Aims In the previous COREA-TAXUS trial, a 6-month adjunctive use of celecoxib reduced target-lesion revascularization (TLR) without increased thrombotic risk. We aimed to confirm the effects of 3-month celecoxib in patients receiving drug-eluting stent (DES) implantation in the larger prospective, randomized trial.
Methods and results Patients (n = 909) treated for native coronary lesions were randomized into four groups: the control or the celecoxib group with stratification by stents: paclitaxel-eluting stent (PES) or zotarolimus-eluting stent (ZES). In the celecoxib group, 200 mg of celecoxib was given twice daily for 3 months after the procedure. The primary endpoint was in-stent late loss (LL) at 6 months. In-stent LL was significantly lower in the celecoxib group than the control group (0.64 ± 0.54 vs. 0.55 ± 0.47 mm, P = 0.02). The trend of LL reduction in the celecoxib group was maintained in the ZES and PES subgroups, although it did not reach statistical significance. There was a trend towards the reduced clinically driven TLR in the celecoxib group (5.7 vs. 3.2%, log-rank P = 0.09), but adverse cardiac events rate did not differ between the two groups (composite of cardiac death, non-fatal myocardial infarction, and TLR; 8.6 vs. 7.7%, log-rank P = 0.84). Non-fatal myocardial infarction and cardiac death occurred in 1.6% of the patients in the celecoxib group when compared with 0.2% in the control group (log-rank P = 0.03).
Conclusion Three-month adjunctive celecoxib would be useful to reduce LL of DES. However, this study may raise the concern about increased thrombotic risk with celecoxib even in patients receiving dual anti-platelet therapy.
Drug-eluting stents (DES) reduce the incidence of angiographic in-stent restenosis and repeat revascularization.1,2 However, the majority of complications in percutaneous coronary intervention (PCI) with DES is still in-stent restenosis.3,4 Celecoxib is a selective inhibitor of cyclo-oxygenase 2 (COX-2) that is commonly used as an anti-inflammatory agent. However, it also exerts anti-proliferative effects and induces apoptosis in cancer cells.5,6 We have shown that celecoxib is a potential inhibitor of neointimal formation by blocking injury-induced Akt activation.7 Our previous COREA-TAXUS trial8 (the Effect Of Celecoxib on REstenosis after coronary Angioplasty with TAXUS stent) showed that the adjunctive use of celecoxib for 6 months after TAXUS-EXPRESSTM stent implantation in patients with coronary artery disease significantly decreased in-stent luminal late loss (LL). The rate of adverse cardiac events within the first 6 months after stent implantation was lower in the celecoxib group than in the control group. Furthermore, we recently confirmed in the COREA-TAXUS long-term follow-up study9 that celecoxib treatment for 6 months provided better clinical outcomes for 2 years than the control group. At 2 years, the rate of adverse cardiac events was consistently lower in the celecoxib group (6.9%) than in the control group (19.7%).
In the DES system, most of the drugs, such as paclitaxel or zotarolimus, elute in the first 4 weeks to inhibit the early proliferation of smooth muscle cells after stent implantation. Celecoxib may inhibit the proliferation of not only vascular smooth muscle cells but also endothelial cells10 and endothelial progenitor cells.11 Therefore, the use of celecoxib in the later healing phase in DES implantation might delay the healing of the endothelial layer after coronary intervention and interrupt the balance between prostaglandin and thromboxane, which can increase the thrombotic risk. We hypothesize that shortened celecoxib treatment from 6 to 3 months might be sufficient to reduce neointimal growth and might reduce the potential thrombotic risk of celecoxib.
To confirm this hypothesis, we evaluated the efficacy and safety of 3-month duration of adjunctive celecoxib treatment in reducing neointimal hyperplasia in patients with coronary implantation of the Taxus-libertéTM PES (Boston Scientific Corporation, Natick, MA, USA) or the EndeavorTM ZES (Medtronic Vascular, Santa Rosa, CA, USA).
