OUP user menu

The impact of the introduction of drug-eluting stents on the clinical practice of surgical and percutaneous treatment of coronary artery disease

Ron T. van Domburg, Pedro A. Lemos, Johanna J.M. Takkenberg, Tommy K.K. Liu, Lex A. van Herwerden, Chourmouzios A. Arampatzis, Pieter C. Smits, Joost Daemen, Angeliek C. Venema, Patrick W. Serruys, Ad J.J.C. Bogers
DOI: http://dx.doi.org/10.1093/eurheartj/ehi088 675-681 First published online: 6 January 2005

Abstract

Aims Sirolimus-eluting stents (SES) have recently been shown to reduce restenosis in selected patients. The impact of this new stent on the use of coronary bypass graft (CABG) surgery or percutaneous coronary intervention (PCI) in clinical practice is yet unknown. Therefore, we investigated the impact of SES on the clinical practice of CABG and PCI in a series of unselected consecutive patients.

Methods and results Between April and October 2002, a policy of SES implantation for all procedures has been instituted in our hospital. In total, 798 patients were referred to PCI and 275 to CABG (SES group). A control group was composed of all interventions (806 PCI and 314 CABG) performed during the preceding 6 months (pre-SES). The main outcome was the occurrence of major adverse cardiac events (MACE) at 15 months. In the SES era, a significant shift was noted in the PCI group towards more multi-vessel stenting (28 vs. 24%; P<0.05), more bifurcation stenting (18 vs. 7%; P<0.0001), and the use of more stents (1.9 vs. 1.5; P<0.05). In the PCI elective patients, a shift was noted towards more three-vessel disease (pre-SES: 16% vs. SES: 23%; P=0.02). Furthermore, we observed a shift in the CABG group towards more impaired LV function (pre-SES: 34% vs. SES: 41%; P=0.02) and towards more three-vessel disease (pre-SES: 67% vs. SES: 75%; P=0.03). Overall, the cumulative MACE percentages at 1 year after coronary revascularization (PCI and CABG combined) decreased from 16.8 to 13.8% (P=0.03). The cumulative MACE percentages in the pure SES group and the pre-SES bare metal stent group at 12 months were 15.6 and 19.8%, respectively (P<0.01).

Conclusion The introduction of the SES has certainly had an impact on the treatment strategy of coronary artery disease (CAD). Increased use of these stents allows more complex coronary anatomy to be treated by PCI, and results in lower repeat revascularization rates.

  • Drug-eluting stent
  • Coronary bypass graft
  • Percutaneous coronary intervention
  • Prognosis

Introduction

Although both coronary artery bypass graft (CABG) surgery and percutaneous coronary intervention (PCI) relieve symptoms of angina in patients with severe ischaemia, the most appropriate treatment remains a matter of debate for most of the patients. An important element of the ongoing discussion is the occurrence of restenosis after PCI, the main limitation to this treatment strategy. Sirolimus-eluting stent (SES) implantation has recently been demonstrated to markedly reduce the incidence of restenosis and repeat revascularization in selected patients.13 The SES was introduced in our hospital in April 2002.4,5 The impact of this treatment on the overall clinical management of revascularization is yet unknown. The aim of this study was to investigate the impact of the drug-eluting stent in clinical practice. Therefore, we evaluated all PCI and CABG procedures during the first 6 months of SES use, and in the 6 months preceding the introduction of SES in our clinic.

