A, Mean time between stent placement and surgery for all patients. B, Mean time between stent placement and surgery for patients with bare metal stents vs those with drug-eluting stents. C, Proportion of patients with recent myocardial infarction by stent type. D, Proportion of patients with recent congestive heart failure by stent type. Shaded areas represent the 95% confidence interval for the smoothed line.
Shaded areas represent the 95% confidence interval for the smoothed line.
Graham LA, Singletary BA, Richman JS, Maddox TM, Itani KMF, Hawn MT. Improved Adverse Postoperative Outcomes With Revised American College of Cardiology/American Heart Association Guidelines for Patients With Cardiac Stents. JAMA Surg. 2014;149(11):1113-1120. doi:10.1001/jamasurg.2014.2044
In 2007, the American College of Cardiology/American Heart Association guidelines were revised for patients with cardiac stents in need of subsequent surgery to recommend delaying elective noncardiac surgery by 365 days in patients with drug-eluting stents (DESs).
To examine the effect of the guidelines on postoperative major adverse cardiac events (MACEs) in subsequent noncardiac surgery.
Design, Setting, and Participants
Patients receiving a cardiac stent between fiscal years 2005 and 2010 were identified by International Classification of Diseases, Ninth Revision diagnosis codes in the Veterans Affairs Medical SAS data sets. The Veterans Affairs Surgical Quality Improvement Program data were used to identify subsequent operations in the 2 years following stenting. The preguideline period was defined as fiscal years 2005 through 2007 and the postguideline period was defined as fiscal years 2008 through 2010. Surgery patients admitted through the emergency department or transferred from another hospital were excluded from analyses. Overall, 16 634 elective noncardiac operations were identified (8034 [48.3%] in the preguideline period; 8600 [51.7%] in the postguideline period).
Main Outcomes and Measures
Composite 30-day postoperative MACEs. We used χ2 tests to examine differences in bivariate frequencies and used logistic models to examine adjusted associations with 2-year postoperative MACEs.
The median time to surgery was 364 days (interquartile range, 184-528 days). A total of 11 026 operations (66.3%) followed DES placement, and 5608 (33.7%) followed bare metal stent placement. After the guidelines’ publication, surgery timing increased following DES placement from 323 to 404 days (P < .001) and decreased following bare metal stent placement from 402 to 309 days (P < .001). In addition, postoperative MACE rates decreased from 4.2% to 3.3% (P = .002). After adjusting for cardiac risk factors and procedure characteristics, there was an overall absolute risk reduction of 0.9% for MACEs (odds ratio = 0.74; 95% CI, 0.62-0.89). On further examination of trends across time, MACE rates with DES placement began to decrease prior to the guidelines’ publication from 5.5% in 2005 to 4.3% in 2006 and remained stable through 2010. In contrast, MACE rates with bare metal stent placement increased from 4.3% in 2005 to 8.0% in 2007 but decreased to 4.8% following the guidelines’ publication.
Conclusions and Relevance
After the guidelines’ publication, noncardiac surgery was delayed in patients with DESs but not bare metal stents. With a 26% reduction in MACEs following the guidelines, it would appear that the guidelines did improve postoperative outcomes; however, when examined over time, it becomes evident that there are many more factors influencing management of patients with cardiac stents in need of subsequent surgery.
Approximately 1 in 5 patients with a cardiac stent will undergo noncardiac surgery in the 2 years following placement of their stent.1 The current guidelines recommend delaying elective noncardiac surgery by 365 days in patients with drug-eluting stents (DESs) or 42 days in patients with bare metal stents (BMSs) based on case series and expert opinion, but an increasing number of cohort studies have shown that the appropriate timing for surgery following DES placement could possibly be shortened.2,3 In addition to delaying surgery, the guidelines also recommend continuing the use of aspirin for urgent procedures within the following year; more recent research has suggested that antiplatelet therapy management may not be as highly associated with major adverse cardiac events (MACEs) as originally thought, but aspirin still remains associated with a risk of major bleeding following surgery.4- 6 These guidelines advise on the perioperative management of an increasing number of patients every year. In the Veterans Health Administration of Veterans Affairs (VA) in 2012, there were approximately 16 000 percutaneous coronary interventions resulting in a cardiac stent placement, with 3200 of these patients expected to undergo noncardiac surgery in the 12 months following their stent placement. Surgical decision making in this population must balance the increased risk of cardiac events against the bleeding risk attributable to antiplatelet therapy in the perioperative period.
The main objective of this study was to examine the association of the American College of Cardiology/American Heart Association guidelines with MACEs following noncardiac surgery within the VA. Prior survey research by our study team has shown that most VA clinicians were aware of and agreed with the current guidelines, but to our knowledge no studies have examined how the guidelines may have influenced practice habits at a patient level.7 With this in mind, we first set out to determine whether the time between stent placement and surgery changed following the guidelines’ publication on October 21, 2007, with the hypothesis that the time between DES placement and noncardiac surgery would increase in accordance with the recommendation to delay surgery in these patients. Second, we sought to examine the possible effect of the guidelines on adverse postoperative outcomes of myocardial infarction (MI), a revascularization procedure, and death in a VA population.
