Author + information
- Received June 5, 2016
- Revision received August 28, 2016
- Accepted September 27, 2016
- Published online December 12, 2016.
- Gro Egholm, MD, PhDa,b,∗ (, )
- Steen Dalby Kristensen, MD, DMSca,
- Troels Thim, MD, PhDa,
- Kevin K.W. Olesen, MDa,
- Morten Madsen, MScb,
- Svend E. Jensen, MD, PhDc,
- Lisette O. Jensen, MD, PhD, DMScd,
- Henrik T. Sørensen, MD, PhD, DMScb,
- Hans E. Bøtker, MD, PhD, DMSca and
- Michael Maeng, MD, PhDa
- aDepartment of Cardiology, Aarhus University Hospital, Aarhus, Denmark
- bDepartment of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark
- cDepartment of Cardiology, Aalborg University Hospital, Aalborg, Denmark
- dDepartment of Cardiology, Odense University Hospital, Odense, Denmark
- ↵∗Reprint requests and correspondence:
Dr. Gro Egholm, Department of Cardiology, Aarhus University Hospital, Palle Juul-Jensens Boulevard 99, 8200 Aarhus N, Denmark.
Background Guidelines recommend postponing surgery for at least 6 months after treatment with a drug-eluting stent by percutaneous coronary intervention (DES-PCI).
Objectives The goal of this study was to evaluate the surgical risk associated with DES-PCI compared with that in nonstented patients without ischemic heart disease (IHD).
Methods Between 2005 and 2012, a total of 22,590 patients underwent DES-PCI in western Denmark. By record-linking the Western Denmark Heart Registry and the Danish National Patient Register, we evaluated 4,303 DES-PCI–treated patients with a surgical procedure and compared them with a control group of patients without previous IHD undergoing similar surgical procedures (n = 20,232). Events of interest were myocardial infarction (MI), cardiac death, and all-cause mortality within 30 days after surgery.
Results Surgery in DES-PCI–treated patients was associated with an increased risk of MI (1.6% vs. 0.2%; odds ratio [OR]: 4.82; 95% confidence interval [CI]: 3.25 to 7.16) and cardiac death (1.0% vs. 0.2%; OR: 5.87; 95% CI: 3.60 to 9.58) but not all-cause mortality (3.1% vs. 2.7%; OR: 1.12; 95% CI: 0.91 to 1.38). When stratified for time from PCI to surgery, only surgery within the first month was associated with a significant increased risk of events.
Conclusions Patients requiring surgery within 12 months after DES-PCI had an increased risk of MI and cardiac death compared with patients without IHD. The increased risk was only present within the first month after DES-PCI, suggesting that surgery might be undertaken earlier than currently recommended.
Drug-eluting coronary stent (DES) implantation has become the standard of care in patients undergoing percutaneous coronary intervention (PCI) (1). Compared with bare-metal stent (BMS) implantation, earlier generations of DES-PCI delay the healing process after stent implantation (2,3). Consequently, patients are instructed to take dual antiplatelet therapy (DAPT) for 6 to 12 months to reduce the risk of stent thrombosis (4,5). A systematic review has estimated that approximately 4% to 15% of all patients undergoing DES-PCI need surgery within 12 months after stent implantation (6). The risk of adverse cardiac events ranges from 3% to 11% depending on the definition of adverse cardiac events, the timing of surgery after DES-PCI, and the different types of surgery (6–9).
Surgery within the first 6 months after DES-PCI is associated with an increased risk of adverse events compared with surgery performed >6 months after DES-PCI (10,11). Therefore, current guidelines from the European Society of Cardiology (ESC) recommend postponing elective surgery for 6 to 12 months after DES-PCI for stable coronary artery disease and for 12 months in patients with acute coronary syndrome (ACS) (4). U.S. guidelines recommend delaying surgery for 12 months after DES-PCI unless the risk associated with surgical delay is considered larger than the risk of stent thrombosis (5).
Using Danish population-based registries and individual-based record linkage of Danish registries, we present the results of a population-based comparison of the risk associated with surgery among DES-PCI–treated patients compared with patients without ischemic heart disease (IHD) undergoing a similar operation.
