OUP user menu

Randomized comparison of everolimus- and paclitaxel-eluting stents: pooled analysis of the 2-year clinical follow-up from the SPIRIT II and III trials

Yoshinobu Onuma, Patrick W. Serruys, Neville Kukreja, Susan Veldhof, Julie Doostzadeh, Sherry Cao, Gregg W. Stone,
DOI: http://dx.doi.org/10.1093/eurheartj/ehp599 1071-1078 First published online: 29 January 2010


Aims To investigate the clinical impact of the following observations in the randomized SPIRIT II and III trials: an incremental increase in in-stent neointima between 1 and 2 years with the everolimus-eluting stent (EES) but not with the paclitaxel-eluting stent (PES) in SPIRIT II; a tendency of lower stent thrombosis in EES than in PES among those who first discontinued a thienopyridine after 6 months.

Methods and results A pooled analysis was performed using the 2-year clinical data from the SPIRIT II and III trials randomizing a total of 1302 patients with de novo coronary artery lesions either to EES or to PES. Inclusion and exclusion criteria were comparable between two trials. Major adverse cardiac event (MACE) was defined as cardiac death, myocardial infarction, or ischaemia-driven target lesion revascularization (TLR). At 2 years, MACE rates were 7.1% in EES vs. 12.3% in PES, respectively (log-rank P = 0.0014), without late increase in TLR. Among those who first discontinued a thienopyridine after 6 months, Academic Research Consortium (ARC) definite or probable stent thrombosis was 1.1% in EES vs. 1.3% in PES (P = 1.00).

Conclusion The benefits of EES in reducing TLR were robust between 6 months and 2 years. No significant difference in the thrombosis rate among those who first stopped a thienopyridine after 6 months was observed.

  • Everolimus-eluting stent
  • Paclitaxel-eluting stent
  • Randomized trial


Polymer-based sirolimus-eluting stents (SESs) and paclitaxel-eluting stents (PESs) have both been shown to significantly reduce angiographic restenosis and recurrent ischaemia necessitating repeat revascularization, compared with bare-metal stents.1 However, the occurrence of late stent thrombosis is a concern of this technology.25

Everolimus is an anti-proliferative agent that inhibits cell proliferation by inducing cell cycle arrest in the late G1 stage of the cell cycle.6 It is used as immunosuppressive therapy following heart and other solid organ transplantation, and has been shown to delay cardiac allograft vasculopathy.7 With the goal of further enhancing the safety and efficacy of drug-eluting stent (DES), an everolimus-eluting stent (EES) has been designed in which the anti-proliferative agent is released from a thin (7.8 µm), non-adhesive, durable, biocompatible fluoropolymer coated onto a low profile [0.0032 in. (81.3 µm) strut thickness], flexible cobalt chromium stent. Pre-clinical studies have shown more rapid endothelialization and reduced expression of platelet-endothelial cell adhesion molecule-1 and increased secretion and mRNA levels of vascular endothelial growth factor at 14 days with EES compared with both SES and PES.8

The clinical efficacy of EES has been tested in the several randomized trials.9 The SPIRIT II trial (n = 300) demonstrated not only non-inferior but also superior in-stent late loss at 6 months with EES compared with PES1012 The subsequent SPIRIT III trial (n = 1002)13 demonstrated a significant reduction in the primary angiographic endpoint of in-segment late loss with EES compared with PES at 8 months and non-inferiority to PES for the clinical endpoint of target vessel failure [TVF; cardiac death, myocardial infarction (MI) or ischaemia-driven target vessel revascularization (ID-TVR)] at 1 year and resulted in a significant reduction in major adverse cardiac event (MACE).14 Nevertheless, certain issues remain unclear. Although angiographic late loss in the diabetic population was less with EES than with PES at 6 and 8 months in the SPIRIT II and III trials, respectively, the TVF or MACE rates were not reduced with EES.13 Secondly, in the SPIRIT III trial among those who discontinued a thienopyridine after 6 months, the rates of protocol-defined stent thrombosis tended to be lower in EES-treated patients than in PES-treated patients (EES 0.4% vs. PES 2.6%, P = 0.10). Thirdly, 2-year follow-up of the SPIRIT II trial suggested late angiographic catch-up of in-stent neointimal hyperplasia or per cent stent volume obstruction (VO) by intravascular ultrasound (IVUS) in EES but not in PES.15 The impact of these findings on clinical outcome is still uncertain.