The study population and definition of study variables have been previously described.1,4 To construct the sample for this analysis, the study population from previous analyses was limited to operations occurring in the 3 years before and 3 years after the guidelines were published (October 1, 2004, to September 30, 2010). Surgery patients admitted through the emergency department or transferred from another hospital were excluded because the guidelines apply only to elective noncardiac surgery. This study was reviewed and approved by the Birmingham VA Institutional Review Board, VA Boston Healthcare System Institutional Review Board, and Colorado Multiple Institutional Review Board and was granted a waiver of informed consent.
The main outcome variable of interest was any MACE in the 30 days following surgery. Postoperative MACEs included a diagnosis of acute MI, a revascularization procedure, or death, all of which were assessed using International Classification of Diseases, Ninth Revision diagnosis codes in the VA Medical SAS data sets in combination with non-VA health care utilization as captured by the VA Information Resource Center’s Center for Medicaid and Medicare Services data for veterans. The Veterans Affairs Surgical Quality Improvement Program data were used to identify subsequent operations in the 2 years following stenting. The main independent variable was the timing of surgery in relation to the guidelines’ publication on October 21, 2007 (start of fiscal year 2008).2 This was examined both as a dichotomous variable—preguideline period (fiscal years 2005-2007) vs postguideline period (fiscal years 2008-2010)—and as a continuous variable to better visualize trends across time that may not necessarily be tied to the guidelines’ publication. As with postoperative outcomes, preoperative comorbidities were determined using the VA Medical SAS data sets in combination with non-VA health care utilization captured by the VA Information Resource Center’s Center for Medicaid and Medicare Services. Recent MI or congestive heart failure (CHF) was defined as an acute MI or a CHF exacerbation diagnosed in the 6 months preceding the surgical procedure. The work relative value units for each procedure were used as a measure of procedure complexity.
Univariate and bivariate frequencies were used to examine differences in population characteristics and MACE rates by timing of surgery in relation to the guidelines’ publication. We used χ2 tests and Wilcoxon rank sum tests to compare bivariate frequencies and continuous variables, respectively. Generalized additive models were used to account for the nonlinear association between time since stent placement and postoperative MACE. The final multivariate model covariates were determined using backward stepwise selection including covariates found to be associated in bivariate analyses. Predicted rates and adjusted odds ratios were calculated from the most parsimonious multivariate model. Finally, LOESS smoothing was used to generate smoothed plots of time between stent placement and surgery as well as MACE rates across time to examine guideline timing as a continuous variable. All analyses were completed using SAS version 9.2 statistical software (SAS Institute, Inc). Smoothed plots were generated using the GGPLOT2 package and R version 3.0.1 statistical software (R Foundation).8- 10
We identified 16 634 noncardiac operations occurring in VA medical centers between October 1, 2004, and September 30, 2010 (8034 [48.3%] in the preguideline period; 8600 [51.7%] in the postguideline period). A total of 11 026 (66.3%) of the operations followed DES placement and 5608 (33.7%) followed BMS placement. Most patients were older white men with a higher prevalence of cardiac risk factors (Table 1). Patients undergoing surgery in the preguideline period were younger (P < .001) and less likely to have diabetes mellitus (P = .004). Preguideline operations had slightly lower relative value units compared with those in the postguideline period (median, 9.6 vs 10.2 relative value units, respectively; P = .01) (Table 2). The proportion of patients with a DES requiring surgery varied across fiscal years (48.0% in 2005, a peak at 77.4% in 2007, and 68.5% in 2010; P < .001), but overall the distribution of operations by stent type was similar in the preguideline and postguideline groups (P = .06). During the entire study period, the median time between stent placement and surgery was 364 days (interquartile range, 184-528 days) with 4256 patients (53.0%) undergoing surgery during the first year following their cardiac stent placement in the preguideline period compared with 4121 patients (47.9%) in the postguideline period. Likewise, the median time to surgery between the 2 periods differed significantly with a shorter time to surgery in the preguideline period (349 days) than in the postguideline period (378 days) (P < .001) (Table 2).
Changes across time in the median days to surgery and cardiac risk factors by stent type are presented in Figure 1. For patients with a DES, time to surgery showed a consistent increase from 2005 to early 2008, after which it decreased and stabilized at just over 365 days. The time between BMS placement and surgery decreased in 2006 prior to the guidelines’ release directly in contrast to the increasing time to surgery for patients with a DES (Figure 1B). Concurrent with the decreasing time to surgery, we observed an increase in major cardiac risk factors of recent MI or CHF exacerbation at the time of surgery in patients with a BMS (Figure 1C and D).