We conducted a matched-cohort study on the basis of patient-level record linkage of the Western Denmark Heart Registry (WDHR) and the Danish National Patient Register (DNPR) evaluating patients with DES-PCI undergoing a surgical procedure within 12 months after the index DES-PCI procedure. The study compared the risk of myocardial infarction (MI), cardiac death, and all-cause mortality between DES-PCI–treated patients and patients without IHD, sampled from the general population, who underwent the same surgical procedure.
Study setting and population
All patients treated with a DES-PCI procedure between May 2005 and January 2012 were identified by using the WDHR (12). The registry assembles patient and procedural information from all coronary procedures performed at the 3 large PCI centers in western Denmark (Odense University Hospital, Aarhus University Hospital, and Aalborg University Hospital) covering the region’s population of approximately 3 million inhabitants, corresponding to 55% of the Danish population.
For each patient, the first PCI procedure performed during the study period was included as the index procedure. Because >90% of the PCI procedures at our centers are currently performed with DES, patients treated with balloon dilation alone or with BMS were not included.
The Danish Civil Registration System provides each citizen with a unique personal identification number on birth or immigration; we were thus able to collect individual-level linked information throughout the national and regional administrative registries (13). Data from WDHR were linked with data from the DNPR to identify patients who required surgery within 12 months after the DES-PCI. The DNPR contains data on all Danish hospital admissions since 1995, including dates of hospitalization, outpatient visits, and discharge diagnosis/surgical procedures coded according to the International Classification of Diseases-10th Revision (ICD-10). The Danish National Health Service provides tax-supported health care, guaranteeing access to general practitioners and hospitals, as well as partial reimbursement for prescribed medications. This approach ensures a high coverage in the registries used (14).
Surgery was categorized as low- to high-risk procedures according to the latest version of the ESC guidelines on noncardiac surgery (4) and could be performed under local, spinal, or general anesthesia. The ESC guidelines recommend risk assessment on the basis of estimates of 30-day risk of MI and cardiovascular death after noncardiac surgery. In this study, cardiac surgery was considered a high-risk procedure. Noninvasive endoscopic procedures, closed orthopedic repositions, heart transplantations, or insertion of left ventricular assist devices were not included. Surgical procedures were grouped according to the surgery performed (endocrine, ophthalmic, otorhinolaryngologic, oral and maxillofacial, heart and central vascular; respiratory, mammary, abdominal surgery minor/major, urologic minor/major, gynecologic, orthopedic minor/major, peripheral vascular, dermatologic, or neurosurgery) (Online Table 1).
Patients without IHD were identified by using the DNPR and sampled from the general population. First, we included all patients with no history of IHD, defined as diagnosis ICD-10 codes I20 to I25 and without a prescription for anticoagulation therapy. Second, we matched DES-PCI–treated patients versus patients without IHD. Each DES-PCI–treated patient was matched individually with at least 1 and up to 5 patients without IHD. The matching was based on type of surgical procedure according to surgical procedure codes (ICD-10), inpatient/outpatient admission at surgery, age (±5 years), and sex. DES-PCI–treated patients undergoing anticoagulation therapy (warfarin and novel oral anticoagulant agents) were not included because these patients were likely in a different risk category.
Patient characteristics and event definitions
The stents used in the cohort were both first-generation DES; sirolimus-eluting stent (Cypher, Cordis, Johnson & Johnson, New Brunswick, New Jersey) and the paclitaxel-eluting stent (Taxus, Boston Scientific Corp., Marlborough, Massachusetts) and second-generation DES; zotarolimus-eluting stent (Endeavor, Medtronic, Minneapolis, Minnesota) and the everolimus-eluting stent (Xience V, Abbott Vascular, or PROMUS, Abbott's privately labeled Xience V Everolimus Eluting Coronary Stent, Boston Scientific Corp.) and second-generation DES. Demographic and patient characteristics were retrieved from the WDHR and the Danish Register of Medical Product Statistics for DES-PCI–treated patients; the latter contains information on all prescriptions dispensed in Danish pharmacies since 1994. Patients were considered to be in treatment if a prescription was redeemed 100 days before surgery. Demographic and patient characteristics were retrieved for DES-PCI–treated patients and for patients without IHD. The Charlson Comorbidity Index was used to quantify comorbidity in the 2 groups, and data were retrieved from the DNPR (15). We used both the Charlson Comorbidity Index and a modified Charlson Comorbidity Index, excluding peripheral artery disease (PAD), diabetes mellitus, and renal disease, allowing these factors to be evaluated separately.
