MACCE indicates major adverse cardiovascular and cerebrovascular events.
eFigure 1. Patient selection flow diagram
eFigure 2. Frequency of perioperative MACCE by RCRI score
eFigure 3. Risks of major adverse cardiovascular and cerebrovascular events over time, adjusted for demographics, clinical covariates, and surgical subtypes
eFigure 4. Sensitivity Analyses of trends in MACCE excluding urgent surgeries (Panel A), excluding surgeries performed >72 hours from hospital admission (Panel B), and excluding vascular surgery (Panel C)
eFigure 5. Trends in perioperative major adverse cardiovascular and cerebrovascular events (Panel A), perioperative mortality (Panel B), perioperative myocardial infarction (Panel C), and perioperative ischemic stroke (Panel D) by sex
eFigure 6. Trends in perioperative major adverse cardiovascular and cerebrovascular events (Panel A), perioperative mortality (Panel B), perioperative myocardial infarction (Panel C), and perioperative ischemic stroke (Panel D) by race/ethnicity
Customize your JAMA Network experience by selecting one or more topics from the list below.
Smilowitz NR, Gupta N, Ramakrishna H, Guo Y, Berger JS, Bangalore S. Perioperative Major Adverse Cardiovascular and Cerebrovascular Events Associated With Noncardiac Surgery. JAMA Cardiol. 2017;2(2):181–187. doi:10.1001/jamacardio.2016.4792
What are the trends in perioperative major adverse cardiovascular events (MACCE) after noncardiac surgery in the United States?
In this analysis of patients undergoing in-hospital major noncardiac surgery from 2004 to 2013, the frequency of MACCE declined from 3.1% to 2.6% driven by a decline in frequency of perioperative death and acute myocardial infarction but an increase in perioperative ischemic stroke from 0.52% in 2004 to 0.77% in 2013.
Perioperative MACCE occurs in 1 of every 33 hospitalizations for noncardiac surgery; despite reductions in the rate of death and acute myocardial infarction among patients undergoing major noncardiac surgery in the United States, perioperative ischemic stroke increased over time.
Major adverse cardiovascular and cerebrovascular events (MACCE) are a significant source of perioperative morbidity and mortality following noncardiac surgery.
To evaluate national trends in perioperative cardiovascular outcomes and mortality after major noncardiac surgery and to identify surgical subtypes associated with cardiovascular events using a large administrative database of United States hospital admissions.
Design, Setting, Participants
Patients who underwent major noncardiac surgery from January 2004 to December 2013 were identified using the National Inpatient Sample.
Main Outcomes and Measures
Perioperative MACCE (primary outcome), defined as in-hospital, all-cause death, acute myocardial infarction (AMI), or acute ischemic stroke, were evaluated over time.
Among 10 581 621 hospitalizations (mean [SD] patient age, 65.74 [12.32] years; 5 975 798 female patients 56.60%]) for major noncardiac surgery, perioperative MACCE occurred in 317 479 hospitalizations (3.0%), corresponding to an annual incidence of approximately 150 000 events after applying sample weights. Major adverse cardiovascular and cerebrovascular events occurred most frequently in patients undergoing vascular (7.7%), thoracic (6.5%), and transplant surgery (6.3%). Between 2004 and 2013, the frequency of MACCE declined from 3.1% to 2.6% (P for trend <.001; adjusted odds ratio [aOR], 0.95; 95% CI, 0.94-0.97) driven by a decline in frequency of perioperative death (aOR, 0.79; 95% CI, 0.77-0.81) and AMI (aOR, 0.87; 95% CI, 0.84-0.89) but an increase in perioperative ischemic stroke from 0.52% in 2004 to 0.77% in 2013 (P for trend <.001; aOR 1.79; CI 1.73-1.86).
Conclusions and Relevance
Perioperative MACCE occurs in 1 of every 33 hospitalizations for noncardiac surgery. Despite reductions in the rate of death and AMI among patients undergoing major noncardiac surgery in the United States, perioperative ischemic stroke increased over time. Additional efforts are necessary to improve cardiovascular care in the perioperative period of patients undergoing noncardiac surgery.