This mini-COREA trial was a prospective, randomized, open-label, multicentre trial based at five centres in South Korea: the Seoul National University Hospital, the Seoul National University Bundang Hospital, the Kwangju Veterans Hospital, the Korea University Guro Hospital, and Konyang University Hospital. Patients were enrolled between March 2006 and June 2009. Patients are at least 30 years of age who have angina pectoris or a positive stress test with native coronary artery lesions for which DES implantation was feasible. Reasons for exclusion were: acute or recent ST-segment elevation myocardial infarction (within 4 weeks); left main coronary artery disease; hepatic dysfunction (aspartate aminotransferase or alanine aminotransferase >120 IU/L); renal dysfunction (serum creatinine >2.0 mg/dL); severe congestive heart failure (NYHA class >2); left ventricular ejection fraction <30%; haemodynamically unstable condition; definite intracoronary thrombus; contraindication or history of allergy to aspirin, clopidogrel, or celecoxib; expected survival <2 years due to other medical conditions. Patients already taking warfarin, any COX-2 inhibitor, or non-steroidal anti-inflammatory drugs (NSAIDs) were also excluded. All patients gave written informed consent and the institutional review board at each centre approved this study.
Procedural anticoagulation was achieved with unfractionated heparin as per the standard of care, with glycoprotein IIb/IIIa inhibitors used per operator discretion. Patients were administered 300 mg aspirin before catheterization. A ≥300 mg oral dose of clopidogrel was recommended before the procedure. The protocol recommended use of aspirin 100 mg daily indefinitely and clopidogrel 75 mg daily for a minimum of 6 months; a longer duration of clopidogrel use was permitted at the discretion of the treating physicians. Other medications were prescribed as per the standard of care.
Patients were randomized into one of four groups: celecoxib/Taxus-libertéTM stent group, control/Taxus-libertéTM stent group, celecoxib/EndeavorTM stent group, and control/EndeavorTM stent group. Patients in the celecoxib group were given 200 mg CelebrexTM twice daily for the 3 months. To reduce the potential risk of thrombotic complication, the duration of celecoxib treatment was reduced to 3 months. However, to avoid reduction in the anti-proliferative effect of celecoxib, a 400 mg daily drug dose was used because we believed that the anti-proliferative effect of celecoxib is dose-dependent and further reduction in dose may result in reduction in efficacy.12,13 In the case of multivessel intervention, allocated DES were used for all lesions. Blood samples were taken immediately before and 1 month after stent implantation. Coronary angiograms were obtained before and after the intervention within a single session, and 6 months after the procedure. Quantitative coronary angiography (QCA) was done by two independent angiographers who were not aware of the patients' celecoxib status and analysis was done with CAAS II QCA research Version 2.0.1 (PIE Medical Imaging, Maastricht, The Netherlands). Measurements were obtained at both the stented segment for in-stent analysis, and proximal and distal 5 mm native reference segment adjacent to stent and stented segment for in-segment analysis.
The primary endpoint was comparison of the in-stent LL between the celecoxib and control groups assessed by QCA at 6-month follow-up in the whole study population and subgroups such as Taxus-libertéTM or EndeavorTM subgroup. The secondary endpoint was composite of cardiac death, myocardial infarction, and need for revascularization of the target lesion, at 6 months after coronary intervention. All deaths were regarded as cardiovascular, unless there was documented evidence of a clear non-cardiovascular cause. Myocardial infarction was defined as the presence of at least two of the following: ischaemic symptoms; concentrations of cardiac enzymes (creatine kinase and its MB isoenzyme, CK-MB) at least twice their upper normal limits; and new electrocardiographic changes compatible with myocardial infarction. Clinically driven target-lesion revascularization (TLR) was defined as revascularization done in lesions with 50–70% diameter stenosis that was related to ischaemic symptoms, objective signs, or both, or as revascularization done in lesions with 70% or more diameter stenosis, even in the absence of ischaemic symptoms or signs. Stent thrombosis was defined as definite or probable ST according to the Academic Research Consortium criteria.14
Primarily, our study was designed to evaluate effects of celecoxib in DES [either paclitaxel-eluting stent (PES) or zotarolimus-eluting stent (ZES)]. To evaluate the effect of celecoxib on ZES and PES, respectively, the in-stent LL of PES (Taxus-libertéTM) and ZES (EndeavorTM) was estimated to be 0.65 ± 0.60 and 0.61 ± 0.46 mm on the basis of the results of the COREA-TAXUS trials8 and ENDEAVOR II trials,15 respectively. This trial was designed to have 80% power to detect an expected 25% reduction in LL with celecoxib, with a two-sided α-level of 0.05. Presuming an angiographic follow-up rate of 80%, we planned to enrol 900 patients: 540 patients in the PES group and 360 in the ZES group.