MI within the first 30 days’ post-surgery was defined as the listing of MI as the primary or secondary discharge diagnosis (ICD-10 code I21) and acute hospital admission collected from the DNPR. This definition of MI has a sensitivity of 94% and a specificity of 98% (16). The discharge diagnosis of MI is decided by the discharging physician in accordance with the universal definition of MI (17). Cardiac death within the first 30 days after surgery was defined as death resulting from IHD, sudden cardiac death, death from ventricular tachycardia, death caused by heart failure, and sudden death undefined (ICD-10 codes I20 to I25, I46, I47.2, I50, and R96, respectively); this information was retrieved from the Danish Registry of Causes of Death (18). All other deaths were recorded as noncardiac deaths.
All-cause mortality was identified within 30 days after surgery from the Civil Registration System (19). Events were included from any Danish hospital.
Events were summarized as 30-day cumulative incidence and reported as counts (percentages). Patient demographic and surgical characteristics were stated as counts (percentages), mean ± SD, and age and time from PCI to surgery as median (interquartile range). Crude and adjusted odds ratios with 95% confidence intervals were computed for each adverse event (MI, cardiac death, and all-cause mortality).
To evaluate risk factors in DES-PCI–treated patients, bivariate logistic regression models were used. When assessing the impact of time from index PCI to surgery, Kaplan-Meier curves were determined to evaluate events as a function of time. Thereafter, we divided time from PCI to surgery according to period, as follows: 0 to 30 days, >1 to 2 months, >2 to 3 months, >3 to 6 months, >6 to 9 months, and >9 to 12 months (reference) for patients treated with DES-PCI. Subsequently, the information from these analyses was used to dichotomize time from index PCI to surgery into ≤1 month versus >1 to 12 months when evaluating risk factors in DES-PCI–treated patients.
Cumulative incidence proportion was calculated for patients without IHD matched with DES-PCI–treated patients with surgery within 0 to 30 days, >1 to 2 months, and >2 to 12 months after stent implantation. Patient-level matching was maintained for each of these groups; that is, a specific control subject was matched with a specific DES-PCI–treated patient and, thus, only appears in 1 of these groups Conditional logistic regression analyses were used to assess the risk of DES-PCI in the surgical setting. Covariates used in the comparative models were emergency versus elective surgery; they were further adjusted for PAD, diabetes mellitus, and renal disease, and categorized for the Charlson Comorbidity Index.
Statistical analyses were performed by using Stata version 13 (Stata Corp., College Station, Texas) and SAS software version 9.4 (SAS Institute, Inc., Cary, North Carolina). The study was approved by the Danish Data Protection Agency. Registry-based studies do not require ethical approval in Denmark.
Among 22,590 DES-PCI–treated patients, 5,157 (23%) individuals undergoing surgery within 12 months of the index PCI were identified. DES-PCI–treated patients receiving oral anticoagulation therapy (n = 430) were excluded, yielding a total of 4,727 patients treated with DES-PCI who were eligible for matching with patients without IHD. Because the comparison was limited to patients without IHD, 424 patients undergoing coronary artery bypass grafting were also excluded (Figure 1). For this unmatched DES-PCI population, the rates of MI (n = 11), cardiac death (n = 9; information limited to 408 patients), and all-cause mortality (n = 20) were 2.6%, 2.2%, and 4.7%, respectively.