To further explore these issues, we performed a pooled analysis of the 2-year clinical data from the SPIRIT II and III trials to examine the comparative long-term outcomes of EES and PES. Poolability was justified on the basis of comparable inclusion and exclusion criteria with similar baseline and angiographic characteristics and endpoint definitions between the two studies.


Protocol entry criteria

The designs of the SPIRIT II and III trials have been described previously.10,13 In brief, both were prospective, multicentre, single-blind, randomized controlled clinical trials in which 300 and 1002 patients (in SPIRIT II and III, respectively) were randomized to receive the EES (XIENCE™ V, Abbott Vascular, Santa Clara, CA, USA) or the PES (TAXUS® EXPRESS2™, Boston Scientific, Natick, MA, USA) (see Supplementary material online, Appendix for site names). In Spirit II, both TAXUS® Express2™ (73% of lesions) and TAXUS® Liberte® (27% of lesions) were used as control. Patients were eligible for the study if they were aged 18 years and above, with a diagnosis of stable or unstable angina or inducible ischaemia. Additional key eligibility criteria were the presence of either one or two de novo native coronary artery lesions (maximum one lesion per epicardial coronary artery) with a diameter stenosis of ≥50 and <100%, with a lesion length of ≤28 mm and a reference vessel diameter of 2.5–4.25 and 2.5–3.75 mm in SPIRIT II and III, respectively. Patients were excluded from enrolment if they presented with acute or recent MI, had a left ventricular ejection fraction <30%, restenotic lesions or lesions located in the left main coronary artery, were awaiting a heart transplant, or had a known hypersensitivity or contraindication to aspirin, heparin, bivalirudin, clopidogrel or ticlopidine, cobalt, chromium, nickel, tungsten, everolimus, paclitaxel, acrylic, and fluoropolymers. Angiographic exclusion criteria were target lesion(s) in aorto-ostial, left main stem, within 2 mm of the origin of the left anterior descending or left circumflex coronary artery, bifurcation lesions with either the sidebranch >50% stenosed or >2 mm in diameter or requiring pre-dilatation, lesion located within a bypass graft, lesions with heavy calcification, or a visible thrombus within the target vessel.

The studies were approved by the Ethics Committee at each participating institution, and eligible patients gave their written informed consent. Following the confirmation of angiographic criteria, telephone randomization was performed in randomly alternating blocks of four and eight patients in SPIRIT II, or three and six patients in SPIRIT III using an automated voice response system, stratified by the presence of diabetes, planned dual vessel treatment, and study site. Protocol-specified angiographic follow-up was planned in all patients at 6 months for SPIRIT II and at 8 months in a subgroup of 564 SPIRIT III patients. Two-year angiographic follow-up was planned in a pre-defined subgroup of 152 patients from the SPIRIT II trial and not in the SPIRIT III trial.

Medication administration and clinical follow-up

Patient preparation and pharmaceutical treatment during the procedure were to be in accordance with standard hospital practice. The use of GPIIb/IIIa inhibitors was left to the discretion of the physician. All patients were to receive 75 mg clopidogrel for a minimum of 6 months and ≥75 mg of aspirin daily for a minimum of 1 or 5 years for Spirit II and III, respectively, following the procedure; a longer duration of clopidogrel use was permitted per the discretion of the treating physician. Clinical follow-up was scheduled at 1, 6 (9 months only in SPIRIT III), 12 months, 2 years and then yearly through 5 years.

Data management

Clinical study monitors verified 100% of case report form data on-site. For each study, an independent committee blinded to treatment allocation adjudicated all MACEs after review of the original source documentation. A clinical events committee blinded to randomization performed a post hoc adjudication of stent thrombosis using the Academic Research Consortium (ARC) definitions.16 Angiographic and IVUS analyses were performed by independent core laboratory technicians blinded to treatment assignment and clinical outcomes using validated methods as described previously. A Data Safety and Monitoring Committee periodically reviewed blinded safety data, each time recommending the studies to continue without modification.

