three years. In addition, at the five-year clinical follow-up, anginal status and use of medications were assessed. Additional information was obtained by telephone interview or via the referring physician when needed. An independent committee adjudicated clinical events and electrocardiograms.
Pre-specified analyses were performed on diabetics versus non-diabetics and two- versus three-vessel disease. In addition, post-hoc analyses were performed on the following subgroups: proximal left anterior descending versus non-proximal left anterior descending lesions, renal status, gender, and age.
Clinical end points and effectiveness
The primary end point was defined as the absence of any of the following major adverse cardiac and cerebral events (MACCE) within 12 months after randomization: death (all-cause mortality), cerebrovascular accident, documented non-fatal myocardial infarction adjudicated by either new abnormal Q-wave or predefined enzymatic changes, or repeat revascularization by coronary stenting or CABG (4,5).
Secondary objectives of the study were to compare both strategies at three and five years. The MACCE were counted from the time of randomization, whereas the clinical status and medications were assessed at predetermined times of one, two, three, and five years post-procedure. Of 1,205 patients enrolled in the trial, complete follow-up was available at five years in 590 of 600 (98.3%) stent patients and 584 of 605 (96.6%) CABG patients (Fig. 1).
Statistical analysis was performed with SAS 6.12 software (SAS Institute Inc., Chicago, Illinois). Binary outcome variables are reported as frequencies and percentages and were compared in terms of relative risk with 95% confidence intervals calculated by the formula of Greenland and Robins (6). The Fisher exact test was used for categorical variables. All analyses were based on the intention-to-treat principle, and statistical tests were two-tailed. Event-free survival was estimated using the Kaplan-Meier method, and differences were assessed using the log-rank test. The sample size calculation to achieve adequate power for an inferiority study was based on the difference in event-free survival at one year (4). For this five-year report on late outcomes, no new calculations were performed.
Table 1shows the baseline and procedural characteristics of the ARTS trial's randomized patients. The randomized groups were similar with respect to their demographic and anatomic characteristics. Five patients, one assigned to stenting and four assigned to surgery, did not undergo coronary revascularization and instead continued to receive only medical therapy (4). The average interval between randomization and treatment was 27 ± 39 days (range, 0 to 362 days) for patients in the surgery group and 11 ± 16 days (range, 0 to 173 days) for patients in the stenting group. Three patients died while waiting for surgery, 6 patients randomly assigned to stent implantation were instead treated surgically, and 19 patients randomly assigned to bypass surgery were instead treated with stent implantation. A total of 99% of patients in the stenting group (593 patients) and 93% in the surgery group (579 patients) received the assigned treatment. An equivalent anatomical degree of revascularization was achieved in each group.
During the initial hospital stay, after complicated or unsatisfactory angioplasty procedures, 14 patients assigned to stent implantation underwent bypass surgery, 3 urgently and 11 electively. Conversely, two patients underwent an angioplasty procedure after surgical revascularization during their initial hospital stay (Fig. 1).
Five-year clinical outcome
At five years, there were 48 deaths in the stent group and 46 deaths in the surgical group, which represents 8.0% and 7.6% of the respective cohorts (p = 0.83; relative risk [RR], 1.05; 95% confidence interval [CI], 0.71 to 1.55) (Table 2, Fig. 2). The incidence of cardiac death was not significantly different between the groups (Table 3).Of the 94 deaths, 6 occurred within 30 days after a repeat revascularization procedure.
The incidence of death, stroke, or myocardial infarction was not significantly different among the groups (18.2% in the stent group vs. 14.9% in the surgical group; p = 0.14; RR, 1.22; 95% CI, 0.95 to 1.58).
The incidence of repeat revascularization was significantly higher in the stent group (30.3%) than in the CABG group (8.8%; p < 0.001; RR, 3.46; 95% CI, 2.61 to 4.60). At the end of five years, 10.5% of patients originally assigned to stenting required CABG and 23.2% underwent a second percutaneous coronary intervention (PCI). Conversely, 1.2% of patients assigned to CABG required re-operation and 8.3% required revascularization with PCI. In the stent group, the majority of re-interventions (69%) took place within the first year, whereas in the CABG group, the majority of the re-interventions (57%) occurred after the first year. The overall MACCE-free survival at five years was 58.3% in the stent group and 78.2% in the CABG group (p < 0.0001).