All primary and secondary endpoints were analysed on the modified intention-to-treat basis; patients who completed allocated DES implantation were included for analysis. We compared continuous variables using the Student's t-test, and analysed categorical variables using the χ2 test. The Mann–Whitney U-test was used to compare non-normally distributed data, such as highly sensitive C-reactive protein concentrations. Serial changes of continuous variables at 6-month follow-up were compared by use of a paired t-test. Event-free survival was analysed from the time of stent implantation to the first event according to the Kaplan–Meier method and the difference was evaluated by the log-rank test as pre-specified in the protocol. SPSS version 18.0 was used for all statistical analyses and P < 0.05 (two-sided) was considered statistically significant.
Figure 1 shows the trial profile. Among the initially randomized patients, six patients failed to implant DES for the target lesion, six patients withdrew consent, and one patient lost to follow-up. Therefore, the modified intention-to-treat analysis population consisted of 896 patients: 442 assigned to celecoxib and 454 to control. The overall angiographic follow-up rate in our study was 88% and it was higher than the presumed rate, 80%.
Trial profile. ITT, intention-to-treat; FU, follow up.
Table 1 shows the baseline clinical, angiographic, and procedural characteristics of the 903 patients who entered the study. Baseline characteristics were not different between the celecoxib and control groups except the stent profile. The patients in the celecoxib group received longer stent (P = 0.02). At discharge, statin was used by 452 (96%) patients in the control group and 430 (97%) in the celecoxib. And cilostazol in addition to dual anti-platelets with aspirin and clopidogrel was used by 185 (41%) patients in the control group and 171 (39%) in the celecoxib group.
Baseline clinical, angiographic, and procedural characteristics
(n = 454)
(n = 449)
Mean age (years)
63.3 ± 9.1
63.2 ± 9.0
Current smoker (%)
Previous myocardial infarction (%)
Previous revascularization (%)
Non-ST elevation myocardial infarction
Lesion and procedural characteristics
Left anterior descending artery (%)
Left circumflex artery (%)
Right coronary artery (%)
Type B2/C lesion (%)
Type C lesion (%)
Total occlusion (%)
Bifurcation lesion (%)
Moderate-to-heavy calcification (%)
Mean number of stents per lesion
1.13 ± 0.41
1.21 ± 0.50
Mean stent diameter (mm)
3.09 ± 0.41
3.14 ± 0.38
Mean stent length per lesion (mm)
28.4 ± 12.6
30.7 ± 15.7
Adjunctive balloon dilatation (%)
Mean max. inflation pressure, overall (atm)
14.1 ± 3.2
13.8 ± 3.1
Blood pressure and blood tests
Blood pressure and results of blood tests were similar in the celecoxib and control groups at baseline and at 1 month except the lipid profile (Table 2). The mean levels of total and LDL cholesterol were higher in the control group at baseline and 1-month follow-up. However, the change of total and LDL cholesterol for 1 month did not differ between the celecoxib and control groups (total cholesterol, 40.7 ± 39.7 vs. 41.0 ± 39.7, P = 0.82; LDL cholesterol, 38.3 ± 34.4 vs. 39.3 ± 33.1, P = 0.71).
Baseline QCA measurements did not differ between the two groups (Table 3). In the whole study population, the mean in-stent LL was smaller in the celecoxib group (0.55 ± 0.47 mm) than in the control group (0.64 ± 0.54 mm) (0.09 mm absolute reduction, 14% relative reduction, P = 0.02). In-segment LL and restenosis rate did not differ between two groups.
The mean in-stent LL was analysed also in the subgroup treated with different stents, the PES group and the ZES group. Celecoxib showed a consistent tendency to reduce LL in both DES, which did not reach statistical significance (see Supplementary material online, Figure S1). For the PES group, in-stent LL was 0.49 ± 0.45 mm in the celecoxib group and 0.56 ± 0.50 mm in the control group (absolute reduction 0.07 mm; 13% relative reduction, P = 0.12). For the ZES group, in-stent LL was 0.65 ± 0.50 mm in the celecoxib group and 0.76 ± 0.58 mm in the control group (absolute reduction 0.11 mm; 14% relative reduction, P = 0.07).