Demographic and patient characteristics for DES-PCI–treated patients (n = 4,303) and patients without IHD (n = 20,232) are listed in Table 1. There were major differences in drug therapy related to IHD, such as statins, beta-blockers, nitrates, proton pump inhibitors, P2Y12 inhibitors, and aspirin (the latter 2 comprising the components of DAPT). The surgical characteristics were matched, and the rate of emergency surgery was 25.6%. In the patients treated with DES-PCI, 61% of DES used were first generation, and 56% of the DES-PCI procedures were performed in patients with ACS. Comorbidity, as assessed by using the Charlson Comorbidity Index, was more frequent in patients treated with DES-PCI, of whom 76.9% had at least 1 comorbidity compared with 27.6% in patients without IHD (Table 2). Thus, the DES-PCI–treated patients had more comorbidities than patients without IHD. The different surgical fields and their related events are displayed in Online Table 2.
Risk factors and risks with surgery
First, we evaluated the possible risk factors associated with events among the patients treated with DES-PCI. Surgery was more common within the first 3 months (34%) than at >3 to 6 months (25%), >6 to 9 months (21%), and >9 to 12 months (20%) after index PCI. Time from index PCI was significantly associated with the risk of events. The risks of surgery for each of the first 3 months as well as for the periods >3 to 6 months, >6 to 9 months, and using surgery >9 to 12 months after DES-PCI as reference, are presented in Online Table 3, which shows that surgery within the first month after DES-PCI was associated with a significantly increased risk of all events. Beyond the first month, however, no significant differences were observed compared with surgery at >9 to 12 months. According to the unadjusted ORs of clinical, PCI-related, or surgical risk factors related to post-operative adverse events (Table 3), the factors associated with the highest risk of events were timing of surgery (≤1 month vs. >1 to 12 months), emergency surgery (acute vs. elective), and ACS as an indication for DES-PCI, PAD, and renal disease.
Compared with certain stent-related factors (DES generation, total stent length per patient, and number of stents implanted per patient), diabetes mellitus, sex, and Charlson Comorbidity Index score were less strongly associated with events (Table 3).
We also evaluated the risk of surgery and subsequent events in DES-PCI–treated patients compared with patients without IHD (Table 4). Surgery in DES-PCI–treated patients was associated with an increased risk of MI (64 [1.6%] vs. 49 [0.2%]; OR: 4.82; 95% CI: 3.25 to 7.16) and cardiac death (44 [1.0%] vs. 31 [0.2%]; OR: 5.87; 95% CI: 3.60 to 9.58) but not all-cause mortality (137 [3.1%] vs. 551 [2.7%]; OR: 1.12; 95% CI: 0.91 to 1.38). After adjusting for emergency surgery, the risk of MI and cardiac death increased among DES-PCI–treated patients compared with patients without IHD, whereas all-cause mortality remained almost unaffected. Further adjustment for comorbidity by using the modified Charlson Comorbidity Index with separate inclusion of PAD, diabetes mellitus, and renal disease had very little effect on the risk estimates.
When stratifying for timing of surgery in relation to index DES-PCI, surgery performed within 1 month after index DES-PCI was associated with an increased risk in DES-PCI–treated patients compared with patients without IHD. However, the risk associated with surgery was also increased in patients without IHD undergoing similar operations with a matched degree of urgency, indicating that part of the increased risk was related to the type and urgency of surgery and not solely explained by surgery early after PCI (Central Illustration, Figure 2).
Comparing patients treated with PCI >1 to 2 months and >2 to 12 months versus their respectively matched patients without IHD, we found that surgery in these groups had a low and similar risk (Central Illustration, Figure 2).
The present study compared risk associated with surgery in patients treated with DES-PCI versus patients without IHD, the latter sampled from the general population after excluding patients with previous IHD. The main findings were as follows: 1) surgery in patients treated with DES-PCI was associated with an increased 30-day risk of MI; 2) surgery among DES-PCI–treated patients did not increase 30-day all-cause mortality; and 3) beyond the first month after stent implantation, DES-PCI–treated patients had the same perioperative risk as surgery in patients without IHD.
The current guidelines on noncardiac surgery from the ESC (4) and the American College of Cardiology/American Heart Association (5) recommend delaying surgery after DES-PCI for 6 to 12 months. Our results suggest that surgery after DES-PCI might be performed earlier without increased risk.
