Stratification is by stent indication of myocardial infarction (A), unstable angina (B), and revascularization not associated with acute coronary syndrome (non-ACS) (C). Solid lines indicate estimated rates; shaded areas, 95% CIs. Density plots on the right show percentage of operations by time between percutaneous coronary intervention and operation.
MI indicates myocardial infarction; non-ACS, revascularization not associated with acute coronary syndrome.
eFigure. Adjusted odds of MI by stent indication and time from PCI
Holcomb CN, Hollis RH, Graham LA, Richman JS, Valle JA, Itani KM, Maddox TM, Hawn MT. Association of Coronary Stent Indication With Postoperative Outcomes Following Noncardiac Surgery. JAMA Surg. 2016;151(5):462–469. doi:10.1001/jamasurg.2015.4545
Current guidelines for delaying surgery after coronary stent placement are based on stent type. However, the indication for the stent may be an important risk factor for postoperative major adverse cardiac events (MACE).
To determine whether the clinical indication for a coronary stent is associated with postoperative MACE.
Design, Setting, and Participants
Retrospective cohort study in patients at US Veterans Affairs hospitals who had a coronary stent placed between January 1, 2000, and December 31, 2010, and underwent noncardiac surgery within the following 24 months. The association between the indication for stent and postoperative MACE rates was examined using logistic regression to control for patient and procedure factors.
Three subgroups of stent indication were examined: (1) myocardial infarction (MI); (2) unstable angina; and (3) revascularization not associated with acute coronary syndrome (non-ACS).
Main Outcomes and Measures
Composite 30-day postoperative MACE rates including all-cause mortality, MI, or revascularization.
Among 26 661 patients (median [IQR] age, 68 [61.0-76.0] years; 98.4% male; 88.1% white) who underwent 41 815 surgical procedures within 24 months following coronary stent placement, the stent indication was MI in 32.8% of the procedures, unstable angina in 33.8%, and non-ACS in 33.4%. Postoperative MACE rates were significantly higher in the MI group (7.5%) compared with the unstable angina (2.7%) and non-ACS (2.6%) groups (P < .001). When surgery was performed within 3 months of percutaneous coronary intervention, adjusted odds of MACE were significantly higher in the MI group compared with the non-ACS group (odds ratio [OR] = 5.25; 95% CI, 4.08-6.75). This risk decreased over time, although it remained significantly higher at 12 to 24 months from percutaneous coronary intervention (OR = 1.95; 95% CI, 1.58-2.40). The adjusted odds of MACE for the unstable angina group were similar to those for the non-ACS group when surgery was performed within 3 months (OR = 1.11; 95% CI, 0.80-1.53) or between 12 and 24 months (OR = 1.08; 95% CI, 0.86-1.37) from stent placement. Stent type was not significantly associated with MACE regardless of indication.
Conclusions and Relevance
Surgery in patients with a coronary stent placed for MI was associated with increased postoperative MACE rates compared with other stent indications. The risk declined over time from PCI, and delaying surgery up to 6 months in this cohort of patients with stents may be important regardless of stent type.
Current clinical practice guidelines for optimal timing to delay noncardiac surgery following percutaneous coronary intervention (PCI) differs by stent type: 4 weeks following bare-metal stent (BMS) placement and 1 year after drug-eluting stent (DES) placement.1 The reason thought to drive this difference is the longer time to stent endothelialization in DES compared with BMS, which can increase the risk for stent thrombosis. However, large observational studies have shown that the highest risk for major adverse cardiac events (MACE) following surgery after recent PCI is confined to the first 6 months after stent placement with no significant difference in the rate of postoperative MACE by stent type.2- 5 Given that stent types do not explain the inverse relationship between time from PCI and perioperative MACE risk, we explored whether the indication for stent placement may provide additional information regarding the risk of MACE after noncardiac surgery.