Methods

Study design and patient selection

Since 16 April 2002, a policy of SES implantation (Cypher; Johnson & Johnson-Cordis unit, Cordis Europe NV, Roden, The Netherlands) for all procedures in all suitable PCI candidates has been instituted in our hospital.4 Essentially, all patients with multivessel coronary artery disease [except patients with a preference for either surgery or PCI, patients with single vessel disease, and patients with impending myocardial infarction (MI)] are being discussed in the team discussion, irrespective of being presented to the cardiac surgeon or the intervention cardiologist. In our referral area, no other surgical or interventional facilities were available at the time of this investigation. In this regard non-cardiac risk factors for either treatment are evaluated and put into perspective with the risk–benefit ratio of either surgery or percutaneous intervention for the coronary disease. The decision is facilitated by both guidelines and experience. As a result, until 15 October 2002, 798 consecutive patients were referred to PCI (of whom 563 patients received only SES) and 275 to CABG (SES group). For comparison, a control group was composed of all interventions (806 PCI, of whom 718 with a bare stent, and 314 CABG) performed over the same period immediately prior to 16 April 2002 (pre-SES group). No patients were excluded. Therefore, the control and the SES groups are constituted by two sequential cohorts, primarily defined by the interventional strategy applied (conventional bare stent or SES implantation, respectively). The primary endpoint was the occurrence of major adverse cardiac events (MACE) at 15 months defined as death, post-discharge non-fatal MI, or all coronary revascularizations regardless of the site of restenosis. Rescue CABG or rescue PCI was defined as if occurred within 24 h after the index procedure. Peri-procedural MI was diagnosed by a rise in the creatine kinase (CK) level to more than twice the upper normal limit with an increased CK-MB.

Follow-up

In-hospital outcome information was obtained from an electronic clinical database for patients maintained in our hospital and by review of the hospital records for those discharged to secondary hospitals. During the follow-up, recordings of all repeat interventions (surgical and percutaneous) and re-hospitalizations were collected. Survival status at 6 months and 15 months was assessed by written inquiries to the municipal civil registries at 6 months and at 15 months. In 40% of the PCI patients [acute MI (AMI), left main stenting, bifurcation stenting, and chronic total occlusions] a scheduled angiographic control was performed at 6 months follow-up. Questionnaires were sent at 6 months and 1 year to all living patients, requesting the occurrence of clinical events. General practitioners and peripheral hospitals were directly approached whenever necessary for additional information. For patients who moved abroad, an effort was made to contact the local civil registries of their new residences by Internet and E-mail. Of five patients who moved abroad, we were unable to retrieve survival status and the follow-up was censored at the date of last patient contact. Follow-up was complete for all other patients (99.6%).

Statistical analysis

All consecutive procedures were included in the pre-SES period and in the SES period utilizing a dynamic registry design as previously described by Rothman and Greenland.6 The person-year that contributes to the index procedure ends when a first MACE occurs. Any eventual repeat PCI then acts as a new index procedure and the person-year contributes to the cohort. This design permits the inclusion of patients with in-stent restenosis in both study periods, allowing the evaluation of the impact of each particular re-intervention on the subsequent outcomes. Because the STENT and CABG populations are not comparable, analyses were performed on the CABG and PCI population separately. This type of analysis leads to lower statistical power and a higher type I error. Continuous variables were presented as mean±standard deviation (SD) and were compared using the Student's unpaired t-test. Categorical variables were presented as counts and percentages and compared with Fisher's exact test. The level of significance was a P-value less than 0.05 (two-tailed). The cumulative incidence of adverse cardiac events was estimated according to the Kaplan–Meier method. Among patient subgroups, the log-rank test was used to compare survival curves between the pre-SES and SES period. In order to investigate the impact of SES on the clinical practice of coronary intervention, we combined the STENT and CABG populations into one population in Figure 1.

Figure 1 (A) Cumulative death and post-discharge myocardial infarction percentages; (B) cumulative revascularization; and (C) cumulative major adverse cardiac events (death, post-discharge myocardial infarction, revascularization) after revascularization (PCI and CABG combined) according to the pre-SES and SES era 12 months after the index procedure.

The clinical profile (demographic and procedural details) was recorded during the index procedure. Pre-selected variables were: treated diabetes (NIDDM or IDDM), treated hypertension, treated hypercholesterolaemia, smoking habits, chronic obstructive pulmonary disease (pharmacologically treated), renal impairment (blood creatinine >150 µmol/L), family history, previous coronary interventions, prior MI, extent of vessel disease, and qualitative left ventricular (LV) function (normal if ≥50%, impaired if <50%).

In the PCI group a sub-analysis was performed according to patients who would have been eligible for inclusion in the drug-eluting stent randomized controlled trials (RCT) such as the SIRIUS study.3 The following criteria were used for exclusion from RCTs: a primary PCI for AMI, age older than 80 years, left main stenting, small vessels (≤2.5 mm), long stented length (>36 mm), restenotic lesions, bifurcation stenting, and stenting of the saphenous vein graft.

