Risk of Noncardiac Surgery and Adverse Events After Stents
Of 41,989 surgeries, we matched 49% to two patients in the nonsurgical cohort on the basis of stent type and cardiac risk factors. Differences between the matched and nonmatched surgeries are presented in Online Table 1 http://content.onlinejacc.org/data/Journals/JAC/931875/09072_mmc1.docx?v=635544260354200000. On the basis of available VA Clinical Assessment, Reporting, and Tracking Program information, the majority of stents in both cohorts during the study period were first-generation DES (58.0%), with 58.4% sirolimus-eluting and 42.9% paclitaxel-eluting in the surgical cohort versus 55.9% and 44.8% in the nonsurgical cohort, respectively (Online Table 2 http://content.onlinejacc.org/data/Journals/JAC/931875/09072_mmc1.docx?v=635544260354200000). Demographics and comorbidities of the surgical and nonsurgical cohort after matching are presented in Table 1. Approximately two-thirds of stents in both cohorts were placed for acute coronary syndromes. The surgical cohort had higher rates of elective stent placement compared with the nonsurgical cohort (33.9% vs. 31.4%). The matched surgical procedures included a broad representation of surgical procedure types and most surgeries were performed at a VA hospital (Table 2). To examine secular trends and evolving stent technology over the study period, we determined the rates of the composite cardiac endpoint by the fiscal year of stent placement (Online Figure 1 http://content.onlinejacc.org/data/Journals/JAC/931875/09072_mmc1.docx?v=635544260354200000). Postoperative adverse cardiac event rates remained relatively stable over time for both surgical and nonsurgical cohorts and by stent type. The cumulative incidence of adverse cardiac events over the 24 months following coronary stent placement was similar in both the surgical and nonsurgical cohorts when stratified by stent type indicating that both cohorts were evenly matched on risk for adverse cardiac events (Figure 1). When stratified by stent type, the surgical cohort had lower adverse cardiac event rates compared to the nonsurgical cohort following BMS placement (18.6% vs. 19.0%) but higher rates were observed with DES (16.2% vs. 15.1%) over the 24-month time period.
(Enlarge Image)
Figure 1.
Cumulative Incidence of Adverse Cardiac Events Following Coronary Stent Placement
The cumulative incidence of adverse cardiac events (includes myocardial infarction and/or revascularization) over 24 months from time of coronary stent placement. Results are stratified by surgical versus nonsurgical cohort and by stent type. BMS = bare-metal stent(s); DES = drug-eluting stent(s).
To better refine the incremental risk associated with surgery, we restricted the comparison to adverse cardiac events occurring in the 30-day interval following surgery for the surgical cohort and the same post-stent period for the matched nonsurgical cohort. We observed higher rates of the composite cardiac endpoint (3.1% vs. 1.9%), MI (2.5% vs. 1.1%), and all-cause mortality (1.4% vs. 0.4%) (all p <0.001) in the surgical population compared with those who did not undergo surgery. There was no difference in revascularization rates (1.1% vs. 1.0%; p = 0.37). The RD for the surgical cohort during the 30-day period following surgery for MI was 1.4% (95% CI: 1.2% to 1.7%) and 1.0% (95% CI: 0.9% to 1.2%) for all-cause mortality compared to the nonsurgical cohort (Table 3).
The 30-day adverse cardiac event rates for both the surgical cohort and nonsurgical cohorts decrease over time as time from stent placement increases (Figure 2A). Rates for both cohorts were highest immediately post-stent and decreased rapidly in the first 6 months regardless of stent type. The risk difference, adjusted for surgical characteristics associated with adverse cardiac events, is plotted in Figure 2B. The adjusted risk difference is highest in the first 6 weeks following stent implantation, decreasing sharply in the first 6 months, and leveling off to 1% for the remainder of the 24 months. This pattern is similar for surgery following DES or BMS implantation.
(Enlarge Image)
Figure 2.
Outcomes and Risk Differences in Surgical Versus Nonsurgical Cohorts
Smoothed 30-day plot of (A) and risk differences (adjusted for procedure type, admission status, and work [relative value unit as a measure of surgical complexity]) in (B) rates of the composite cardiac endpoint (myocardial infarction and/or revascularization) over the 24 months from the time of coronary stent placement. Gray areas represent 95% confidence intervals for the adjusted risk differences. Abbreviations as in Figure 1.