Clinical endpoints and definitions

Ischaemia-driven target vessel (or lesion) revascularization [ID-TVR (ID-TLR)] was defined as a revascularization at the target vessel (or lesion) associated with an angiographic diameter stenosis ≥50% by core lab quantitative coronary angiography (QCA) with a positive functional ischaemia study (exercise testing, fractional flow reserve, or coronary flow reserve) or ischaemic symptoms; or a diameter stenosis ≥70% by core lab QCA with or without ischaemic symptoms or a positive functional study. Target vessel failure was defined as the occurrence of either cardiac death, MI, or ID-TVR. Major adverse cardiac event was defined as the occurrence of either cardiac death, MI, or ID-TLR. Myocardial infarction was defined as either the development of new pathologic Q-waves ≥0.4 s in duration in two or more contiguous leads, or an elevation of creatine phosphokinase levels to >2.0 times normal with positive creatine phosphokinase-MB. Stent thrombosis was prospectively defined by the study protocols as an acute coronary syndrome with angiographic evidence of thrombus within or adjacent to a previously treated target lesion, or in the absence of angiography, any unexplained death or acute MI with ST-segment elevation or new Q-waves in the distribution of the target lesion occurring within 30 days post-procedure. Stent thrombosis was also categorized in a post hoc analysis according to the definitions proposed by the ARC for definite, probable, and possible stent thrombosis.16

Statistical methods

Continuous variables that were normally distributed are presented as mean ± SD and compared using the t-test. Continuous variables that did not distribute normally are presented as median (lower and upper quartile range) and compared by the Wilcoxon rank sum test. Categorical variables are presented as percentages and were compared using Fisher's exact test. This pooled analysis of SPIRIT II and III was post hoc. All analyses are by intention-to-treat, utilizing all patients randomized in the study, regardless of the treatment actually received. Patients lost to follow-up in whom no event had occurred before the follow-up windows were not included in the denominator for calculations of binary endpoints. Relative risk was calculated as the event rate of EES divided by the event rate of PES arm. The incidence of events over time was studied with the use of the Kaplan–Meier method, whereas log-rank tests were applied to evaluate differences between the treatment groups. Patients lost to follow-up were considered at risk until the date of last contact, at which point they were censored. In addition, a landmark analysis was conducted at 12 months following randomization. Outcomes between 12 and 24 months were then estimated using the Kaplan–Meier method and compared with the log-rank test. Hazard ratios (HRs) of EES vs. PES group were calculated by using Cox's proportional hazard model. To search for the interaction between the subgroup and treatment effect, the logistic regression model was constructed including treatment, subgroup, and treatment*subgroup and P-values were calculated by the Wald χ2 statistics. A two-sided α = 0.05 was used for all statistical tests. All statistical analyses were performed by SAS version 9.1.3 (SAS Institute, Cary, NC, USA).


Baseline clinical and angiographic characteristics were comparable between groups as shown in Table 1. Of the 892 patients randomized to the EES arm, 854 patients were subsequently treated with EES, whereas 386 of the 410 patients assigned to PES treatment actually received PES. In the SPIRIT II trial, 2-year follow-up was completed in 94.2% (210/223) and 93.5% (72/77) of the EES and PES arms, respectively. In the SPIRIT III trial, 2-year follow-up was completed in 93.6% (626/669) and 89.8% (299/333) of the EES and PES cohorts, respectively. In total, 93.7% (836/892) of the EES patients and 90.5% (371/410) of the PES patients completed the 2-year follow-up. The countries or areas with low follow-up rate (≤85%) have enrolled a relatively small number of patients (≤34 patients).

View this table:
Table 1

Patient baseline characteristics

Everolimus-eluting stentPaclitaxel-eluting stent
Number of patients892410
Age in years (mean ± SD)62.9 ± 10.562.6 ± 10.1
Male (%)70.368.2
Diabetes (%)27.927.1
 Treated with insulin (%)7.15.7
Hypertension (%)74.072.3
Hypercholesterolaemia (%)72.872.1
Current smoker (%)25.323.8
Prior MI (%)23.719.3
Unstable angina (%)20.826.5
Number of lesions1032474
 LAD (%)41.143.8
 LCX (%)28.026.4
 RCA (%)30.729.6
 LMCA (%)0.10.2
 RVD in mm [median (IQ range)]2.73 (2.42, 3.05)2.77 (2.43, 3.04)
 MLD in mm [median (IQ range)]0.87 (0.58, 1.15)0.89 (0.58, 1.18)
 %DS [median (IQ range)]66.2 (57.0, 77.2)66.9 (56.2, 77.6)
 Lesion length in mm [median (IQ range)]12.9 (10.0, 17.6)13.0 (10.5, 17.2)
  • There were no significant differences between the groups. SD, standard deviation; IQ, inter-quartile; MI, myocardial infarction; LAD, left anterior descending artery; LCX, left circumflex artery; RCA, right coronary artery; LMCA, left main coronary artery; RVD, reference vessel diameter; MLD, minimal luminal diameter; %DS, per cent diameter stenosis.

