Author + information
- Received March 10, 2010
- Accepted March 18, 2010
- Published online September 1, 2010.
- Peter B. Berger, MD⁎,⁎ (, )
- Neal S. Kleiman, MD‡,
- Michael J. Pencina, PhD§,
- Wen-Hua Hsieh, PhD§,
- Steven R. Steinhubl, MD⁎,
- Allen Jeremias, MD∥,
- Ali Sonel, MD†,
- Kevin Browne, MD¶,
- Greg Barseness, MD#,
- David J. Cohen, MD, MSc⁎⁎,
- EVENT Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Peter B. Berger, Geisinger Center for Clinical Studies, 100 North Academy Avenue, MC 40-00, Danville, Pennsylvania 17822-4400
Objectives This study sought to determine the frequency of noncardiac surgery and adverse post-operative events among patients who recently received a drug-eluting stent (DES) following noncardiac surgery.
Background Little is known about frequency of and risks associated with noncardiac surgery after DES implantation.
Methods In the EVENT (Evaluation of Drug-Eluting Stents and Ischemic Events) registry, consecutive patients who underwent attempted stent placement at 42 hospitals between July 2004 and September 2005 were enrolled and followed for 1 year. In this study, we analyzed patients who received ≥1 DES to determine the frequency of noncardiac surgery and cardiac death, myocardial infarction, or stent thrombosis in the following week.
Results Among 4,637 DES recipients, 206 (4.4%) underwent major noncardiac surgery in the following year (median days to surgery: 179 [interquartile range 112 to 266 days; range 13 to 360 days]). Overall, stent use averaged 1.5 per patient. The most frequent operations were orthopedic (36%), abdominal (31%), and vascular (20%). Compared with patients who did not undergo surgery, those who did were older, more likely to be women, and have had a prior stroke; the frequencies of prior myocardial infarction, prior coronary artery bypass graft, and diabetes were similar, as were left ventricular ejection fraction and indication for percutaneous coronary intervention. In the 7 days after surgery, 4 patients had a cardiac death, myocardial infarction, or stent thrombosis (1.9% [exact 95% confidence interval (CI): 0.5% to 4.9%]). The risk of the composite outcome was increased 27-fold in the week following noncardiac surgery compared with any other week after stent implantation (hazard ratio [HR]: 27.3 [95% CI: 10.0 to 74.2], p < 0.001).
Conclusions The frequency of major noncardiac surgery in the year after DES placement is >4%. Although the overall risk of adverse outcomes was less than previously reported when surgery is performed months after DES placement, it is significantly increased in the week after major noncardiac surgery.
Coronary revascularization before noncardiac surgery is believed to decrease the peri- and post-operative risk in selected patients (1). However, the risk of coronary thrombosis appears to be increased in patients who undergo a noncardiac surgical procedure in the weeks and months after a percutaneous coronary intervention (PCI) in which a bare-metal stent is used (2,3). Little is known, however, about the frequency with which noncardiac surgery is required after a PCI procedure or the risk of ischemic events among patients undergoing noncardiac surgery shortly after placement of a drug-eluting stent (DES) (4).
Accordingly, we performed this analysis of the EVENT (Evaluation of Drug-Eluting Stents and Ischemic Events) registry to determine the frequency with which cardiac surgery is required in the year after PCI in which a DES is placed and the risk of adverse cardiac events in the week following the surgical procedure.
A detailed description of the EVENT registry has been published previously (5). Briefly, the EVENT registry was launched shortly after the approval of the first DES in the U.S. to assess the practice of PCI utilizing intracoronary stents of any kind. Unselected patients undergoing attempted stent implantation were enrolled. Four waves of enrollment of approximately 2,500 patients per wave were performed. Forty-two U.S. medical centers (74% academic, median PCI volume 1,400/year) participated in the first 2 waves. Clinical and procedural data were prospectively collected on standardized case report forms and submitted to a data coordinating center. Creatine kinase and creatine kinase-myocardial band (CK-MB) levels were assessed at baseline and every 8 h for a minimum of 2 post-procedure samples and assayed using each site's clinical laboratory. Patients were contacted by telephone 6 and 12 months after the index PCI; only 176 of 2,321 patients (7.6%) were lost to follow-up in Wave I, and 185 of 2,316 patients (8.0%) were lost to follow-up in Wave II. Events noted at follow-up for this analysis included cardiac death, myocardial infarction (MI), stent thrombosis, and performance of a noncardiac surgical procedure. The use of aspirin and clopidogrel was also assessed at each time point. The EVENT study protocol was approved by the ethical review committee at all participating institutions, and all patients provided written informed consent.