Ethical approval was obtained from the hospital ethics committee and the study was carried out in accordance with the Helsinki Declaration.

Results

CABG

The CABG patients were 64 years old and predominantly male (83%) (Table 1). Only a small minority had single vessel disease (6%) and left main disease was present in 19% of the patients. A shift was observed in the CABG group towards more three-vessel disease (pre-SES: 67% vs. SES: 75%; P=0.03) and towards more impaired LV function (pre-SES: 34% vs. SES: 41%; P=0.02). All other clinical characteristics were similar.

View this table:
Table 1

Baseline characteristics of 589 patients undergoing CABG and 1604 patients undergoing PCI (%)

Number of patientsCABGPCI
Pre-SESSESP-valuePre-SESSESP-value
314275806798
Age, years ±SD 64±10 64±110.8 61±10 62±110.5
Male, n (%)261 (83)228 (83)0.9580 (72)559 (70)0.4
Diabetes, n (%) 72 (23) 66 (24)0.9129 (16)136 (17)0.8
Hypertension, n (%)141 (45)132 (48)0.6306 (38)319 (40)0.3
Dyslipidemia, n (%)195 (62)182 (66)0.5451 (56)431 (54)0.8
Chronic obstructive pulmonary disease, n (%) 22 (7) 17 (6)0.5 40 (5) 32 (4)0.1
Renal impairment, n (%) 19 (6) 22 (8)0.2 24 (3) 24 (3)0.5
Family history, n (%) 50 (16) 74 (27)<0.001218 (27)231 (29)0.4
Previous CABG, n (%) 41 (13) 33 (12)0.4 97 (12) 96 (12)0.9
Previous PCI, n (%) 44 (14) 55 (20)0.05218 (27)207 (26)0.8
Previous MI, n (%)151 (48)127 (46)0.7322 (40)271 (34)0.03
Unstable anginaa119 (38)105 (38)0.9274 (34)297 (37)0.4
AMI  0  0129 (16)156 (20)0.5
Three-vessel disease,b n (%)210 (67)206 (75)0.03157 (20)179 (22)0.4
 In stable angina 66 (16) 78 (23)0.02
 In unstable angina 58 (21) 78 (26)0.1
 In AMI 33 (25) 23 (15)0.03
Left main disease 63 (20) 47 (17)0.4 32 (4) 32 (4)0.9
Impaired LVEF,c n (%)107 (34)113 (41)0.02137 (17)112 (14)0.7

aUnstable angina: according to Braunwald classification.

bTriple-vessel disease in the PCI group was split up according to patients who underwent elective angioplasty (stable angina), unstable angina and patients who underwent primary PCI for AMI.

cImpaired LVEF: left ventricular ejection fraction <50%.

CCS: Canadian Cardiovascular Society class.

The angiographic and procedural characteristics are shown in Table 2. Vein grafts without arterial grafts were used in only 6%. The left anterior descending (LAD) artery was grafted in 90%, mostly by using the left internal mammary artery and the mean number of distal anastomoses was 3.5.

View this table:
Table 2

Procedural characteristics of 589 patients undergoing CABG and 1604 patients undergoing PCI (%)

Pre-SESSESP-value
CABG
Number of patients314275
Treated vessel, n (%)
 RCA270 (86)245 (89)0.6
 LAD276 (88)253 (92)0.2
 LCX261 (83)237 (86)0.6
 LM 63 (20) 47 (17)0.3
Grafts, n (%)0.3
 Venous only 19 (6) 17 (6)
 Arterial only 41 (13) 19 (7)
 Venous and arterial257 (82)242 (88)
Grafts, n (%)
 LIMA292 (93)259 (94)0.6
 RIMA 38 (12) 14 (5)0.01
 Vein273 (87)256 (93)0.03
Number of distal anastomoses (mean±SD)3.4±1.13.5±1.00.5
PCI
Number of patients806798
Treated vessel, n (%)
 RCA303 (38)295 (37)0.7
 LAD431 (54)455 (57)0.5
 LCX271 (34)271 (34)0.9
 LM 32 (4) 32 (4)0.9
Pure SES stent, n (%)  0630 (79)
Bifurcation stenting, n (%) 56 (7)144 (18)0.0001
Multivessel stenting, n (%)192 (24)223 (28)0.03
Stent use, n (%)694 (87)766 (96)0.001
Number of implanted stents (mean±SD)1.5±0.51.9±0.60.03
Total stented length (mm) (mean±SD)29±2139±220.01
Brachy therapy, n (%) 40 (5) 16 (2)0.02
Glycoprotein IIbIIIa inhibitors, n (%)223 (28)120 (15)0.0001
Clopidogrel, n (%)734 (92)734 (92)0.9