Clinical outcomes

Kaplan–Meier estimates of TVF, MACE rate, and the components through 758 days are presented in Table 2; event rates from 1 to 2 years are shown in Table 3. Kaplan–Meier estimates of survival with events are depicted in Figure 1. At 2 years, TVF rates were 10.4% for EES vs. 14.7% for PES [HR (95% CI) = 0.69 (0.50, 0.96)]. Major adverse cardiac event rates were 7.1% for EES vs. 12.3% for PES [HR (95% CI) = 0.55 (0.38–0.80)]. The observed reduction in TVF and MACE rates in patients randomized to EES compared with PES were driven by lower rates of non-Q-wave MI (2.7 vs. 5.0%, log-rank P = 0.037) and TLR by PCI (3.9 vs. 6.3%, log-rank P = 0.044), with no differences in the rates of Q-wave MI (0.3 vs. 0.5%, respectively, log-rank P = 0.66) or cardiac death (0.9 vs. 1.3%, respectively, P = 0.56). Stent thrombosis rates were comparable by both the pre-specified protocol and post hoc ARC definitions (Table 4). The length of thienopyridine therapy was not different between EES and PES groups (487.4 ± 241.4 vs. 500.0 ± 243.7 days, P = 0.38). When stratified by the timing of thienopyridine discontinuation, the incidence of ARC-defined stent thrombosis (definite or probable) did not significantly vary with the two stent types at any time period (discontinuation before 6 months: EES 1.8% vs. PES 3.8%, P = 0.59; discontinuation between 180 and 758 days EES 1.1% vs. PES 1.3%, P = 1.00; never discontinued 0.8 vs. 0.6%, P = 1.00). The P-values for interaction of treatment (EES or PES) by first discontinuation of thienopyridine before 180 days (yes or no), treatment by first discontinuation between 180 and 758 days, and treatment by never discontinued were 0.83, 0.88, and 0.48, respectively.

Figure 1

Kaplan–Meier survival curves stratified according to treatment arm of everolimus- or paclitaxel-eluting stent. (A) The composite of cardiac death or myocardial infarction, (B) ischaemic target lesion revascularization, (C) the composite of cardiac death, any myocardial infarction, or ischaemia-driven target lesion revascularization (D) stent thrombosis (per-protocol). CI, confidence interval; HR, hazard ratio.

View this table:
Table 2

Event rates at 2 years

EES (%)PES (%)Hazard ratio95% CILog-rank P-value
MACE (cardiac death, MI, ID-TLR)–0.800.0014
TVF (Cardiac Death, MI, ID-TVR)10.414.70.690.50–0.960.027
 All death2.43.30.720.36–1.450.36
 Cardiac death0.91.30.720.24–2.200.56
 Any MI3.15.60.550.31–0.960.034
  Q-wave MI0.30.50.670.11–4.030.66
  Non-Q-wave MI2.75.00.540.30–0.970.037
 Any ischaemia-driven TLR4.16.80.590.36–0.960.031
 All TLRa5.310.10.510.34–0.770.0012
 Any ischaemia-driven TVR7.99.90.770.52–1.130.18
 All TVRb9.212.70.710.50–1.010.054
  • Event rate was estimated by the Kaplan–Meier method. EES, everolimus-eluting stent; PES, paclitaxel-eluting stent; CI, confidence interval; MACE, major adverse cardiac event; MI, myocardial infarction; ID, ischaemia-driven; TLR, target lesion revascularization; CABG, coronary artery bypass graft; PCI, percutaneous coronary intervention; TVR, target vessel revascularization.

  • aAll TLR includes both ischaemia-driven and non-ischaemia-driven TLR.