Clinical outcomes analysis
The cumulative Kaplan–Meier estimates of the rates of cardiac events (log-rank P = 0.84) and TLR (log rank P = 0.45: mean follow-up 6.8 ± 1.6 months) did not differ in two groups (Figure 2A and B, and Table 4). There was a trend towards the reduced clinically driven TLR in celecoxib group (5.7 vs. 3.2%, log-rank P = 0.09: figures not shown). The revascularization rates of patients who failed to obtain a 6-month follow-up angiogram were similar between the control and the celecoxib group; two events in the control group (3 and 4 months after intervention, respectively) and one event in the celecoxib group (4 months after intervention).
Cumulative incidence curves are shown for all major cardiovascular events (A), target-lesion revascularization (B), and hard endpoints (C) including non-fatal myocardial infarction or cardiac death during follow-up period.
However, cumulative incidence of hard endpoints including non-fatal myocardial infarction and cardiac death was higher in the celecoxib group than in the control group (1.6 vs. 0.2%, log-rank P = 0.03). In the celecoxib group, there were four cases of myocardial infarction due to stent thrombosis (one fatal and three non-fatal) and two sudden cardiac deaths. In the control group, there was just one non-fatal myocardial infarction due to stent thrombosis. Most of hard endpoints occurred during initial 3 months after intervention.
Subgroup analysis did not show heterogeneity of celecoxib treatment effects on adverse cardiac event (Figure 3).
Pre-specified subgroup analysis for cardiac adverse events at 6 months. LAD, target lesion in the left anterior descending coronary artery; DES, drug-eluting stent; P int, P-value for interaction.
Discontinuation of celecoxib
Table 5 shows the reasons for discontinuation of celecoxib in the 28 patients (6% of that group) who stopped taking the drug before the designated 3 months. Adverse reactions that caused discontinuation of celecoxib were resolved after stopping celecoxib without sequel. There were no other serious side effects in patients who continued to take celecoxib. Of the patients who discontinued celecoxib by their preference, 15 did so within 1 month. Of 28 patients who stopped taking the drug before the designated 3 months, only two events occurred, all at several months after discontinuation of celecoxib. One patient (3.6%) died 2 months after the discontinuation of celecoxib and the other (3.6%) received TLR 5 months after the discontinuation of celecoxib. These two patients discontinued celecoxib within a month after PCI due to skin rash.
aCardiac death (n = 1): 2 months after the discontinuation of celecoxib; target-lesion revascularization (n = 1): 5 months after the discontinuation of celecoxib.
We have shown that the adjunctive use of celecoxib for 3 months reduces LL of DES in the whole study population. Such an effect of celecoxib to reduce LL was consistently observed in two different DES, PES and ZES, although it did not reach a statistical significance. Adjunctive celecoxib therapy in patients with DES implantation reduced the mean LL by 14% at 6-month follow-up. However, there was no difference in the occurrence of major adverse cardiac events between the two groups, because there was a small but significant increase in the rate of thrombotic events, such as cardiac death and myocardial infarction, in the celecoxib group.