To quantify the incremental risk of surgery over time following stent placement, we stratified 30-day adverse cardiac events post-procedure into three time intervals of <6 weeks, 6 weeks to 6 months, and >6 months from time of stent placement ( Table 4 ). When surgeries were performed in the first 6 weeks following stent placement, the overall rates and associated RDs of the composite cardiac endpoint (9.0%; RD: 2.8%; 95% CI: 0.8% to 4.8%), MI (7.5%; RD: 3.6%; 95% CI: 1.8% to 5.4%), and death (3.2%; RD: 2.5%; 95% CI: 1.4% to 3.6%) were markedly higher compared to the later time intervals. Across the three time intervals, a decreasing trend in the rates and associated RDs was observed for all of the adverse event endpoints, excepting revascularization.
Given that the risk for adverse cardiac events associated with surgery was stabilized at 6 months from stent placement, we sought to determine which procedure types would most benefit from delaying surgery until 6 months after stent deployment. Accordingly, we compared the RD between two different stent to surgery time periods: 6 weeks to 6 months and 6 months to 24 months following stent placement. Nonelective surgeries (RD6wk–6mo: 12.3%; RD6mo–24mo: 7.6%), vascular (RD6wk–6mo: 3.8%; RD6mo-24mo: 2.7%), and respiratory (RD6wk–6mo: 3.0%; RD6mo–24mo: 3.3%) procedures confer the highest RD for adverse cardiac events, but risk was not significantly decreased between the two time periods (Table 5). Compared to earlier surgery, digestive procedures (RD6wk–6mo: 4.6%; RD6mo–24mo: 0.9% [p = 0.003]), high-risk surgeries (RD6wk–6mo: 6.5%; RD6mo–24mo: 2.1% [p = 0.01]), and elective inpatient cases (RD6wk–6mo: 3.7%; RD6mo–24mo: 1.9% [p = 0.01]) are associated with a significant decrease in the RD when performed more than 6 months following stent. Surgery performed on patients with a DES more than 6 months from time of placement was associated with a significant decrease in risk difference (RD6wk–6mo: 2.1%; RD6mo–24mo: 0.9% [p = 0.01]), whereas the risk difference for patients with a BMS did not significantly change (RD6wk–6mo: 1.9%; RD6mo–24mo: 1.0% [p = 0.13]). This difference by stent type was still observed after adjusting for procedure characteristics during each time period (DES: p = 0.03; BMS: p = 0.14).
Results
Matched Variables and Characteristics of the Surgical and Nonsurgical Cohorts
Of 41,989 surgeries, we matched 49% to two patients in the nonsurgical cohort on the basis of stent type and cardiac risk factors. Differences between the matched and nonmatched surgeries are presented in Online Table 1 http://content.onlinejacc.org/data/Journals/JAC/931875/09072_mmc1.docx?v=635544260354200000. On the basis of available VA Clinical Assessment, Reporting, and Tracking Program information, the majority of stents in both cohorts during the study period were first-generation DES (58.0%), with 58.4% sirolimus-eluting and 42.9% paclitaxel-eluting in the surgical cohort versus 55.9% and 44.8% in the nonsurgical cohort, respectively (Online Table 2 http://content.onlinejacc.org/data/Journals/JAC/931875/09072_mmc1.docx?v=635544260354200000). Demographics and comorbidities of the surgical and nonsurgical cohort after matching are presented in Table 1. Approximately two-thirds of stents in both cohorts were placed for acute coronary syndromes. The surgical cohort had higher rates of elective stent placement compared with the nonsurgical cohort (33.9% vs. 31.4%). The matched surgical procedures included a broad representation of surgical procedure types and most surgeries were performed at a VA hospital (Table 2). To examine secular trends and evolving stent technology over the study period, we determined the rates of the composite cardiac endpoint by the fiscal year of stent placement (Online Figure 1 http://content.onlinejacc.org/data/Journals/JAC/931875/09072_mmc1.docx?v=635544260354200000). Postoperative adverse cardiac event rates remained relatively stable over time for both surgical and nonsurgical cohorts and by stent type. The cumulative incidence of adverse cardiac events over the 24 months following coronary stent placement was similar in both the surgical and nonsurgical cohorts when stratified by stent type indicating that both cohorts were evenly matched on risk for adverse cardiac events (Figure 1). When stratified by stent type, the surgical cohort had lower adverse cardiac event rates compared to the nonsurgical cohort following BMS placement (18.6% vs. 19.0%) but higher rates were observed with DES (16.2% vs. 15.1%) over the 24-month time period.
(Enlarge Image)
Figure 1.
Cumulative Incidence of Adverse Cardiac Events Following Coronary Stent Placement
The cumulative incidence of adverse cardiac events (includes myocardial infarction and/or revascularization) over 24 months from time of coronary stent placement. Results are stratified by surgical versus nonsurgical cohort and by stent type. BMS = bare-metal stent(s); DES = drug-eluting stent(s).