In total, 617 patients (3.7%) experienced a postoperative MACE within 30 days of noncardiac surgery (Table 3). Most MACEs were cardiac (n = 527 [85.4%]), with 107 deaths reported in the first 30 postoperative days. Of the 107 deaths, 15.9% occurred following a cardiac event (13.4% with DES and 20.0% with BMS; P = .37). Postoperative MACEs differed in relation to the guidelines, with 4.2% prior to the guidelines’ publication compared with 3.3% after the guidelines were released (P = .002). Most MACEs were postoperative MI (n = 450), and 107 postoperative MACEs (17.3%) resulted in death. When stratified by stent type, MACE rates were slightly higher among patients with a BMS compared with patients with a DES, but this did not reach statistical significance (4.1% vs 3.5%, respectively; P = .10) and differed significantly by whether the surgery occurred in the first or second year following cardiac stent placement (4.8% vs 2.6%, respectively; P < .001).
The final adjusted model included operation date, stent type, time between stent placement and surgery, procedure type, work relative value units, admission status, recent MI, recent CHF, and age at time of surgery. In the adjusted model, we observed an overall absolute risk reduction (ARR) of 0.9% following publication of the guidelines (odds ratio = 0.74; 95% CI, 0.62-0.89) (Table 4). The reduction in MACEs varied significantly by stent type and time between stent placement and surgery. Patients with a BMS undergoing surgery in the first year following stent placement not only had the highest preguideline MACE rates after adjustment but also had one of the largest reductions (5.9% in the preguideline period vs 4.8% in the postguideline period) (Table 3). Even after adjustment, the greatest ARR was seen in patients with a BMS undergoing surgery in the first year (ARR = 1.1%) or second year (ARR = 1.2%) following stent placement. On further examination of predicted rates by fiscal year, MACE rates with DES placement began to decrease prior to the guidelines’ publication from 5.5% in 2005 to 4.3% in 2006 and remained stable through 2010. In contrast, MACE rates with BMS placement increased from 4.3% in 2005 to 8.0% in 2007 but decreased to 4.8% following the guidelines’ publication (Figure 2).
In secondary analyses, we repeated the predicted MACE rates using a cutoff of 6 months for our surgery timing interval from stent placement. Postoperative MACE rates following operations within 6 months in patients with a DES decreased from 6.1% in the preguideline period to 5.2% in the postguideline period, while the decrease was much smaller for surgical procedures in the 6 to 12 months after stent placement (3.2% in the preguideline period to 3.1% in the postguideline period). For patients with a BMS, the change in postoperative MACE rates from the preguideline period to the postguideline period was more consistent across all surgery timing intervals. In addition, most of the increase in postoperative MACE rates between fiscal years 2005 and 2007 was seen in the surgical procedures occurring within the 6 months following a BMS placement, which also carry the highest risk of postoperative MACE (fiscal year 2005, 3.3%; fiscal year 2006, 8.4%; fiscal year 2007, 9.3%).
This study provides evidence of increases in timing between stenting and surgery in association with the adoption of the American College of Cardiology/American Heart Association guidelines. At the same time, there was a slight reduction in adverse postoperative events in patients with a cardiac stent. From the beginning of the study, prior to the guidelines’ publication through early 2008, the average time between stent placement and surgery for patients with a DES increased in a manner consistent with the guidelines’ recommendation to delay elective surgery. This phenomenon is in line with previous research showing high awareness and agreement with the guidelines among VA surgeons, anesthesiologists, and cardiologists.11 In addition, prior research has shown a decrease in the incidence of noncardiac surgery within 12 months of cardiac stent placement following the guidelines’ release.1 This decrease was more evident in patients with a DES, with 14.9% of patients with a DES undergoing surgery within 12 months in the preguideline period compared with only 8.2% after October 2007.1
These trends in timing of surgery following DES placement suggest a high level of awareness among health care professionals treating patients with cardiac stents and an increased use of evidence-based practices as reports of selecting stents based on projected need of urgent surgery began to appear in peer-reviewed literature as early as 2004.12 Early case reports and case series suggesting an increased risk of stent thrombosis following noncardiac surgery in patients with DES began to appear in the literature in late 2004 through early 2007, but even in 2007 the evidence was limited to case reports and single-institution studies.13- 18 While our study shows a shift in cardiac comorbidities and time between stent placement and surgery, other studies have also shown that while the number of stents placed across time has stabilized, the proportion of patients receiving a DES decreased drastically between 2005 and 2006 when the case reports and case series of adverse perioperative events were beginning to be published.19- 21 While the publication of the guidelines marks a summary of this information into 1 document, it does not necessarily mark the time at which the information was officially available to health care professionals in the field. Our results, along with previous studies, suggest that VA health care professionals were already aware of what the published literature was showing and began to change practices according to their interpretation of the literature.