LIMA: left internal mammary artery; RIMA: right internal mammary artery; RCA: right coronary artery; LAD: left anterior descending; LCX: left circumflex artery; LM: left main.

PCI

The PCI patients were 62 years old, 71% were male (Table 1). Left main disease was rare in the PCI group (4%). No shift in impaired LV function was observed from pre-SES to SES (17 and 14%, respectively). Also, no shift was observed in the incidence of three-vessel disease in the total PCI group (21% in both groups). However, in those PCI patients who underwent an elective intervention we also noted a shift towards more three-vessel disease (pre-SES: 16% vs. SES: 23%; P=0.02). All other clinical characteristics were similar in CABG and PCI except for more prior MI in the pre-SES PCI group as compared with the SES PCI group (40 vs. 34%). In the PCI group. Eighteen per cent of the patients underwent direct coronary angioplasty for acute MI.

SES were solely implanted in 79% (pure-SES group). In the remaining (21%), another type of balloon (31%) or bare stent (69%) was utilized due to unavailability of an appropriate, especially larger sized stent. SES patients had more bifurcation stenting (18 vs. 7%; P<0.0001), more multivessel stenting (28 vs. 24%; P<0.05), and more stents used (1.9 vs. 1.5; P<0.05) compared with the pre-SES PCI group. This increase of bifurcation stenting in the SES period did not result in higher in-hospital complications (death, MI, or revascularization) (pre-SES: 3.8% vs. SES: 2.0%) as might be expected and also the increase of multivessel stenting did not lead to more in-hospital complications (pre-SES: 3.0% vs. SES: 2.1%). The mean total length of stents implanted per patient increased from 29 to 39 mm (P<0.05) and stent usage increased from 87 to 96%. Unexpectedly, the use of peri-procedural glycoprotein IIb/IIIa inhibitors during PCI decreased from 28 to 15% (P<0.0001).

Thirty-days outcome

CABG

CABG, the incidence of major adverse cardiac events in the first 30 days was similar between the pre-SES and SES periods (Table 3). Early mortality rate was 1.0% (n=6) and MI in the first 30 days occurred in 16 patients. Early re-CABG was necessary in three patients and early recanalization with PCI post-CABG was performed in eight patients.

View this table:
Table 3

Thirty-day and 12-month outcomes of 589 patients undergoing CABG and 1604 patients undergoing PCI (%)

CABGPCI
Pre-SESSESPre-SESSES
(n=314)(n=275)(n=806)(n=798)
Thirty days
Death, n (%)3 (0.9)3 (1.1)13 (1.6)15 (1.9)
Myocardial infarction, n (%)9 (3.0)7 (2.6)10 (1.2)11 (1.4)
Elective re-CABG, n (%)1 (0.3)1 (0.3)5 (0.6) 0
Rescue (re-CABG), n (%)2 (0.6)08 (1.0)7 (0.9)
Elective (re-)PCI, n (%)5 (1.5)3 (1.1)24 (3.0)7 (0.9)
Rescue (re-)PCI, n (%) 004 (0.5)2 (0.2)
Twelve months
Death, n (%)11 (3.5)6 (2.2)31 (3.9)33 (4.1)
MI (post-discharge), n (%)2 (0.6)2 (0.7)8 (1.0)15 (1.9)
(Re-)CABG, n (%)4 (1.3)1 (0.3)40 (5.0)14 (1.8)
(Re-)PCI, n (%)2 (0.6)2 (0.7)8 (1.0)15 (1.9)

Rescue=same day.