In this analysis, consecutive patients enrolled in the first 2 waves of EVENT who received ≥1 DES were analyzed to determine the frequency of noncardiac surgery in the year after the PCI procedure and the frequency of adverse events in the 7 days after the surgical procedure. The 7-day period was determined a priori, in advance of this analysis, in view of the fact that essentially all patients who undergo noncardiac surgery would be expected to be on dual antiplatelet therapy 1 week after surgery and that events occurring >7 days after surgery may not have been truly related to the surgical procedure. Patients who received a bare-metal stent in addition to a DES (n = 186) were included in the analysis; patients receiving only a bare-metal stent (n = 339, 6.7%), those in whom no stent could be implanted (n = 48, 0.9%), patients requiring coronary artery bypass surgery before discharge from the index PCI procedure (n = 16, 0.32%), and those who died during the index hospitalization (n = 1, 0.22%) were not included in this analysis. Only major noncardiac surgical procedures in which a significant surgical incision was required from which bleeding would result were included in this analysis. Patients who underwent coronary artery bypass surgery or valve surgery (n = 67), pacemaker and defibrillator placement (n = 46), and patients who underwent surgery whose nature could not be determined (n = 50) were prospectively excluded from this analysis. Patients who underwent minor surgical procedures (n = 27), such as minor dermatological procedures, endoscopic procedures, joint aspirations, and cataract surgery, were prospectively considered not to have undergone major noncardiac surgery and were included in the control group.
Definitions, adjudication of events, and study outcomes
The main outcome measure was a composite of cardiac death, MI, or definite stent thrombosis in the week after noncardiac surgery. The definition of MI unrelated to a revascularization procedure was a clinical syndrome compatible with MI associated with a typical rise and fall in either CK (with a positive CK-MB or troponin) or CK-MB to at least twice the upper limit of normal. Stent thrombosis was independently adjudicated and, for the purposes of this study, considered to have occurred when it met the Academic Research Consortium's definition of definite stent thrombosis (6).
Data are presented as means, with standard deviations, or medians, with 25th and 75th percentiles. Comparisons between the 2 groups were performed using a chi-square analysis for discrete variables and Student t test for continuous variables. Cox proportional hazards regression was performed with the composite of cardiac death, MI, or stent thrombosis as primary outcome and the presence of noncardiac surgery as time-dependent exposure, with a focus only on the first 7 days after the noncardiac surgery. We attempted to confirm the results of the proportional hazards model using exact methods and obtained similar results. In addition to a univariate model described previously, we ran the same model but adjusted for covariates measured at the time of stent implantation (and not updated during follow-up), which included age, sex, use of at least 1 Taxus stent (Boston Scientific, Natick, Massachusetts), congestive heart failure, creatinine >2.0 mg/dl, left anterior descending lesion location, bifurcation lesion, total stent length, number of lesions, and maximum balloon diameter. Similar analyses were conducted for cardiac death alone and cardiac death or MI as end points.
Baseline and procedural characteristics
Among the 4,637 DES recipients in the first 2 waves of the EVENT study, 206 patients (4.4%) underwent major noncardiac surgery in the first year after their index PCI. In only 7 of the 4,637 patients (0.15%) was a planned surgical procedure listed as an indication for the PCI.
The baseline characteristics of patients in the overall study population, and of patients who did and did not undergo major noncardiac surgery, are presented in Table 1. Overall, 34% had diabetes mellitus, 37% had had a prior MI, 21% had undergone a prior coronary artery bypass graft, 52% had multivessel disease, 33% had ≥1 lesion treated during the index procedure, and more than 90% of patients treated received a stent.