  • bAll TVR includes both ischaemia-driven and non-ischaemia-driven TVR.

View this table:
Table 3

Event rate between 1 and 2 years

EES (%)PES (%)Hazard ratio95% CILog-rank P-value
MACE (cardiac death, MI, ID-TLR)–1.560.40
TVF (cardiac death, MI, ID-TVR)–1.390.35
 All death1.21.60.740.27–2.040.56
 Cardiac death0.40.31.330.14–12.780.80
 Any MI0.81.70.510.17–1.510.22
  Q-wave MI0.10.30.450.03–7.120.56
  Non-Q-wave MI0.71.40.520.16–1.710.28
 Any ischaemia-driven TLR1.11.10.860.29–2.520.79
 All TLRa1.–2.860.99
 Any ischaemia-driven TVR2.52.60.870.42–1.790.70
 All TVRb2.–2.440.75
  • Event rates were calculated using the Kaplan–Meier method. EES, everolimus-eluting stent; PES, paclitaxel-eluting stent; CI, confidence interval; MACE, major adverse cardiac event; MI, myocardial infarction; ID, ischaemia-driven; TLR, target lesion revascularization; CABG, coronary artery bypass graft; PCI, percutaneous coronary intervention; TVR, target vessel revascularization.

  • aAll TLR includes both ischaemia-driven and non-ischaemia-driven TLR.

  • bAll TVR includes both ischaemia-driven and non-ischaemia-driven TVR.

View this table:
Table 4

Stent thrombosis rates at 2 years

EES (n = 892)PES (n = 410)Relative risk (95% CI)P-value
ARC definition
 Acute stent thrombosis (<1 days)
   Definite0.1% (1/892)0.0% (0/407)NC (NC)1
   Probable0.0% (0/892)0.0% (0/407)NC (NC)NA
   Possible0.0% (0/892)0.0% (0/407)NC (NC)NA
   Definite/probable0.1% (1/892)0.0% (0/407)NC (NC)1
   Definite/probable/possible0.1% (1/892)0.0% (0/407)NC (NC)1
 Subacute stent thrombosis (1–30 days)
   Definite0.2% (2/890)0.2% (1/407)0.91 (0.08, 10.06)1
   Probable0.0% (0/890)0.0% (0/407)NC (NC)NA
   Possible0.0% (0/890)0.0% (0/407)NC (NC)NA
   Definite/probable0.2% (2/890)0.2% (1/407)0.91 (0.08, 10.06)1
   Definite/probable/possible0.2% (2/890)0.2% (1/407)0.91 (0.08, 10.06)1
 Late stent thrombosis (31–393 days)
   Definite0.2% (2/872)0.3% (1/394)0.90 (0.08, 9.94)1
   Probable0.1% (1/872)0.5% (2/394)0.23 (0.02, 2.48)0.2301
   Possible0.5% (4/872)0.5% (2/394)0.90 (0.17, 4.91)1
   Definite/probable0.3% (3/872)0.8% (3/394)0.45 (0.09, 2.23)0.3826
   Definite/probable/possible0.8% (7/872)1.3% (5/394)0.63 (0.20, 1.98)0.5315
 Very late stent thrombosis (394–758 days)
   Definite0.2% (2/841)0.3% (1/372)0.88 (0.08, 9.73)1
   Probable0.2% (2/841)0.5% (2/372)0.44 (0.06, 3.13)0.5906
   Possible0.5% (4/841)0.3% (1/372)1.77 (0.20, 15.78)1
   Definite/probable0.5% (4/841)0.8% (3/372)0.59 (0.13, 2.62)0.4444
   Definite/probable/possible1.0% (8/841)1.1% (4/372)0.88 (0.27, 2.92)0.7643
 Overall stent thrombosis (0–758 days)
   Definite0.8% (7/847)0.8% (3/376)1.04 (0.27, 3.98)1
   Probable0.4% (3/847)1.1% (4/376)0.33 (0.07, 1.48)0.2111
   Possible0.9% (8/847)0.8% (3/376)1.18 (0.32, 4.44)1
   Definite/probable1.2% (10/847)1.6% (6/376)0.74 (0.27, 2.02)0.5889
   Definite/probable/possible2.1% (18/847)2.4% (9/376)0.89 (0.40, 1.96)0.8333
 Acute (<1 day)0.1 (1/892)0 (0/407)NC1
 Subacute (1–30 days)0.2 (2/890)0 (0/407)NC1
 Late (31–393 days)0.3 (3/867)0.8 (3/392)0.45 (0.09–2.23)0.38
 Very late (394–758 days)0.5 (4/837)0.8 (3/372)0.59 (0.13–2.63)0.45
 Overall stent thrombosis (0–758 days)1.2 (10/838)1.6 (6/374)0.74 (0.27–2.03)0.59
  • EES, everolimus-eluting stent; PES, paclitaxel-eluting stent; CI, confidence interval; NC, not calculatable.