PCI

Twenty-eight patients (1.7%) died before discharge (24 patients presented with AMI of whom 14 with cardiogenic shock). Within 30 days 21 patients (1.3%) underwent a MI. Early coronary bypass surgery in the pre-SES period was performed in 1.6% (rescue in 1.0%) and 0.9% (all rescue) in the SES. Early re-PCI was necessary in 3.5% (rescue in 0.5%) and 1.1% (rescue in 0.2%) in the pre-SES and SES periods, respectively (P=0.002). Of the 37 PCI patients (16 AMI) who underwent an early re-PCI within 30 days, the majority (25 patients, 67%) was planned as staged procedures. In eight patients (sub-)acute stent thrombosis was the reason for re-PCI and four patients underwent a direct re-PCI for a recurrent infarct.

One-year outcome

At 12 months there was no difference in mortality rates and in MI rates. As a result, the cumulative death or MI percentages were similar (5%) after revascularization (PCI and CABG combined) (Figure 1A). However, the cumulative revascularization percentages decreased from 13.5 to 9.7% (P=0.01) (Figure 1B). Overall, the cumulative rate of MACE after revascularization (PCI and CABG combined) decreased from 16.8 to 13.8% (P=0.03) (Figure 1C). This was due to the introduction of the SES: the cumulative MACE percentages in the pre-SES bare metal stent group and the pure SES group at 12 months were 19.8 and 15.6%, respectively (P=0.03) (Figure 2). The cumulative incidence of MACE percentages after CABG in the pre-SES and SES period also showed a slight decrease (pre-SES: 7.4% and SES: 5.9%; P=NS). The increase of impaired LV function and multivessel disease in the CABG group did not lead to more cardiac events at 12 months. In particular, the 12 months cumulative mortality rates were 3.1% in the pre-SES period and 3.8% in the SES period (P=0.8), and MACE percentages were respectively 5.1 and 4.8% (P=0.5). Also, in CABG patients with three-vessel disease, the cardiac event rates were similar in the pre-SES and SES periods.

Figure 2 (A) Cumulative revascularization percentages, and (B) cumulative one-year adverse events (death, post-discharge myocardial infarction, revascularization) after CABG and PCI according to the bare-stent (implantation of a bare metal stent in the pre-SES period), pure-stent [only SES(s) implanted without bare metal stents], pre-SES CABG and SES CABG.

One-year outcome according to eligibility for RCT

More than 70% of our patients (n=1157) would not have been eligible for inclusion in the randomized controlled clinical trials. However, also in this patient population, a benefit of SES was observed in less MACE (pre-SES: 20.2%, SES: 14.4%; P=0.02) (Figure 3A). In those patients who would have been eligible for inclusion in the clinical trials, SES patients had less repeat revascularizations.

Figure 3 One-year adverse events (death, post-discharge myocardial infarction, revascularization) after PCI according to patients who (A) would not have been eligible for inclusion in randomized controlled trials (RCT) in the pre-SES and SES era, and (B) would have been eligible for inclusion in randomized controlled trials in the pre-SES and SES era.