Coronary stents are placed in different clinical scenarios, including acute coronary syndrome (ACS), which can be subdivided into either acute myocardial infarction (MI) or unstable angina. Additionally, coronary stents can be placed for stable angina. For patients undergoing surgery, a history of ACS with MI is a known risk factor for postoperative reinfarction6- 9 and current guidelines recommend delaying noncardiac surgery for at least 6 weeks from the time of MI.1 Studies predating PCI found that the risk of reinfarction extended to 6 months following MI,6 and more contemporary studies have determined that the risk for postoperative reinfarction extends beyond the recommended 6-week post-MI delay.10,11 However, coronary revascularization may decrease the risk.12 Few studies have compared the risk associated with surgery following coronary stent placement based on the indication for the coronary stent.4
To better understand the factors contributing to cardiac risk in patients who have undergone recent PCI and require noncardiac surgery, we comparatively examined the postoperative MACE associated with 3 distinct subgroups of stent indication: (1) MI; (2) unstable angina; and (3) non-ACS revascularization. We hypothesized that risk for postoperative MACE would be associated with the acuity of presentation at the time of PCI prior to surgery, with MI carrying the highest risk, followed by unstable angina and then non-ACS stent indication regardless of stent type.
We conducted a retrospective cohort study of patients undergoing noncardiac surgery at US Veterans Affairs (VA) hospitals within 2 years following coronary stent implantation to examine the relationship between indication for coronary stent implantation and postoperative outcomes of MI, revascularization, or death. The study protocol was reviewed and approved by the local institutional review boards of the Birmingham VA Hospital, the VA Boston Healthcare System, and the VA Eastern Colorado Health Care System. The requirement for informed consent was waived owing to the retrospective nature of the study.
Coronary stents were identified in the National Patient Care Database and the VA Clinical Assessment, Reporting, and Tracking Program data. Noncardiac surgery occurring in the VA was obtained from VA Surgical Quality Improvement Program data. Data from the Centers for Medicare and Medicaid Services acquired through the VA Information Resource Center was used to identify non-VA noncardiac surgical procedures. Demographic characteristics, comorbidities, and operative characteristics were obtained from the National Patient Care Database and the Centers for Medicare and Medicaid Services. Date of death was obtained from the VA Vital Status File.
We identified all coronary stents implanted in the VA between January 1, 2000, and December 31, 2010, using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes (36.06 for BMS or 36.07 for DES) and direct abstraction from the VA Clinical Assessment, Reporting, and Tracking Program data files. The PCI care episodes were defined as a single visit to the catheterization laboratory for a PCI procedure, where 1 or more stents were implanted. Noncardiac surgical procedures were defined using Current Procedural Terminology (CPT) codes 10000 to 32999 and 34000 to 69999. We excluded minor operations such as endoscopic procedures (CPT codes 43200-43272, 45300-45392, 46600-46608) and minor musculoskeletal procedures such as application of a cast and joint aspiration (CPT codes 29000-29750). Operations performed under local or monitored anesthesia were excluded from analyses. The level of analysis was individual surgical procedures. Patients who underwent coronary artery bypass grafting or had evidence of PCI performed at a non-VA hospital following the index PCI performed at a VA hospital were excluded. If more than 1 surgical procedure occurred in a hospitalization, only the first surgical procedure was included and subsequent surgical procedures within 30 days were excluded. Only 3.5% of patients had separate surgical procedures following different PCI episodes. Further details on the construction of the study cohort have been published.2
The main dependent variable for our study was the occurrence of MACE defined as MI, revascularization, or death within 30 days of noncardiac surgery. Postoperative MI was defined as the occurrence of an ICD-9-CM diagnosis code for MI either during the hospital stay but not the principal ICD-9-CM code for the admission or following discharge but still within 30 days of surgery (ICD-9-CM codes 410.x1 or VA Surgical Quality Improvement Program nurse-abstracted ST-segment elevation MI). Revascularization was defined similarly using ICD-9-CM procedure codes or CPT codes (CPT codes 92980-92982, 92984, 92995, 92996; ICD-9 procedure codes 00.66, 36.01, 36.02, 36.05, 36.06, 36.07). An analysis of the secondary end point of MI to examine the risk of reinfarction was performed.
The independent variable of interest for this analysis was the indication for the coronary stent placement. Stent indication was determined using the primary ICD-9-CM diagnosis code for the coronary stent placement procedure. Coronary stents placed for MI were defined by ICD-9-CM codes 410.xx, while coronary stents placed for chest pain (ICD-9-CM code 786.50 or 786.59) or intermediate coronary syndrome (ICD-9-CM codes 411.xx) were defined as unstable angina indications. All other stent indications were defined as non-ACS. Patients with codes for both unstable angina and MI were coded as having an MI stent indication. Patients who had an MI or unstable angina episode following the PCI but preceding surgery were excluded from analyses. This resulted in the exclusion of 2552 operations.