Thirty-day Post-surgical Outcomes of the Surgical Cohort Compared to Nonsurgical Cohort
To better refine the incremental risk associated with surgery, we restricted the comparison to adverse cardiac events occurring in the 30-day interval following surgery for the surgical cohort and the same post-stent period for the matched nonsurgical cohort. We observed higher rates of the composite cardiac endpoint (3.1% vs. 1.9%), MI (2.5% vs. 1.1%), and all-cause mortality (1.4% vs. 0.4%) (all p <0.001) in the surgical population compared with those who did not undergo surgery. There was no difference in revascularization rates (1.1% vs. 1.0%; p = 0.37). The RD for the surgical cohort during the 30-day period following surgery for MI was 1.4% (95% CI: 1.2% to 1.7%) and 1.0% (95% CI: 0.9% to 1.2%) for all-cause mortality compared to the nonsurgical cohort (Table 3).
Incremental Risk of Noncardiac Surgery for Adverse Cardiac Events Across Time From Coronary Stent
The 30-day adverse cardiac event rates for both the surgical cohort and nonsurgical cohorts decrease over time as time from stent placement increases (Figure 2A). Rates for both cohorts were highest immediately post-stent and decreased rapidly in the first 6 months regardless of stent type. The risk difference, adjusted for surgical characteristics associated with adverse cardiac events, is plotted in Figure 2B. The adjusted risk difference is highest in the first 6 weeks following stent implantation, decreasing sharply in the first 6 months, and leveling off to 1% for the remainder of the 24 months. This pattern is similar for surgery following DES or BMS implantation.
(Enlarge Image)
Figure 2.
Outcomes and Risk Differences in Surgical Versus Nonsurgical Cohorts
Smoothed 30-day plot of (A) and risk differences (adjusted for procedure type, admission status, and work [relative value unit as a measure of surgical complexity]) in (B) rates of the composite cardiac endpoint (myocardial infarction and/or revascularization) over the 24 months from the time of coronary stent placement. Gray areas represent 95% confidence intervals for the adjusted risk differences. Abbreviations as in Figure 1.
Incremental Risk of Noncardiac Surgery Stratified by <6 Weeks, 6 Weeks to 6 Months, and >6 Months
To quantify the incremental risk of surgery over time following stent placement, we stratified 30-day adverse cardiac events post-procedure into three time intervals of <6 weeks, 6 weeks to 6 months, and >6 months from time of stent placement ( Table 4 ). When surgeries were performed in the first 6 weeks following stent placement, the overall rates and associated RDs of the composite cardiac endpoint (9.0%; RD: 2.8%; 95% CI: 0.8% to 4.8%), MI (7.5%; RD: 3.6%; 95% CI: 1.8% to 5.4%), and death (3.2%; RD: 2.5%; 95% CI: 1.4% to 3.6%) were markedly higher compared to the later time intervals. Across the three time intervals, a decreasing trend in the rates and associated RDs was observed for all of the adverse event endpoints, excepting revascularization.
Evaluation of Procedure Characteristics Associated With Reduced Incremental Risk for Surgery >6 Months Post-stenting
Given that the risk for adverse cardiac events associated with surgery was stabilized at 6 months from stent placement, we sought to determine which procedure types would most benefit from delaying surgery until 6 months after stent deployment. Accordingly, we compared the RD between two different stent to surgery time periods: 6 weeks to 6 months and 6 months to 24 months following stent placement. Nonelective surgeries (RD6wk–6mo: 12.3%; RD6mo–24mo: 7.6%), vascular (RD6wk–6mo: 3.8%; RD6mo-24mo: 2.7%), and respiratory (RD6wk–6mo: 3.0%; RD6mo–24mo: 3.3%) procedures confer the highest RD for adverse cardiac events, but risk was not significantly decreased between the two time periods (Table 5). Compared to earlier surgery, digestive procedures (RD6wk–6mo: 4.6%; RD6mo–24mo: 0.9% [p = 0.003]), high-risk surgeries (RD6wk–6mo: 6.5%; RD6mo–24mo: 2.1% [p = 0.01]), and elective inpatient cases (RD6wk–6mo: 3.7%; RD6mo–24mo: 1.9% [p = 0.01]) are associated with a significant decrease in the RD when performed more than 6 months following stent. Surgery performed on patients with a DES more than 6 months from time of placement was associated with a significant decrease in risk difference (RD6wk–6mo: 2.1%; RD6mo–24mo: 0.9% [p = 0.01]), whereas the risk difference for patients with a BMS did not significantly change (RD6wk–6mo: 1.9%; RD6mo–24mo: 1.0% [p = 0.13]). This difference by stent type was still observed after adjusting for procedure characteristics during each time period (DES: p = 0.03; BMS: p = 0.14).