Worldwide, more than 300 million noncardiac surgeries are performed each year.1 Major adverse cardiovascular and cerebrovascular events (MACCE), including myocardial infarction and ischemic stroke, are a significant source of perioperative morbidity and mortality.2,3 Cardiovascular complications prolong inpatient hospitalization, increase medical costs, and are the leading cause of perioperative death.4,5 Over the past decades, there have been improvements in perioperative risk stratification, advances in surgical and anesthetic technique, and ongoing efforts to improve perioperative cardiovascular care with large therapeutic trials in the perioperative period.6-10 At the same time, the rising burden of cardiovascular risk factors in the population undergoing noncardiac surgery may attenuate improvements in perioperative outcomes over time.11
Despite the significant burden perioperative events place on the national health care system, trends in perioperative MACCE among patients hospitalized for major noncardiac surgery have not been reported in the contemporary era.5 We sought to evaluate national trends in perioperative cardiovascular outcomes and mortality after noncardiac surgery and to identify surgical predictors of in-hospital perioperative cardiovascular events using a large administrative database of hospital admissions from the United States.
Patients age 45 years or older requiring major noncardiac surgery between January 2004 and December 2013 were included in this analysis. Patients were identified using the Healthcare Cost and Utilization Project’s (HCUP) National Inpatient Sample (NIS), a large administrative database of discharge-level data from a 20% stratified sample of all hospitals in the United States with deidentified data from approximately 8 million hospitalizations per year.12 Patients were included if they had a principal Clinical Classifications Software (CCS) procedure code representing a major therapeutic operating room procedure (HCUP Procedure Class 4). Principal CCS procedure codes represent an aggregate of relevant primary International Classification of Diseases, Ninth Revision (ICD-9) procedure codes by surgical subtype. Patients who underwent cardiac procedures (n = 1 655 567), cardiac surgery and cardiac transplantation (n = 582 726), bone marrow transplantation (n = 18 151), ophthalmologic surgery (n = 13 342), radiation therapy (n = 9817), dental surgery (n = 1779), and nonoperating room procedures (n = 386) as a principal in-hospital procedure were excluded. Major noncardiac surgery CCS procedure codes were clustered into 13 major surgical subtypes: breast, endocrine, otolaryngology, general, genitourinary, gynecologic, neurosurgery, obstetrics, orthopedic, skin and burn, thoracic, noncardiac transplant, and vascular surgery.
The primary outcome was major adverse cardiovascular and cerebrovascular events (MACCE), defined as in-hospital all-cause death, acute myocardial infarction (AMI), or ischemic stroke; AMI was identified using CCS diagnosis code 100. Acute ischemic stroke was identified using ICD-9 diagnosis codes 433.x1, 434.x1, 436, and 437.1.13 Other outcomes evaluated were complete heart block using ICD-9 diagnosis code 426.0, cardiogenic shock using ICD-9 diagnosis code 785.51, and cardiac arrest using ICD-9 diagnosis code 427.5.
Continuous variables were reported as mean (SD) and compared using the t test and analysis of variance tests for multiple comparisons. Categorical variables were reported as percentages and compared by χ2 tests. Analyses of proportions over time were performed using the Cochran-Armitage test for trend. Multivariable logistic regression models were generated to estimate odds of perioperative cardiovascular events, adjusted for patient demographics, cardiovascular risk factors, and relevant comorbidities. Models included age, sex, race/ethnicity, obesity, tobacco use, hypertension, hyperlipidemia, diabetes mellitus, chronic kidney disease, end stage renal disease, coronary artery disease, prior revascularization with either percutaneous coronary intervention or coronary artery bypass surgery, peripheral artery disease, valvular heart disease, congestive heart failure, prior venous thromboembolism, chronic lung disease, alcohol abuse, malignancy, anemia, elective or urgent hospitalization, surgery type and year of hospitalization as covariates. To facilitate data presentation, patient characteristics and adverse-event rates were reported in 2-year intervals: 2004 to 2005, 2006 to 2007, 2008 to 2009, 2010 to 2011, and 2012 to 2013. Sampling weights were applied to calculate rates for trend analyses and to determine national incidence estimates.14 Unweighted data were used in all other analyses, unless otherwise specified. Statistical analyses were performed using SPSS 20 (IBM SPSS Statistics). Statistical tests are 2-sided and P values less than .05 were considered statistically significant.
Several sensitivity analyses were performed to validate the study findings. To confirm established associations between validated risk factors and perioperative MACCE in this national data set, the incidence of perioperative MACCE was determined for subgroups by modified Revised Cardiac Risk Index (RCRI) score using ICD-9 codes for ischemic heart disease, heart failure, prior transient ischemic attack or stroke, chronic kidney disease, diabetes, and high-risk surgery.7 Due to the introduction of an ICD-9 diagnosis code for prior stroke or transient ischemic attack in mid-2007, modified RCRI scores were calculated from 2008 to 2013 data. To exclude cases in which MACCE may have preceded the primary noncardiac surgery, a sensitivity analysis was performed excluding patients who were hospitalized urgently or emergently. Similarly, a sensitivity analysis was performed in the cohort of patients who underwent noncardiac surgery within the first 72 hours of hospital admission. Finally, due to the established risks of perioperative stroke in patients undergoing major vascular surgery,15 a sensitivity analysis was performed excluding this high-risk cohort.