The type of noncardiac surgery was classified using the primary CPT code as the following: integumentary, 10040 to 19999; musculoskeletal, 20000 to 29999 (except amputation reclassified under vascular); respiratory, 30000 to 32999; vascular, 34000 to 37799, plus 27290, 27295, 27598, 27880 to 27899, and 28801 to 28825; digestive, 40000 to 49999; genitourinary, 50000 to 58999; nervous, 61000 to 64999; or eye/ear, 65000 to 69999. Procedures with CPT codes not listed herein were categorized as other. The 2011 Centers for Medicare and Medicaid Services work relative value unit for the primary CPT code was used to measure procedure complexity.
To quantify the patient’s cardiac risk at the time of surgery, we calculated the 6-point Revised Cardiac Risk Index score using administrative ICD-9-CM diagnosis codes for congestive heart failure, stroke, MI, and diabetes along with laboratory data for serum creatinine level and CPT codes for high-risk surgery classification.13 An insulin prescription in the VA Decision Support System Pharmacy data within 12 months of surgery was used to identify insulin-dependent diabetes in patients with an ICD-9-CM code for diabetes. Patients received a point if they had 1 or more serum creatinine values higher than 2 mg/dL (to convert to micromoles per liter, multiply by 88.4) in the year prior to surgery. The Revised Cardiac Risk Index score was analyzed as both an ordinal and categorical variable: low risk (1 point), moderate risk (2 points), or high risk (≥3 points). Additional comorbidities at the time of surgery were identified via VA Surgical Quality Improvement Program nurse abstraction or queried from the VA National Patient Care Database using ICD-9-CM diagnosis codes.
Univariate and bivariate frequencies were examined to determine factors associated with coronary stent indication. Unadjusted frequencies were compared using χ2 test. The most parsimonious adjusted logistic model for postoperative MACE was developed using backward stepwise selection including variables found to be associated with both stent indication and MACE in bivariate analyses. To develop the model, a spline-fit term was used to account for the nonlinear association of MACE and time since stent placement.2,14 An interaction term for stent indication and time since stent placement was included in the final adjusted model to assess differences in MACE by stent indication. Stent type, indicating whether the patient had DES, BMS, or both stent types, was forced into the model because of prior clinical relevance. To analyze MACE rates associated with defined periods, time from PCI was categorized into 0 to 3, 3 to 6, 6 to 12, and 12 to 24 months based on risk patterns observed on unadjusted smooth plots. Additional covariates in the final model for MACE included age, history of congestive heart failure in the prior 6 months, acute renal failure, Revised Cardiac Risk Index, number of PCI episodes in the prior 2 years, history of peripheral vascular disease in the prior 2 years, relative value units as a spline-fit term, elective case status, and procedure type. The same methods were applied to examine the outcome limited to postoperative MI. The statistical threshold for significance was set at P = .05 for a 2-tailed test.
All univariate, bivariate, and multivariate statistics were calculated using SAS version 9.2 statistical software (SAS Institute, Inc). Smoothed plots for postoperative MACE by time between surgery and stent placement were created with R package GGPLOT2.15 Smoothed trends were fitted using the loess algorithm.
A total of 26 661 patients (median [IQR] age, 68 [61.0-76.0] years; 98.4% male; 88.1% white) met study inclusion criteria and underwent 41 185 surgical procedures within 24 months of PCI. Patient demographic characteristics and comorbidities are shown in Table 1, and stent and surgical characteristics are shown in Table 2. Among the 41 185 surgical procedures, 32.8% were performed after a stent was placed for MI, 33.8% occurred following an unstable angina stent indication, and the remaining 33.4% occurred following a non-ACS stent indication. Overall, the MI group was sicker with higher Revised Cardiac Risk Index scores and higher rates of congestive heart failure and acute renal failure within 6 months of surgery. The majority of surgical procedures were outpatient (65.8%), followed by elective inpatient procedures in 29.2% and nonelective procedures in only 5.0% of the admissions. Patients with non-ACS stent indication had higher rates of surgery performed within 3 months of stent placement (14.1%) compared with MI (11.7%) and unstable angina (11.0%) stent indications.
The 30-day postoperative MACE rate in the cohort was 4.2% (Table 3). Overall, 2.3% of the cohort experienced a postoperative MI, 1.0% underwent coronary revascularization, and 1.6% died within 30 days of surgery. The MI stent indication group had significantly higher MACE rates (7.5%) compared with the unstable angina (2.7%) and non-ACS (2.6%) groups (P < .001) (Table 3). The MI group had a 5.0% reinfarction rate after surgery, 1.3% underwent a revascularization procedure, and 2.2% died within 30 days of surgery. The unstable angina group and non-ACS groups had similar rates of MI (1.0% vs 0.9%, respectively) but the unstable angina group had a lower mortality rate (1.1% for the unstable angina group; 1.4% for the non-ACS group).