Compared with patients who did not undergo surgery, those who did were older (age 67 vs. 64 years, p < 0.001), more likely to be women (39% vs. 31%, p = 0.016), and have had a prior stroke (13% vs. 8%, p = 0.021). The frequencies of prior MI (39% vs. 37%), prior coronary artery bypass graft (25% vs. 20%), and diabetes (36% vs. 34%) were similar, as were left ventricular ejection fraction and the indication for PCI.
The timing of surgery, and the angiographic and procedural characteristics of the study population and in patients who did or did not undergo surgery, are displayed in Figure 1 and Table 2. Overall, stent use averaged 1.5 per patient; 53% of patients received a Taxus stent and 51% of patients received a Cypher stent (Cordis Corporation, Bridgewater, New Jersey); 7% received at least 1 bare-metal stent in addition to a DES.
The medications prescribed at discharge after the index PCI are listed in Table 3. Data about how aspirin and clopidogrel were managed in surgical patients were not collected in such a way that it could be determined which patients remained on therapy during the surgery and those who did not. However, patients who underwent noncardiac surgery were less likely to be taking aspirin and clopidogrel at 1 year (54% vs. 64%, p = 0.001).
The time to surgery is displayed in Figure 1. The median time to surgery was 179 days (112, 266) (range 13 to 360 days). The operations performed were orthopedic (36%), abdominal (31%), vascular (20%), ear/nose/throat (4%), thoracic (3%), urologic (1%), intracranial (1%), and other (4%).
Major adverse events
In the first 7 days after surgery, 4 of the 206 patients (1.9% [exact 95% confidence interval (CI): 0.5% to 4.9%]) suffered a major cardiac event: 1 died and 3 suffered an MI. No patient suffered definite stent thrombosis.
There was no difference in the timing of surgery among patients who did and did not suffer a cardiac event (median: 156 vs. 181 days, p = 0.40).
To determine whether cardiac risk was increased in the immediate peri-operative period, we performed a time-updated proportional hazards analysis in which the period of interest was the 7 days following noncardiac surgery. Among patients who did not undergo surgery, the risk of an event in any week of follow-up was 0.67% (exact 95% CI: 0.56% to 0.078%); the risk per week of cardiac death was 0.01%, of MI was 0.14%, and of stent thrombosis was 0.002%. In a univariate analysis, the performance of noncardiac surgery was associated with a 27-fold increased risk of ischemic complications during the week following surgery compared with any other week after stent implantation among patients who did not undergo surgery (hazard ratio [HR]: 27.3 [95% CI: 10.0 to 74.2], p < 0.001). The increased risk remained similar in multivariable-adjusted analyses (25.7 [95% CI: 9.4 to 70.0], p < 0.0001). Only congestive heart failure, creatinine >2.0 mg/dl, and the number of lesions treated were significantly related to the primary outcome of interest in addition to our main exposure (Table 4).
The most important findings of this study were that more than 4% of patients required a major surgical procedure in the year after placement of a DES. Among such patients, the risk of cardiac death, MI, or stent thrombosis was 1.9% (exact 95% CI: 0.5% to 4.9%) in the week following surgery, lower than suggested by previously reported case series when surgery was performed earlier, generally in the first 2 months, after stent placement. However, the risk of a major adverse event in the week after surgery in patients who had received a DES within the prior year was nonetheless much higher in a time-dependent proportional hazards analysis than in patients in whom surgery was not performed.
The first small study of patients undergoing noncardiac surgery after bare-metal stent placement reported a 34% mortality (8 of 25 patients) rate when surgery was performed in the 2 weeks after surgery; there were no deaths among the 15 patients who underwent surgery 3 to 8 weeks after bare-metal stent placement (2). A much larger study of 207 patients subsequently identified a 3% to 5% rate of major adverse events when surgery was performed anytime during the 6 weeks after stent placement; no adverse events occurred when surgery was performed beyond this time frame (3). These studies, and all such analyses, are inherently biased in that patients who required surgery more rapidly were undoubtedly sicker in 1 or more ways. It is interesting, however, that a similar analysis of the risk of noncardiac surgery among 197 patients who underwent pre-operative balloon angioplasty did not reveal a high rate of death or MI when surgery was performed shortly after the procedure (although repeat revascularization was required more frequently, as is expected when balloon angioplasty is compared with stent placement) (7).