At five years, there was a significant difference in the presence of anginal symptoms between the two treatment groups (21.2% of the stent patients vs. 15.5% of the CABG patients, p < 0.05). More of the stent patients were on short-acting nitrates (6.1% vs. 2.4%, p = 0.003), long-acting nitrates (19.6% vs. 11.6%, p < 0.001), beta-blocker therapy (53.9% vs. 46.5%, p = 0.016), and calcium-channel antagonists (29.1% vs. 18.9%, p < 0.001).
Patients with diabetes
In patients with diabetes, those who underwent stenting had a mortality rate of 13.4%, versus 8.3% in those who underwent CABG (p = 0.27; RR, 1.61; 95% CI, 0.71 to 3.63) (Tables 4 and 5)⇓⇓. Within the stent group, diabetic patients had a significantly higher mortality rate than non-diabetic patients (13.4% vs. 6.8%; p = 0.03; RR, 1.98; 95% CI, 1.11 to 3.52). In stent diabetic patients, death was attributed to a cardiac cause in 50% of cases versus 38% (p = 0.43; RR, 1.32; 95% CI, 0.68 to 2.58) in non-diabetic stent patients. There was no significant mortality difference between the diabetic and non-diabetic patients within the CABG group (8.3% vs. 7.5%; p = 0.8; RR, 1.12; 95% CI, 0.54 to 2.32).
Diabetic patients treated with stenting also had a lower event-free survival at five years than non-diabetic patients. The MACCE rate at five years in diabetic patients treated with stents was 54.5%, versus 38.7% in non-diabetics (p = 0.003). Conversely, there was no significant difference in the five-year MACCE rate between diabetic and non-diabetic patients treated with CABG (25.0% vs. 21.2%, p = 0.42). The difference in MACCE rate between diabetic and non-diabetic patients treated with stenting is largely attributable to the higher rate of repeat revascularization in diabetic patients (42.9% vs. 27.5%, p = 0.002).
Two- versus three-vessel treatment
There was no significant difference in event-free survival rate between patients with two or three vessels treated with stenting (56.7% vs. 60.1%) or CABG (79.4% vs. 75.7%), respectively. However, the event-free survival rate was significantly higher for patients treated with CABG than with stenting for both two and three vessels (p < 0.001 and p = 0.001, respectively).
Other subgroup analyses
There were also no significant differences in event-free survival within the respective treatment groups based on renal function, gender, age, or hypercholesterolemia at the time of randomization. At five years, 65.9% in the stent group and 61.5% in the CABG group were on lipid-lowering agents. Outcomes in patients who were treated for proximal left anterior descending lesions or otherwise were not significantly different stratified by treatment group.
This is the first randomized trial to report on five-year outcomes of patients with multivessel coronary artery disease treated with bare metal stenting versus CABG. Although this study was not specifically powered to detect a difference in five-year mortality, there was no clinically relevant difference (p = 0.83) with these two forms of treatment. This contemporary finding differs from the meta-analysis of previous randomized trials of balloon angioplasty alone versus CABG conducted in patients with multivessel disease, which showed a significantly higher mortality rate with percutaneous treatment at five years (risk difference, 2.3%; 95% CI, 0.29 to 4.3%; p = 0.025) (1).
In this study, mortality in the CABG arm was 7.6% at five years, lower than the composite death rate of 8.9% seen in the CABG patients from the meta-analysis (1), evidence that improved peri-operative management and intra-operative techniques over time have resulted in a reduction in mortality. Similarly, mortality in the stent arm was 8.0%, a risk difference of 0.4% (95% CI, 1.1% to 1.9%; p = 0.83). From one to five years, the risk difference changes from 0.3% in favor of stenting at one year, to 0.9% in favor of stenting at three years, to 0.4% in favor of CABG at five years (all not significant), indicating a strong effect of chance. Furthermore, this difference is not clinically relevant and is much lower than the 2.3% of the meta-analysis.
There was a 3.3% absolute difference in the composite end point of death, stroke, and myocardial infarction in favor of CABG, primarily driven by a higher incidence of myocardial infarctions in the stent arm. Although suggestive, this study was underpowered to detect a significant difference in the end point. Based on this difference, a population of 4,000 patients would be required for statistical significance.
The risk difference for revascularization at five years, as reported in this same meta-analysis, was 38% (95% CI, 30% to 47%). Specifically, the risk difference for subsequent CABG was 24% and for subsequent percutaneous transluminal coronary angioplasty was 23%. The current observed differences in the ARTS trial for any revascularization at five years is 21.5%, for subsequent CABG is 9.3%, and for subsequent PCI is 14.9%. It is worth noting that almost 90% of patients initially treated with stenting did not require CABG over the succeeding five years. The difference in the rate of repeat revascularization between the two groups increases over time from 17.2% at 1 year to 21.5% at five years without a concomitant difference in mortality over this time period. Despite the additional risk of repeat revascularization in the stent group compared with the CABG group, this did not translate into an increase in mortality (Table 3).