Comparisons between mini-COREA and COREA-TAXUS trials
In the previous COREA-TAXUS trial, celecoxib treatment significantly reduced both LL and adverse cardiac events, whereas it reduced LL only in the mini-COREA trial. The mean LL of the celecoxib group in the mini-COREA trial (whole patients: 0.55 ± 0.47 and PES group: 0.49 ± 0.45 mm) was similar to that in the COREA-TAXUS trial (0.49 ± 0.47 mm), but the mean late luminal loss of the control group in the mini-COREA trial (whole patients: 0.64 ± 0.54 and PES group: 0.56 ± 0.50 mm) is lower than that in the COREA-TAXUS trial (0.75 ± 0.60 mm). Consequently, the extent of reduction in in-stent LL by celecoxib was milder in the mini-COREA than in the COREA-TAXUS trial (14 vs. 34%). There were several differences between mini-COREA and COREA-TAXUS trials: first, the treatment protocol of celecoxib was changed. A loading dose of celecoxib was not given before coronary intervention in the celecoxib group and the duration of celecoxib treatment was shortened from 6 to 3 months. Elimination of the loading dose and the reduced duration of treatment may reduce the efficacy of celecoxib. Secondly, the effect of adjunctive celecoxib treatment was evaluated in the patient treated with the new version of DES, Taxus-libertéTM or EndeavorTM. Using different DES can be one of reasonable explanation of different outcomes between mini-COREA and previous COREA-TAXUS trial. Regarding PES, Taxus-libertéTM stent used in the current study showed better angiographic results than TAXUS-EXPRESSTM stent used in the previous study.16 Improved LL in the control group might attenuate the anti-restenotic effects of celecoxib. Additionally, the celecoxib group received slightly but significantly longer and more stents than the control group in the mini-COREA trial. And the benefit of celecoxib was attenuated in-segment LL rather than in-stent LL. This may attenuate the preventive effects of celecoxib on TLR. Thirdly, cilostazol was more frequently used in the mini-COREA trial. Cilostazol has mild but significant anti-restenotic effects.17,18 Triple anti-platelets including cilostazol were given to 40% of the patients. These might reduce neointimal growth in the current study population and attenuate effects of celecoxib. Fourthly, in the current study, lipid management is stricter as reflected by the higher prescription rate of statin and the lower achieved level of LDL cholesterol, than in the COREA-TAXUS trial. This might reduce LL in the control group and attenuate the anti-restenotic effect of celecoxib.
Comparison of the celecoxib effect between paclitaxel- and zotarolimus-eluting stents
The current study evaluated the effects of adjunctive celecoxib therapy in PES and ZES. Paclitaxel is a microtubule-interfering agent, whereas zotarolimus is a sirolimus analogue, blocking the mammalian target of rapamycin and inducing cell-cycle arrest in the G1 phase.19,20 These two drugs are currently widely used for DES. There are potentials that differences in biological effects of paclitaxel and zotarolimus may differentially interact with anti-proliferative effects of celecoxib. Results from this study showed that there was no difference in the reduction in LL by celecoxib between ZES and PES. The same extent of reduction was observed in the ZES and PES groups (13 vs. 14%). It means that we can expect similar effects of adjunctive therapy with celecoxib in both types of DES, ZES or PES.
Safety issues of celecoxib cyclo-oxygenase 2 inhibitors
Selective COX-2 inhibitors such as celecoxib and rofecoxib might have been associated with an increase in cardiovascular risk in placebo-controlled trials conducted in patients with colorectal polyps.12,21,22 Experts discouraged the use of unopposed COX-2 inhibition. However, there are no consistent evidences of the increased cardiovascular risk by COX-2 inhibition. In large meta-analysis studies,23,24 celecoxib, in contrast to rofecoxib, did not increase the risk of cardiovascular events. So far, only one large randomized-controlled trial (PRECISION) is being performed to evaluate the cardiovascular safety of celecoxib in comparison with the two other commonly used non-selective NSAIDs in patients who are at high cardiovascular risk.25 The conclusion about the cardiovascular risk of COX-2 inhibition needs to be deferred till completion of the PRECISION trial. Additionally, in most previous studies which reported thrombotic risk with COX-2 inhibition, COX-2 inhibitors were taken without aspirin or anti-platelets. The previous study showed that celecoxib treatment does not interfere with the anti-platelet effects of aspirin or clopidogrel in healthy volunteers.26 The safety of celecoxib adjunctive therapy under the coverage of dual anti-platelets in angina patients was previously demonstrated in the COREA-TAXUS trial. But the sample size was too small to evaluate the safety. There were just two thrombotic events (2/267, 0.7%), one non-fatal myocardial infarction due to stent thrombosis in the celecoxib group and one sudden cardiac death in the control group.
The dose and regimen of celecoxib may be associated with increased thrombotic risk. Meta-analysis showed that 400 mg daily of celecoxib is associated with increased thrombotic risk especially in patients with high cardiovascular risk.27 At design of this study, we determined the treatment schedule of celecoxib considering safety and efficacy. Previous studies on adenomatous polyp suggested that the anti-proliferative effect of celecoxib is dose-dependent and further reduction in dose may results in reduction in efficacy.12,13 And we already reported the efficacy and safety of 200 mg twice daily celecoxib in the COREA-TAXUS trial, in which concomitant dual anti-platelet treatment was used. To balance safety and efficacy, the dose of celecoxib was determined as 200 mg twice daily and the duration of treatment was reduced to 3 months.