Investigators from the Cleveland Clinic recently reported the clinical outcome of 606 inpatient surgeries performed on 481 patients with a mean time from PCI to surgery of 1.1 ± 0.9 years (8). Unlike the present study, many patients undergoing cardiac surgery were included in that analysis. Death, MI, or stent thrombosis occurred after 11 (2.0%) surgeries. The risk decreased significantly in the 1 to 6 months after PCI, although some risk persisted when the time between PCI and surgery was even as long 12 months. Independent correlates of the combined end point include emergency surgery, antecedent MI, the pre-operative use of intravenous heparin, and atherosclerotic lesion length treated with DES. Oral antiplatelet status at time of surgery was not a correlate of events.
The current study
Guidelines for the management of patients with coronary disease who require noncardiac surgery recommend that revascularization not be performed prior to a major surgical procedure unless there are strong indications for it, and that if revascularization is required prior to planned surgery, a bare-metal stent should be placed or even balloon angioplasty should be performed rather than placement of a DES (1). If a DES was placed before the need for a surgical procedure was recognized, it has been recommended that at least 1 year be allowed to elapse after the DES was implanted before surgery is performed and that dual antiplatelet therapy be continued perioperatively whenever the surgery can be safely performed during such therapy (1,9). Although these recommendations are entirely reasonable and consistent with the available data, our study is not able to directly address these recommendations. The main conclusions that can be derived from the current study are that the need for surgery in the year after PCI is >4%, emphasizing the importance of patient education prior to hospital discharge after a PCI procedure about the importance of consulting with one's cardiologist regarding the management of dual antiplatelet therapy in the perioperative period. The other main conclusion is that major noncardiac surgery appears to be safer than was believed when performed months after receipt of a DES and likely is safer than when surgery is performed within 2 months after stent placement. Our data do suggest, however, that the risk of a major adverse event in the week after surgery is increased 27-fold above patients not requiring surgery, even though in absolute terms, the risk (1.9% [exact 95% CI: 0.5% to 4.9%]) may be lower than previously believed.
In this analysis, when surgery was performed months after stent placement, there was no evidence that the duration between surgery and DES implantation influenced outcome. However, this relatively small study of 209 patients cannot determine the optimal duration of delay between stent placement and surgery, and the guidelines that recommend delaying entirely elective surgery for at least 1 year would appear to be prudent. Although there were no episodes of proven stent thrombosis among patients after noncardiac surgery, most patients with perioperative MI did not undergo coronary angiography; the possibility that some patients who suffered MI or even death may have suffered stent thrombosis as well cannot be excluded.
Perioperative medical management
Patients who underwent noncardiac surgery were significantly less likely to complete 1 year of dual antiplatelet therapy. Although we were not able to collect detailed data about the management of dual antiplatelet therapy at the time of each surgical episode, the finding that patients who underwent surgery were less likely to be taking clopidogrel 1 year after DES placement suggests that at least some patients probably had their clopidogrel discontinued at the time of surgery and that it was not restarted. Continued patient and physician education regarding the importance of restarting aspirin and clopidogrel (if these medications indeed must be discontinued pre-operatively) is required to prevent catastrophic events in this challenging patient cohort.