From January 2004 to December 2013, 12 863 389 hospitalizations for major surgery were identified among patients 45 years and older. After excluding patients undergoing cardiac surgery, low-risk and nonoperative procedures, the final study sample consisted of 10 581 621 hospitalizations for major noncardiac surgery (eFigure 1 in the Supplement). This corresponds to an estimated 50 558 529 surgical hospitalizations in the United States during this time period, after applying sampling weights.
Major adverse cardiovascular and cerebrovascular events occurred in 317 479 major noncardiac surgeries (3000 events per 100 000 [3.0%]), corresponding to an estimated 1 510 694 perioperative events in the United States during this time period, after applying sampling weights. Nonfatal AMI occurred in the perioperative period of 80 076 surgeries (757 events per 100 000 [0.76%]), nonfatal stroke occurred in 57 350 (542 events per 100 000 [0.54%]), both nonfatal AMI and nonfatal stroke occurred in 3094 (29 events per 100 000 [0.03%]), and death occurred in 176 959 (1672 events per 100 000 [1.67%]) patients. Demographics and baseline characteristics of patients with and without perioperative MACCE are shown in Table 1. Patients who had a MACCE were older, more likely to be male, and more likely to have cardiovascular risk factors when compared with those who did not experience a MACCE (Table 1). As expected, patients with the highest RCRI scores were most likely to have perioperative MACCE compared with patients with the lowest RCRI scores (10 829 vs 1261 events per 100 000 surgeries; P < .001) (eFigure 2 in the Supplement).
Among all patients undergoing noncardiac surgery, complete heart block was identified in 8264 cases (78 events per 100 000 surgeries [0.08%]), cardiogenic shock occurred in 10 789 cases (102 events per 100 000 surgeries [0.10%]) and cardiac arrest occurred in 34 345 cases (325 events per 100 000 surgeries [0.32%]).
Between January 2004 and December 2013, the number of perioperative MACCE per 100 000 surgeries declined by 580 (95% CI, 559-602), from 3181 to 2601 (P for trend <.001; adjusted odds ratio [aOR], 0.95; 95% CI, 0.94-0.97) (Table 2) (Figure 1). The rate of perioperative mortality per 100 000 surgeries declined by 695 (95% CI, 679-711), from 1978 to 1283 (P for trend <.001; aOR, 0.79; 95% CI 0.77-0.81). The rate of perioperative AMI per 100 000 surgeries decreased by 205 (95% CI, 193-217), from 1004 to 799 (P for trend <.001; aOR, 0.87; 95% CI, 0.84-0.89). Conversely, perioperative ischemic stroke per 100 000 surgeries increased over time by 244 (95% CI, 234-254), from 524 to 768 (P for trend <.001; aOR, 1.79; 95% CI, 1.73-1.86). The odds of MACCE over time, after multivariable adjustment for demographics, clinical covariates, and surgical type are shown in eFigure 3 in the Supplement.
Similar trends in perioperative MACCE, death, AMI, and stroke were observed in a sensitivity analysis of patients who were electively hospitalized for noncardiac surgery, as well as in a sensitivity analysis of patients who underwent the principal noncardiac surgery within the first 72 hours of hospital admission (eFigure 4A and B in the Supplement). After excluding patients who underwent major vascular surgery, similar trends in perioperative MACCE and the individual endpoints were also observed (eFigure 4C in the Supplement).
Patients undergoing vascular surgery (7707 events per 100 000), thoracic surgery (6515 events per 100 000), and transplant surgery (6261 events per 100 000) had the highest rates of perioperative MACCE. The lowest risks were observed in patients undergoing obstetric and gynecologic surgery. The frequency of perioperative MACCE by all surgery types is shown in Figure 2. After multivariable adjustment, thoracic surgery (OR, 2.07; 95% CI, 2.03-2.11), vascular surgery (OR, 1.96; 95% CI, 1.94-1.98), and transplant surgery (OR, 1.80; 95% CI, 1.67-1.95) remained associated with the highest risks of perioperative MACCE when compared with general surgery (Table 3).