In this study, four-fifths of all patients in both groups were free of anginal complaints at five years. Although significantly different, this high proportion of patients free of symptoms is encouraging in this population of patients with chronic multivessel coronary artery disease. Correspondingly, more stent patients than CABG patients were on anti-anginal medications (p < 0.001) at five-year follow-up.
In diabetic patients from three trials comparing balloon angioplasty with surgery, the risk difference for all death was 8.6% in favor of CABG (p = 0.01; 95% CI, 2.2% to 15%) (n = 537 patients) at four years (1). In the present study involving 208 diabetic patients, mortality at five years was 5.1% higher in stent patients compared with CABG patients (p = 0.27). Conversely, in non-diabetic patients the mortality rate was 0.7% lower in the stent cohort. However, the study was not powered to show mortality differences between diabetic and non-diabetic patients.
Repeat revascularization was higher in diabetic patients randomized to the stent arm versus CABG (an absolute difference of 32.5% [42.9% vs. 10.9%, respectively]), compared with non-diabetic patients (a 19.1% absolute difference [27.5% vs. 8.4%, respectively, both p < 0.001]). Based on the available evidence, surgery should continue to be viewed as the preferred therapy for diabetic patients with multivessel disease when using bare metal stents.
The advent of drug-eluting stents has drastically reduced the need for repeat revascularization in both diabetic and non-diabetic patients. The relative reduction in need for re-intervention with drug-eluting stents is very similar in diabetic and non-diabetic patients (7,8). The difference in outcomes seen between bare metal stents versus CABG for the treatment of multivessel disease is likely to narrow substantially with the advent of drug-eluting stents. The U.S. National Institutes of Health is sponsoring a large multicenter trial specifically to evaluate the difference in outcomes in diabetic patients with multivessel coronary disease treated with drug-eluting stents versus CABG. A European multicenter trial comparing drug-eluting stents versus CABG for the treatment of multivessel and left main stem coronary disease in an all-comers population is currently in progress.
Despite the increasing age and concomitant increased co-morbidity of patients presenting for CABG, clinical outcomes have continued to improve (9). This was evident from the lower mortality seen in the CABG arm of this study compared with the older studies. The off-pump coronary bypass technique, developed to minimize the invasiveness of CABG, has in several large retrospective studies suggested a reduction in morbidity and/or mortality when compared with CABG (10). Larger randomized trials are required to address this issue definitively because the three reported randomized prospective studies comparing off-pump coronary bypass with CABG were not large enough to detect a difference in operative mortality or stroke (10). Finally, the routine use of post-procedural medications—aspirin, statins, and control of risk factors—will further improve outcomes in both the CABG and the stent groups in future trials.
This study was supported by Cordis, a Johnson & Johnson company. Dr. Firth is an employee of Cordis. The other authors declare no conflict of interests. The corresponding author has had full access to all of the data in the study and has had the final responsibility for the decision to submit this manuscript for publication.
- Abbreviations and Acronyms
- Arterial Revascularization Therapies Study
- coronary artery bypass grafting
- confidence interval
- major adverse cardiac and cerebral event
- percutaneous coronary intervention
- relative risk
- Received September 27, 2004.
- Revision received December 9, 2004.
- Accepted December 20, 2004.
- American College of Cardiology Foundation
- Hoffman S.N.,
- TenBrook J.A.,
- Wolf M.P.,
- Pauker S.G.,
- Salem D.N.,
- Wong J.B.
- Stables R.H.
- Rodriguez A.,
- Bernardi V.,
- Navia J.,
- et al.
- Serruys P.W.,
- Van Hout B.A.,
- Van den Brand M.J.,
- et al.
- Ferguson T.B. Jr..,
- Hammill B.G.,
- Peterson E.D.,
- DeLong E.R.,
- Grover F.L.,
- for the STS National Database Committee
- ↵Eagle KA, Guyton RA, Davidoff R, et al. ACC/AHA 2004 guideline update for coronary artery bypass graft surgery: a report of the American College of Cardiology /American Heart Association Task Force on Practice Guidelines (Committee to Update the 1999 Guidelines for Coronary Artery Bypass Graft Surgery). Available at: http://www.acc.org/clinical/guidelines/cabg/cabg.pdf. Accessed November 1, 2004.