Postoperative MACE rates by time between stent placement and surgery are shown in Figure 1 for each stent indication. Patients with MI stent indication showed significantly higher risk of MACE with surgery within 3 months of PCI (22.2%) compared with 3 to 6 months (9.4%), 6 to 12 months (5.8%), or 12 to 24 months (4.4%) (P < .001). Surgery close to the time of PCI was also associated with higher MACE rates, although attenuated, in patients with unstable angina (<3 months, 4.7%; 3-6 months, 3.3%; 6-12 months, 2.9%; and 12-24 months, 2.1%; P < .001) or non-ACS stent indications (<3 months, 4.6%; 3-6 months, 3.5%; 6-12 months, 2.2%; and 12-24 months, 2.1%; P < .001).
The adjusted odds of MACE by stent indication and time from PCI are shown in Figure 2. For patients with MI stent indication compared with patients with non-ACS stent indication, the adjusted odds of MACE were significantly higher when the time between surgery and stent placement was within 3 months (odds ratio [OR] = 5.25; 95% CI, 4.08-6.75) and decreased after 3 months, although remaining significantly higher at each interval (3-6 months: OR = 2.45; 95% CI, 1.80-3.35; 6-12 months: OR = 2.50; 95% CI, 1.90-3.28; and 12-24 months: OR = 1.95; 95% CI, 1.58-2.40). The adjusted odds of MACE were similar for patients with unstable angina and non-ACS stent indications at each interval (<3 months: OR = 1.11; 95% CI, 0.80-1.53; 3-6 months: OR = 0.99; 95% CI, 0.68-1.45; 12-24 months: OR = 1.08; 95% CI, 0.86-1.37), although they were marginally significantly higher in the unstable angina group at 6 to 12 months from PCI (OR = 1.36; 95% CI, 1.00-1.85). When only the outcome of MI was analyzed, similar trends were observed (eFigure in the Supplement). Stent type was not a significant predictor of postoperative MACE (P = .06) or MI (P = .14).
In this study, we assessed the risk for MACE following noncardiac surgery in patients who underwent recent coronary stent placement by stent indication. We found that patients with coronary stents placed for MI had the greatest risk of MACE in the 30 days following surgery, and this was mostly attributable to an increased risk of reinfarction. When compared with patients with non-ACS stent indications, the MI group had significantly increased odds of MACE at each interval out to 2 years, with the highest risk occurring within 3 months of PCI. Postoperative MACE rates were similar in the patients who had stents placed for unstable angina and non-ACS indications when surgery occurred at varying points from PCI. Stent type was not an independent predictor of postoperative MACE.
The risk of surgery in patients with prior MI was previously described in a 1972 Mayo Clinic study that showed a 37% risk of reinfarction when surgery occurred within 3 months of MI and a 16% risk when surgery was performed within 3 to 6 months of MI.6 Although rates of perioperative reinfarction have decreased with advancements in perioperative care,16 a more recent study has shown that the rates of reinfarction remain at 32.8% when noncardiac surgery occurs within 30 days of MI and 5.9% when surgery is performed at 3 to 6 months.10 Examining a similar population, a parallel study found that coronary revascularization was associated with a trend for decreased rates of postoperative reinfarction only in patients who had surgery performed more than 6 months following MI.12 In our study of revascularized patients, PCI placed for MI resulted in postoperative MACE rates of 22.2% when surgery was within 3 months of MI and 9.4% within 3 to 6 months. Reinfarction accounted for most of the MACE events. Taken together, these findings suggest that delaying elective surgery beyond 6 months after MI may be more appropriate compared with the currently recommended 6 weeks, regardless of coronary revascularization.