Finally, while our research shows a reduced risk of MACEs in the postguideline period, there is evidence of a shift in the patient population that may affect MACE rates and provides strong evidence of confounding by indication. Overall, patients in the postguideline period were older and more likely to have diabetes. When the population was examined by stent type, we found an increase in the proportion of patients with a recent MI or CHF episode in the 6 months leading up to their surgery for the years surrounding the guidelines’ publication. This is seen concurrent with the shift toward shorter time between stent placement and surgery in patients with a BMS. The risk of MACEs appears to have been shifted to patients with a BMS, who saw a parallel increase in postoperative MACEs in the 3 years preceding the guidelines’ publication (fiscal year 2005, 4.8%; fiscal year 2006, 6.0%; fiscal year 2007, 7.3%). Previous literature has suggested that differences in outcomes of BMS and DES are due to confounding by indication, where sicker patients needing more urgent operations were more likely to receive a BMS and thus were more likely to experience adverse postoperative events. However, both propensity-matching and instrumental variable analyses have failed to provide a clear explanation of confounding by stent type selection.21- 23 While our study cannot adequately account for this confounding, it does provide further evidence of a shift toward higher risk in patients with a BMS.
As with all studies, ours has some limitations. First, this a study of patients undergoing surgery within the VA health care system, predominantly including older men, and thus may not be generalizable to the entire US population. Second, we are unable to take into account confounding by indication, which may have occurred following the publication of new research and guidelines. It is likely that when faced with the decision of which stent to place, cardiologists were more likely to place BMSs in patients with an anticipated need for earlier surgery. Unfortunately, this choice is not measurable with the data we have. Third, the reduced MACE rate for both DES and BMS suggests changes in management of patients with a cardiac stent that occurred outside the guidelines, which our study was not designed to assess. In addition, the guidelines also recommended antiplatelet therapy management strategies to reduce the risk of adverse postoperative events, and our study was not designed to assess this. However, research shows that antiplatelet therapy cessation may not be as strongly associated with adverse cardiac events4- 6 and that aspirin continuation results in a much higher risk of bleeding than previously thought.6 Finally, all-cause mortality was included in the definition of MACE, which may bias the patients with BMSs in the event of confounding by indication. However, we found similar trends for both MI and all-cause death in the population with BMSs, suggesting that most of the deaths had a cardiac cause. In addition, a similar percentage of deaths occurred concurrently with MI or revascularizations by stent type (13.4% with DES vs 20.0% with BMS).
The recommendations of the guidelines were effective at reducing adverse postoperative outcomes in patients with a cardiac stent. In addition to a reduction in adverse postoperative events, we found an increase in time between DES placement and surgery; however, this increase began before the guidelines were published, which suggests influences of other factors such as the initial case reports of stent thrombosis in the early years following introduction of the DES. Finally, there was a shift toward earlier operations with more recent MI and CHF episodes in the population with BMSs, providing further evidence of selection bias that prior studies have struggled to thoroughly describe.22 This shift parallels higher postoperative MACE rates among patients with BMSs; however, in the years following the guidelines, patients with DESs saw a continued decline in postoperative MACE rates, while MACE rates in patients with BMSs appeared to return to baseline. When examined as a dichotomous variable, it would appear that the guidelines were successful at changing practices in surgery timing as well as reducing MACE rates. When examined over time, though, it becomes evident that many more factors influence the management of patients with cardiac stents in need of subsequent surgery. Further research should examine factors influencing postoperative MACEs in patients with BMSs.
Corresponding Author: Mary T. Hawn, MD, MPH, Section of Gastrointestinal Surgery, Department of Surgery, University of Alabama at Birmingham, 1922 Seventh Ave S, KB 428, Birmingham, AL 35294 (email@example.com).
Accepted for Publication: May 5, 2014.
Published Online: September 3, 2014. doi:10.1001/jamasurg.2014.2044.
Author Contributions: Dr Hawn had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Graham, Itani, Hawn.
Acquisition, analysis, or interpretation of data: Graham, Singletary, Richman, Maddox, Hawn.
Drafting of the manuscript: Graham.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Graham, Singletary, Richman.
Obtained funding: Hawn.
Administrative, technical, or material support: Graham, Singletary.
Study supervision: Graham, Maddox, Itani, Hawn.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study is funded by grant IIR 09-347 from VA Health Services Research and Development. Drs Richman and Maddox are supported by career development awards from VA Health Services Research and Development.
Role of the Funder/Sponsor: The funder 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 not necessarily those of the US Department of Veterans Affairs or the US government.
Previous Presentation: This work was presented at the 2014 Annual Meeting of the Association of VA Surgeons; April 6, 2014; New Haven, Connecticut.