There are several important limitations to this analysis that physicians ought to keep in mind when deciding whether or to what extent this analysis ought to influence their practice patterns. The way in which the aspirin and clopidogrel were managed in these patients could not be gleaned from the EVENT registry data collection forms. Because aspirin and clopidogrel reduce the frequency of stent thrombosis, it is likely that the adverse event rate might be influenced by whether these medications were administered perioperatively; the event rate is expected to be higher if patients had not been continued on dual antiplatelet therapy perioperatively, and lower, possibly, if all had been continued on aspirin and clopidogrel. Cardiologists were likely aware, at the time of this study, of the reports of stent thrombosis after surgery in patients with recently placed DES. There may well have been selection bias toward the “healthiest” of patients who were “permitted” to undergo noncardiac surgery, and physicians may have been more aggressive about using aspirin and clopidogrel in such patients than described in previous reports. Definitions of stent thrombosis are all imperfect. It is likely that some cases of stent thrombosis were not recognized clinically; therefore, it is possible that the current data represent an under-reporting of the true frequency of stent thrombosis. Also, the risk of perioperative thrombotic events only partially captures the potential risk of DES-treated patients undergoing surgery. No data were captured to allow for a description of bleeding events in the post-operative period. The risk of performing surgery in patients with coronary disease is known to be higher than in patients without coronary disease; whether the rate of major adverse events is related to the recent implantation of a DES cannot be determined from this study. Randomized trials have indicated that such events are not prevented by the routine performance of revascularization in patients with stable coronary disease; some data suggest it may be increased. Lastly, few adverse events occurred in the study population, particularly among those who underwent surgery. The uncertainty of estimates of risk is reflected in the wide confidence intervals.
In the modern stent era, the frequency of major noncardiac surgery in the year after DES placement is nearly 5%, occurring in approximately 0.5% of patients per month after the first 2 months. Although the overall risk of adverse outcomes appears to be less than previously reported, it is significantly increased in the week after major noncardiac surgery. Improved patient and physician education is warranted to delay truly elective noncardiac surgery and to improve rates of compliance with recommendations for aspirin and clopidogrel therapy during or after the perioperative period when possible.
Dr. Berger's Council on Medical Education–approved scientific symposia have been supported by Bristol-Myers Squibb/Sanofi-Aventis, The Medicines Company, AstraZeneca, Medtronic, Eli Lilly/Daiichi-Sankyo. He has received research funding from Thrombovision, Helena, Accumetrics, AstraZeneca, Haemoscope, The Medicines Company, Corgenix/Aspirinworks, and Eli Lilly/Daiichi-Sankyo and has been a consultant to Accumetrics, The Medicines Company, Novartis/Portola, Guerbet, PlaCor, Eli Lilly/Daiichi-Sankyo; and he owns equity in Lumen Inc. (a company that is developing an embolic protection device). Dr. Kleiman has received research grants from Sanofi-Aventis, Cordis, Eli Lilly, and Schering-Plough. Dr. Pencina is a consultant with Harvard Clinical Research Institute which works with numerous industry clients and serves on DSMB of device/drug trials. Dr. Steinhubl is an employee with stock options of The Medicines Company. He has received research support from Regado Biosciences, The Medicines Company, and Eli Lilly/Daiichi Sankyo; and honoraria for serving on the advisory boards of Sanofi-Aventis/Bristol-Myers Squibb, Eli Lilly/Daiichi Sankyo, Cardax Pharmaceuticals, Portola Pharmaceuticals, and AstraZeneca. Dr. Cohen has received research grant support from Boston Scientific and Abbott Vascular and has received consulting fees from Cordis, Medtronic, Inc., and Eli Lilly.
- Abbreviations and Acronyms
- creatine kinase-myocardial band
- drug-eluting stent(s)
- myocardial infarction
- percutaneous coronary intervention
- Received March 10, 2010.
- Accepted March 18, 2010.
- American College of Cardiology Foundation
- Fleisher L.A.,
- Beckman J.A.,
- Brown K.A.,
- et al.
- Kaluza G.L.,
- Joseph J.,
- Lee J.R.,
- Raizner M.E.,
- Raizner A.E.
- Wilson S.H.,
- Fasseas P.,
- Orford J.L.,
- et al.
- Brilakis E.S.,
- Banerjee S.,
- Berger P.B.
- Anwaruddin S.,
- Askari A.T.,
- Saudye H.,
- et al.
- Grines C.L.,
- Bonow R.O.,
- Casey D.E. Jr..,
- et al.