Discussion

We investigated the impact of the SES on the clinical management of all candidates for revascularization in a large consecutive single centre series. To the best of our knowledge, this is the first study that addresses that topic. The results demonstrate that more cases that were hitherto the domain of the surgeons are now being approached by percutaneous means. Interventional cardiologists are now more confident implanting a drug-eluting stent. In the first 2 months (the first tertile) after the introduction of the drug-eluting stent in our institution the percentage of elective patients with three-vessel disease was similar to those in the same time span of the bare stent period (15 vs. 14%). However, from then on the percentage of elective patients with three-vessel disease increased rapidly (SES: 27% vs. pre-SES: 17%). In the pre-drug-eluting stent era, patients with three-vessel disease were predominantly referred to the surgeons because of the additive risk of having at least one restenosis once several lesions were treated and the probability of repeat procedures was quite high and prohibitive. Once restenosis was reduced to very low levels, as in the drug-eluting stent era, the problem is greatly diminished and operators may start to be more liberal. Compared with the pre-SES era, this resulted in more multivessel dilatations and more bifurcation stenting being performed, more stents were implanted per patient, and in almost all angioplasty procedures, a stent was implanted without an increase in in-hospital complications as might be expected.7 Furthermore, in the elective PCI group, a shift was observed towards more three-vessel disease. An increasing number of patients will be treated with drug-eluting stents and although the problem of restenosis after stenting is not solved yet, it has reduced significantly in our clinical practice. This will all have an impact on both coronary angioplasty and coronary artery bypass surgery. In the CABG population we already observed a shift towards treatment of more three-vessel disease and more patients with an impaired LV function. In addition, in those patients who underwent elective PCI, a shift was noted towards more three-vessel disease in the SES group vs. the pre-SES group. However, due to the waiting lists in The Netherlands for both elective bypass surgery and elective percutaneous intervention, we were not able to investigate the impact of the drug-eluting stent on volume changes. The reduction in the bypass surgery group from 314 patients (pre-SES) to 275 patients (SES) was due to logistic management changes, which resulted in temporary reduction in surgical capacity. However, no patients were referred for PCI because of this temporary reduction. Although beyond the scope of this study, the better clinical outcome with SES relates only to a reduction in repeat PCI at the expense of significant increase of procedure-related costs. Thus, taking into account the present high costs of drug-eluting stents, only a shift of patients from surgery toward PCI could encompass the budgetary burden of SES PCI. On the other hand, the price of CABG is stable and unlikely to drop whereas, with the availability of more drug-eluting stents, it is expected that the price of SES will drop. Future cost-effectiveness evaluation should therefore take into account modification in revascularization techniques favoured by the introduction of SES.

In this study, all patients were included who were referred to our institution for either coronary bypass surgery or stenting without any exclusion. Nevertheless, patients indicated with acute MI are not candidates for primary bypass surgery, but only for primary angioplasty, whereas patients with left main disease or those with extensive complex lesions will primarily be sent to surgery. However, after exclusion of these patients the findings remained unchanged.

This study was not a randomized trial but rather an observational registry of real life clinical practice because the choice of the best treatment was completely left to the clinician's decision. We estimate that around 70% of our PCI population would not have been enrolled in the RAVEL or SIRIUS trials.2,3 However, the results of our study and of other observational studies have been shown to be complementary to randomized controlled trials.8,9 This study shows that in real life, patients with de novo lesions do indeed benefit from SES utilization. Also most patients with de novo lesions who would not be candidates for randomized controlled trials had benefited from the use of SES. For patients with complex in-stent restenosis (ISR), SES implantation was associated with similar outcomes compared with brachytherapy in a non-randomized study from our institution.10 However, the benefit of SES in patients with ISR remains to be proven in randomized controlled trials. In our opinion, coronary artery bypass surgery can no longer be compared with coronary angioplasty. The ARTS study was the latest randomized trial in a highly selected patient population comparing both modalities.11 ARTS showed a 13% absolute reduction in the need for revascularizations in the bypass surgery arm as compared with the stent arm. Clinical practice is running ahead of randomized trials and ongoing randomized trials are easily outdated. To date, our study shows that clinical practice is changing: once restenosis is reduced to very low levels, interventional cardiologists are more daring to apply more complex procedures such as in patients with three-vessel disease or left main disease. Only patients with complex lesions such as diffuse coronary narrowing, small arteries, total occluded vessels and perhaps in-restenotic lesions still undergo bypass surgery. However, this should be tested in true randomized trials of drug-eluting stents vs. CABG in order to establish if this trend is appropriate or not. Until then, the majority of the patients with triple vessel disease and left main disease will continue to undergo coronary bypass surgery. Ultimately, instead of being competitive, CABG and stenting will grow to be complementary to each other.

Conclusions

The significant reduction of coronary re-interventions after SES implantation will certainly have an impact on the invasive treatment of coronary artery disease. Increased use of these stents allows more complex coronary anatomy to be treated by PCI, and results in lower repeat revascularization rates.

Acknowledgements

Study supported by the Erasmus Medical Centre, Rotterdam, The Netherlands, and by an unrestricted institutional grant from Cordis, a Johnson & Johnson Company, Miami Lakes, FL, USA.

References

View Abstract