Subgroup analysis

Regression analysis was performed to explore whether or not the reduction of MACE at 2 years with EES compared with PES was consistent across important subgroups. As shown in Figure 2, the treatment effect of EES compared with PES was consistent in all subgroups and was mainly in favour of EES, except possibly for patients with diabetes, in whom the frequency of MACE was comparable between EES and PES. Among patients with diabetes mellitus, the 2-year MACE rates were 11.3% with EES compared with 7.2% with PES [RR (95% CI) = 1.56 (0.70–3.46)], whereas in those without diabetes, the 2-year MACE rates were 5.9% with EES compared with 15.5% with PES [RR (95% CI) = 0.38 (0.25–0.58)], with a P-value for interaction of 0.002. In diabetic patients, the increased MACE rate in EES is mainly driven by cardiac death (EES 2.1% vs. PES 0%) and ID-TLR (3.8 vs. 1.0%) hierarchically. As the sample size was small for most subgroups, caution should be used in the interpretation of these results and in drawing conclusions.

Figure 2

Subgroup analyses of the 2-year rates of major adverse cardiac events among patients randomized to receive the everolimus-eluting stent vs. the paclitaxel-eluting stent. Probability for interaction represents the likelihood for interaction between the variable and the relative treatment effect. CI, confidence interval; LAD, left anterior descending; IDDM, insulin-dependent diabetes mellitus; NIDDM, non-insulin-dependent diabetes mellitus; EES, everolimus eluting stent; PES, paclitaxel-eluting stent.


In this pooled analysis of 2-year outcomes from the SPIRIT II and III trials, the use of EES compared with PES resulted in significant reductions in MI and ID-TLR. As seen in the event rate curves (Figure 1), the reduction in MI was due to fewer peri-procedural and late events, whereas the reduction in TLR was due to a reduction in restenosis-related events occurring mostly between 4 and 9 months. There were no significant differences between the two stent types in the early or late rates of death and stent thrombosis. Thus, considering composite measures of combined safety and efficacy, both TVF and MACE were significantly reduced at 2 years by treatment with EES rather than PES. In addition, the current analysis suggested the following findings: (i) the hypothesis that EES is associated with fewer very late stent thrombosis than PES was not proven; (ii) the late increase in neointima from 6 months to 2 years in EES was not translated into an increase in clinical event rates; (iii) in the diabetic population, EES might have different treatment effects in comparison to PES.

Serial assessment of angiography and IVUS performed at 6 and 24 months in a subgroup of patients in SPIRIT II showed an increase in angiographic late loss and in-stent neointimal volume in the EES group when compared with the PES group.15 Compared with 6-month QCA, 2-year QCA showed that in-stent late loss had significantly increased in EES from 0.17 ± 0.32 to 0.33 ± 0.37 mm, but was unchanged in the PES group (from 0.33 ± 0.32 to 0.34 ± 0.34 mm). With serial IVUS assessment, neointimal hyperplasia volume and %VO were significantly lower in EES than in PES at 6 months, whereas the differences became no longer significant at 2 years. Despite these results suggesting incremental increase in neointimal hyperplasia noted in EES but not in PES, the present study demonstrates from these two randomized trials that the ID-TLR rates at 2 years with EES are lower than with PES in the non-complex lesions enrolled in these studies.