Surgery after DES-PCI
By evaluating a DES-PCI cohort alone and by comparing it with a matched cohort of patients without IHD, we confirmed that the risk of MI after surgery declined over time from DES-PCI (8,20,21). Similar to previous studies (10,11), we found that surgery performed within the first month after DES-PCI was associated with the highest risk. Others have reported that surgery-associated risk in DES-PCI–treated patients reached a stable level after 3 to 6 months (8,20,22,23). Such temporal comparisons are only valid if the operations performed within the specified time intervals are comparable or if a surgical comparison cohort is included. Because previous studies did not include a surgical comparison cohort, the results are potentially influenced by the type and urgency of the surgical procedures.
To overcome this limitation, we included a matched cohort of patients without IHD. This evaluation of the comparative risk associated with surgery in DES-PCI–treated patients versus patients without IHD revealed an increased overall risk for MI and cardiac death in the patients with previous DES-PCI. However, this difference was highly time dependent and limited to the first month after DES-PCI. Furthermore, the increased risk associated with surgery within the first month after DES-PCI was not solely associated with DES-PCI. Matching for the specific surgical procedure and the urgency of the operation showed that the increased risk associated with surgery after DES-PCI within the first month was in part related to these parameters. Nevertheless, our data confirmed that surgery, if possible, should be delayed for at least 1 month after DES-PCI. Beyond the first 30 days after PCI, the DES-PCI–treated patients exhibited the same low risk as patients without previous IHD, even though the DES-PCI–treated patients had a higher number of comorbid conditions. This novel finding suggests that these patients, in whom delaying surgery might be associated with pain or increased risk of morbidity or mortality, can safely undergo surgical procedures earlier than current guidelines recommend (4,5).
We have been unable to identify any reports using a similar comparative design in patients treated with DES-PCI. In particular, no other study has compared DES-PCI–treated patients versus a control group of patients without previous IHD. However, we did identify a single study, combining DES and BMS, which had a comparison group (21). In that study, data for patients with coronary stents implanted in Veteran Affairs hospitals from 2000 to 2010 were matched with Veteran Affairs Surgical Quality Improvement Program data to identify noncardiac surgery within 24 months of stent placement. Similar to our results, the investigators reported an increased risk of MI but no difference in mortality. The same group of investigators used Veteran Affairs data to compare patients with stents undergoing surgery within 2 years versus patients with stents but without surgery (22). The 2 groups had the exact similar risk of adverse cardiac events during 2 years of follow-up. Patients with stents, however, had a higher risk of adverse cardiac events within the 30-day post-operative period. As in our study, this increased risk was particularly present within the first 4 to 6 weeks.
The main difference between this latter study and our data was the use of a nonsurgical group with IHD (22) versus a surgical non-IHD matched control group. The result obtained from U.S. Veterans Affairs hospitals may not be as representative of an “all-comers” population as our study due to differences in sex (male subjects, 98% vs. 70%), presence of diabetes (49% vs. 17%), and probably a narrower social class (veterans vs. all social classes) (21). Our data are derived from a society with a tax-supported health care system with a guarantee for free access to general practitioners and hospitals. Clearly, therefore, the studies are performed in societies with different access to hospital care and surgery. Moreover, different strategies related to DAPT in relation to surgery are possible. However, the similarities of the results from the 2 studies performed in different settings supported the robustness of these findings.
The major strength of this study was our use of population-based registries, which enabled us to collect procedural and diagnostic information on patients without previous IHD as well as DES-PCI–treated patients. Similarly, the use of registries permitted inclusion of consecutive patients over an 8-year period.
Some limitations must be acknowledged. First, our study had a retrospective design despite being based on prospectively registered data. Second, data on timing of surgery might be affected by confounding by indication because the use of DAPT might have led to delay or even cancellation of some major surgical procedures (23). In addition, we did not have data on the antithrombotic therapy management during the surgery procedures for all 22,590 patients. We did, however, have data from a nested case-control study based on this cohort in which we evaluated the medical charts of 458 patients who had been treated with DES-PCI and underwent surgery during the subsequent 12 months (24). In the control group (n = 230), who did not differ significantly from patients with adverse cardiac events or reoperation for bleeding, 58% of patients received both aspirin and a P2Y12 inhibitor within 3 days before surgery, 3% received a P2Y12 inhibitor only, 13% aspirin only, and 28% neither aspirin nor a P2Y12 inhibitor. Also, the high rate of events within the first month of DES-PCI might have indicated that the concomitant therapy with antiplatelet therapy unmasked malignant conditions or that the PCI-related hospitalization revealed conditions that led to accelerated surgical procedures. Such comorbidities were not accounted for in the matching of DES-PCI–treated patients and patients without IHD, and this issue represents a potential bias.