Compared with the COREA-TAXUS trial, in the current study, the absolute number of thrombotic events increased as sample size enlarged. There were eight events (8/790, 1.0%), one in the control group and seven in the celecoxib group. In the control group, one patient had subacute stent thrombosis and she was successfully revascularized. In the celecoxib group, six patients had thrombotic events when they were taking celecoxib with dual anti-platelet therapy and two of them died. Additionally, one patient had sudden death during sleep 2 months after the discontinuation of celecoxib due to pruritus. All of eight patients were in the use of aspirin and clopidogrel at the time of events. The rate of thrombotic events was significantly higher in the celecoxib group than in the control group (1.6 vs. 0.2%, log-rank P = 0.03). Because the sample size is not enough to evaluate the risk of thrombosis associated with celecoxib, the possibility of a type II error could be high. However, it is suggested that the use of celecoxib may be harmful in the patients who were treated with DES even during the use of dual anti-platelet therapy.
In conclusion, the adjunctive use of celecoxib for 3 months can be useful to reduce LL of DES and probably TLR, but it does not reduce major adverse cardiac events because it may increase thrombotic hard endpoints. And, this study may raise the concern about the increased thrombotic risk with celecoxib even in patients receiving dual anti-platelet therapy. Based on the current data, we did not recommend the routine use of celecoxib in PCI with DES to reduce restenosis and adverse cardiac event.
This study was supported by a grant from the Clinical Research Center for Ischemic Heart Disease (A040152) and a grant from the Innovative Research Institute for Cell Therapy, Seoul National University Hospital (A062260), both sponsored by the Ministry of Health, Welfare & Family, Korea. H.-S.K. is also a professor of Word Class University Program, Molecular Medicine and Biopharmaceutical Sciences, Seoul National University sponsored by the Ministry of Education, Science & Technology, Korea.
Conflict of interest: none declared.
↵† These authors contributed equally to this work.
. The cyclooxygenase-2 inhibitor celecoxib induces apoptosis by blocking Akt activation in human prostate cancer cells independently of Bcl-2. J Biol Chem 2000;275:11397-11403. doi:10.1074/jbc.275.15.11397.
. Effect of Celecoxib On REstenosis after coronary Angioplasty with a Taxus stent (COREA-TAXUS trial): an open-label randomised controlled study. Lancet 2007;370:567-574. doi:10.1016/S0140-6736(07)61295-1.
. Thrombosis in coronary drug-eluting stents: report from the meeting of the Circulatory System Medical Devices Advisory Panel of the Food and Drug Administration Center for Devices and Radiologic Health, December 7–8, 2006. Circulation 2007;115:2352-2357. doi:10.1161/CIRCULATIONAHA.107.688416.
. Randomized, double-blind, multicenter study of the Endeavor zotarolimus-eluting phosphorylcholine-encapsulated stent for treatment of native coronary artery lesions: clinical and angiographic results of the ENDEAVOR II trial. Circulation 2006;114:798-806. doi:10.1161/CIRCULATIONAHA.105.591206.
. Drug-eluting stenting followed by cilostazol treatment reduces late restenosis in patients with diabetes mellitus the DECLARE-DIABETES Trial (A Randomized Comparison of Triple Antiplatelet Therapy with Dual Antiplatelet Therapy After Drug-Eluting Stent Implantation in Diabetic Patients). J Am Coll Cardiol 2008;51:1181-1187. doi:10.1016/j.jacc.2007.11.049.
. Cardiovascular risk and inhibition of cyclooxygenase: a systematic review of the observational studies of selective and nonselective inhibitors of cyclooxygenase 2. J Am Med Assoc 2006;296:1633-1644. doi:10.1001/jama.296.13.jrv60011.
. Risk of acute myocardial infarction and sudden cardiac death in patients treated with cyclo-oxygenase 2 selective and non-selective non-steroidal anti-inflammatory drugs: nested case–control study. Lancet 2005;365:475-481.
. Rationale, design, and governance of Prospective Randomized Evaluation of Celecoxib Integrated Safety versus Ibuprofen Or Naproxen (PRECISION), a cardiovascular end point trial of nonsteroidal antiinflammatory agents in patients with arthritis. Am Heart J 2009;157:606-612. doi:10.1016/j.ahj.2008.12.014.