Subgroup analysis of the combined populations of SPIRIT II and III raise the possibility that there may be a differential treatment effect of stent type according to diabetic status. In non-diabetic patients, a marked reduction in MACE was present with EES compared with PES, whereas in diabetic patients (both insulin-requiring and non-insulin-dependent diabetes mellitus), there were no significant differences in MACE between the two stent types. Of interest, in the SPIRIT II trial, the increase in luminal loss between 6 months and 2 years among EES-treated patients tended to be larger in the diabetic than in the non-diabetic cohort, although this difference did not reach statistical significance (diabetics 0.25 ± 0.43 mm vs. non-diabetics 0.14 ± 0.29 mm, Δ + 0.11 mm, P = 0.36). The same trend was observed in %VO with IVUS: Δ %VO in diabetics was 3.76 ± 6.63, whereas Δ %VO in non-diabetics was 2.12 ± 6.22 (P = 0.44). Theoretically, the different mechanisms of action of the drugs in terms of inhibition of neointimal proliferation could explain the disparity in outcome of EES and PES in the diabetic patients. Insulin exhibits an up-regulation in the PI3-kinase signal transduction pathway which involves phosphorylation and activation of the mammalian target of rapamycin (mTOR).17,18 Everolimus, a natural macrocyclic lactone, inhibits mTOR thereby blocking the cell-cycle during the transition from G1 to S phase: inhibiting protein synthesis by blocking mTOR with everolimus (or any other rapamycin analogue) may be less effective in diabetic patients. On the other hand, paclitaxel might exert the same potency in diabetics as in non-diabetics by inhibiting deconstruction of microtubules, independent of insulin resistance.17

However, between 6 months and 2 years, there was a non-significant increase in TLR in the EES arm compared with the PES arm [EES 4.6% (11/239) vs. PES 2.1% (2/97), P = 0.36). Thus, increased late loss beyond 6 months leading to greater TLR cannot explain the large differences observed in MACE rates between the two stents according to diabetic status. Alternatively, the relatively small size of the diabetic cohort may have falsely led to the noted interaction in this post hoc analysis. Specifically, the 2-year MACE rate among patients treated with PES was lower in diabetic compared with non-diabetic patients, an unexpected finding that may have been due to chance. Examination of numerous underpowered subgroups may lead to false-positive as well as false-negative results.19 Thus, a large, adequately powered randomized trial is required to determine the relative safety and efficacy of EES and PES in patients with diabetes.

The SPIRIT III study suggested that incidence of very late (>1 year) stent thrombosis is less in EES than in PES (0.2 vs. 1.0%, P = 0.10) and that thienopyridine discontinuation after 6 months might be associated with a lower rate of subsequent stent thrombosis with EES than with PES through 2 years of follow-up (0.4 vs. 2.6%), although this difference did not reach statistical significance (P = 0.10).14 In the current analysis, the suggested lower rate of very late stent thrombosis in EES compared with PES was not proven. Between 1 and 2 years, protocol-defined stent thrombosis occurred in 0.5 and 0.8% of the EES and PES patients, respectively, without significant difference (P = 0.45). The current analysis also failed to confirm the lower rate of stent thrombosis in the EES than in PES group when thienopyridine was discontinued after 6 months. The rates of stent thrombosis at 2 years were comparable between EES and PES irrespective of the timing of thienopyridine discontinuation, with no difference observed between the two groups at 2 years. Given the low incidence of stent thrombosis, however, larger studies will be necessary to assess the differential effects of DES on stent thrombosis.

In the current analysis, the 2-year follow-up could not be completed for 7.2% of the patients, although it was mandated by protocol. This is a limitation of the current analysis since we cannot exclude occurrence of events in the patients with lost follow-up. Although the completeness of follow-up is similar between EES and PES (93.7 vs. 90.5%), the slight differences could be relevant.

In conclusion, the current analysis reports the largest cohort with the longest follow-up of patients treated with the EES, which is currently the most widely used DES in the USA and Europe. This study has demonstrated that in patients with mostly stable angina and non-complex coronary artery disease, the EES compared with the PES reduces the rates of MI and TLR, with lower overall TVF and MACE. Additional studies are warranted to confirm the long-term safety and efficacy of the EES in diabetic patients, after thienopyridine discontinuation, and in more complex lesions and patients than studied in the SPIRIT trials to date.


This study was sponsored by Abbott Vascular.

Conflict of interest: S.V., J.D. and S.C. are employees of Abbott Vascular. G.W.S. is a member of advisory board to Abbott Vascular and Boston Scientific.


View Abstract