For patients with coronary stents undergoing surgery, this study sheds insight into the importance of stent indication as a predictor of perioperative risk. In a single-center study of 1923 nonsurgical patients who received PCI, the risk of all-cause mortality in the 6 months following PCI was shown to be significantly higher in patients who underwent PCI for ST-segment elevation MI, but not non–ST-segment elevation MI or unstable angina, when compared with patients with PCI for stable coronary artery disease.17 To our knowledge, only 1 study has assessed perioperative risk of patients with coronary stents by stent indication.4 In this 2010 study using the Scottish Coronary Revascularisation Register, researchers found significantly higher risk of postoperative in-hospital death or MI when patients underwent surgery within 1 month of PCI for ACS (either with or without MI) when compared with patients with PCI for stable coronary artery disease (65% vs 32%).4 Our results support this finding and show that patients who have stents placed for ACS with MI (both ST-segment elevation MI and non–ST-segment elevation MI) are at significantly higher risk compared with patients who had stents placed for non-MI indications, with the highest risk occurring in the first 3 months after PCI. It is unclear whether this finding may be related to greater coronary disease burden, incomplete revascularization, increased susceptibility of the myocardium, or other potential factors.
These findings have significant implications for guidelines regarding the timing of surgery in patients with coronary stents. Current recommendations are based on stent type and do not discuss stent indication: patients with DES should ideally have elective surgery delayed for at least 12 months to avoid interruptions in dual antiplatelet therapy and patients with BMS should have surgery delayed at least 30 days.1 While the concern remains regarding interruptions in antiplatelet therapy and adverse cardiac events, in a nested case-control analysis based on this study’s cohort, we did not find an association between antiplatelet therapy cessation and MACE.2 In addition, other studies have demonstrated that temporary cessation vs continued antiplatelet therapy perioperatively does not significantly alter the risk for subsequent MI or mortality.18- 20 Stent type was not a significant predictor of MACE in this study or other recent large cohort studies.2,4,5 Rates of MACE declined significantly during the first 6 months for the MI stent indication group, while the unstable angina and non-ACS groups showed little change in MACE risk over time from PCI. Thus, it may be important to delay surgery up to 6 months in patients with stents placed for MI regardless of stent type and may be reasonable to operate sooner on patients with stents placed for non-MI indications.
Our findings should be interpreted in light of the observational nature of our study design. First, we do not have information regarding the coronary anatomy, the extent of the vessel disease burden, or echocardiographic findings such as ejection fraction or valvular dysfunction. Second, we used administrative codes to define stent indication subgroups likely leading to misclassification. For example, because the classification is based on the indication at the time of the stent placement, non-MI stent indications may sometimes reflect the occurrence of a delayed revascularization following MI. However, this would bias our results toward the null hypothesis. Third, we were not able to assess medication management in our cohort and therefore cannot adjust for different medical management practices or draw conclusions about the effectiveness of medical therapy in the reduction of postoperative adverse cardiac events in this cohort. Fourth, we could not account for the exact number of coronary stents a patient had at the time of surgery, but we were able to adjust for the number of PCI episodes in the 2 years preceding the index stent placement. Finally, our cohort principally comprises older white men undergoing surgery in the VA system and therefore our findings may not be generalizable to other populations or health care settings.
Patients undergoing noncardiac surgery following coronary stent placement for MI are at increased risk for MACE and reinfarction following surgery compared with patients with stents placed for non-MI indications. Delaying surgery up to 6 months following PCI may be most important for patients with stents placed for MI irrespective of stent type.
Corresponding Author: Mary T. Hawn, MD, MPH, Department of Surgery, Stanford School of Medicine, 300 Pasteur Dr, M121 Alway, Stanford, CA 94305 (firstname.lastname@example.org).
Accepted for Publication: September 23, 2015.
Published Online: December 30, 2015. doi:10.1001/jamasurg.2015.4545.
Author Contributions: Dr Hollis and Ms Graham had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Holcomb, Valle, Hawn.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Holcomb.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Hollis, Graham, Richman.
Obtained funding: Maddox, Hawn.
Administrative, technical, or material support: Holcomb, Graham, Hawn.
Study supervision: Holcomb, Itani, Hawn.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by VA Health Services Research and Development Grant IIR 09-347 from the US Department of Veterans Affairs. Dr Holcomb was supported by grant T32 HS013852-11 from the Agency for Healthcare Research and Quality. Dr Hollis was supported by grant T32 CA091078-13 from the National Institutes of Health. Dr Richman was supported by a VA Health Services Research and Development Career Development Award from the US Department of Veterans Affairs. Dr Valle was supported by grant T32 HL0782 from the National Institutes of Health.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The opinions expressed are those of the authors and do not necessarily reflect those of the US Department of Veterans Affairs.
Previous Presentation: This paper was presented at the 39th Annual Meeting of the Association of VA Surgeons; May 4, 2015; Miami Beach, Florida.