Finally, the risk of being classified as a cardiac death was overall increased among patients with known IHD (DES-PCI–treated patients) compared with patients with no previous IHD. Without a difference in all-cause death, the difference in cardiac death (and noncardiac death) between patients treated with DES-PCI and patients without IHD is striking and is most likely caused by differential classification. It is unclear whether this difference was caused by too many deaths classified as cardiac among DES-PCI–treated patients, too few among patients without IHD, or a combination of these factors.
Patients requiring surgery within 12 months after DES-PCI had an increased risk of MI and cardiac death compared with patients without IHD. The increased risk was only present within the first month after DES-PCI, suggesting that surgery might be undertaken earlier than currently recommended.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: Although surgery within 30 days after implantation of DES increases the risk of MI, cardiac death, and all-cause mortality, surgery performed beyond this period but within 12 months after DES implantation demonstrated a 30-day mortality equivalent to that of patients without IHD.
TRANSLATIONAL OUTLOOK: Further studies are needed to define the optimal approach to antithrombotic therapy in patients undergoing noncardiac surgery in the 12 months after DES implantation.
For supplemental tables, please see the online version of this article.
This work was supported by unrestricted grants from TrygFonden, Knud og Edith Eriksens Mindefond, Snedkermester Sophus Jacobsen og hustru Astrid Jacobsens Fond, Region Midtjyllands Sundhedsvidenskabelige Forskningsfond, and the Department of Cardiology, Aarhus University Hospital. Dr. Egholm has received speaker honorarium from Bayer. Dr. Kristensen is supported by a senior research grant from the Novo-Nordic Foundation. Dr. Thim has received an educational grant from St. Jude Medical; and a travel grant from Biosensors. The work of Dr. Sørensen was funded by the Program for Clinical Research Infrastructure established by the Lundbeck Foundation and the Novo Nordisk Foundation and administered by the Danish Regions. Dr. Bøtker is a member of the advisory board of AstraZeneca and Novo Nordisk. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute coronary syndrome
- bare-metal stent(s)
- confidence interval
- dual antiplatelet therapy
- drug-eluting stent(s)
- Danish National Patient Register
- European Society of Cardiology
- International Classification of Diseases-10th Revision
- ischemic heart disease
- myocardial infarction
- odds ratio
- peripheral artery disease
- percutaneous coronary intervention
- Western Denmark Heart Registry
- Received June 5, 2016.
- Revision received August 28, 2016.
- Accepted September 27, 2016.
- American College of Cardiology Foundation
- Finn A.V.,
- Nakazawa G.,
- Joner M.,
- et al.
- Joner M.,
- Finn A.V.,
- Farb A.,
- et al.
- Kristensen S.D.,
- Knuuti J.,
- Saraste A.,
- et al.
- Levine G.N.,
- Bates E.R.,
- Bittl J.A.,
- et al.
- Cruden N.L.,
- Harding S.A.,
- Flapan A.D.,
- et al.
- Tokushige A.,
- Shiomi H.,
- Morimoto T.,
- et al.
- Sorensen H.T.,
- Sabroe S.,
- Olsen J.
- Egholm G.,
- Madsen M.,
- Thim T.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- Jaffe A.S.,
- et al.
- Helweg-Larsen K.
- Pedersen C.B.
- Saia F.,
- Belotti L.M.,
- Guastaroba P.,
- et al.
- Holcomb C.N.,
- Graham L.A.,
- Richman J.S.,
- et al.
- Holcomb C.N.,
- Graham L.A.,
- Richman J.S.,
- et al.
- Graham L.A.,
- Hollis R.H.,
- Richman J.S.,
- Hawn M.T.
- Egholm G.,
- Thim T.,
- Olesen K.K.,
- et al.