Perioperative MACCE was more frequent among men than women (3521 events per 100 000 surgeries [3.5%] vs 2611 events per 100 000 [2.6%]; aOR, 1.17; 95% CI, 1.16-1.18; P < .001), with higher in-hospital mortality (1984 events per 100 000 surgeries [2.0%] vs 1439 events per 100 000 [1.4%]; aOR, 1.22; 95% CI, 1.20-1.23; P < .001), AMI (1118 events per 100 000 surgeries [1.1%] vs 832 events per 100 000 [0.8%]; aOR, 1.14; 95% CI, 1.13-1.16; P < .001) and ischemic strokes (771 events per 100 000 surgeries [0.8%] vs 592 events per 100 000 [0.6%]; aOR, 1.05; 95% CI, 1.04-1.07; P < .001). Declines in perioperative MACCE, in-hospital death, and AMI, and increases in perioperative acute ischemic stroke were observed over time in both sexes (eFigure 5 in the Supplement).
Rates of perioperative MACCE varied by race and ethnicity. Non-Hispanic black patients experienced significantly more MACCE than non-Hispanic white patients (3795 events per 100 000 surgeries [3.8%] vs 2892 events per 100 000 [2.9%]; aOR, 1.14; 95% CI, 1.13-1.16). Black patients also had higher rates of perioperative death (aOR, 1.24; 95% CI, 1.22-1.26) and stroke (aOR, 1.17; 95% CI, 1.14-1.20) than non-Hispanic white patients, although differences in the rates of adverse events narrowed over time. Trends in MACCE by race and ethnicity are shown in eFigure 6 in the Supplement.
In this analysis of 10.5 million in-hospital major noncardiac surgeries, perioperative MACCE occurred in 3.0% of patients, with vascular, thoracic, and noncardiac transplant surgeries conferring the highest risks of cardiovascular morbidity and mortality. The rate of perioperative MACCE declined from 2004 to 2013, due to reductions in perioperative death and AMI. In contrast, rates of perioperative ischemic stroke increased during the study timeframe. Men had higher risk of perioperative MACCE than women in unadjusted and multivariable adjusted models. In analyses of perioperative events by race and ethnicity, non-Hispanic black patients had the highest rates of perioperative death and ischemic stroke in comparison to other racial groups.
Since the publication of the initial Goldman multifactorial index of cardiac risk in 1977,6 adverse cardiovascular events have been recognized as a major cause of perioperative morbidity and mortality. Nearly 40 years later, efforts to identify patients at the greatest risk for perioperative MACCE and to reduce morbidity and mortality following noncardiac surgery remain ongoing. In this contemporary analysis, perioperative MACCE still occurs in 1 in every 33 hospitalizations for noncardiac surgery, corresponding to approximately 150 000 perioperative events each year in the United States. To our knowledge, this is the first study to report national data on the cardiovascular outcomes of in-hospital major noncardiac surgery in the modern era, with multivariable adjusted predictors of perioperative death, ischemic stroke, AMI, and the composite of MACCE. Furthermore, this is the largest analysis of time trends in perioperative cardiovascular outcomes in the United States.
The observed reductions in overall perioperative MACCE are encouraging. These may be due to improved surgical case selection, advances in the management of cardiovascular risk factors and disease, improved surgical techniques, including increased use of minimally invasive surgical interventions, improved anesthetic techniques, enhanced intraoperative monitoring, and advanced postoperative critical care. Reductions in myocardial infarction during the study are surprising, given the increase in the sensitivity of modern cardiac biomarkers necessary for the diagnosis of myocardial infarction during the study timeframe.16 However, declines in AMI may be attributed to improved pharmacologic and percutaneous strategies in the management of coronary artery disease, and the recognition of thrombotic risks early post-percutaneous coronary intervention.17-19
Stroke incidence rates have declined steadily over the past decades in the United States.20,21 In the context of these national trends, the observed increase in the rate of perioperative ischemic stroke is a concerning finding that warrants additional study. The rising rates of stroke in the perioperative period may be attributable to an increased prevalence of cardiovascular risk factors of surgical patients, carotid stenosis or cerebrovascular disease, atrial arrhythmias, or changes in intraoperative hemodynamic management.22 Alternatively, increases in perioperative ischemic strokes may be related to increased use of perioperative β blockade.9 Importantly, in these analyses, we found a strong association between perioperative stroke and patients undergoing vascular surgery. While vascular surgery patients are likely to have the highest atherosclerotic burden and the greatest risk for ischemic complications following surgery, this strong association also raises the possibility that perhaps some patients in this cohort underwent a major vascular procedure or surgery as a consequence of ischemic stroke. After excluding patients who underwent vascular surgery, an increase in the rate of perioperative ischemic stroke over time was still observed. This trend also persisted in sensitivity analyses of patients who were electively hospitalized for surgery, and among patients who underwent the principal noncardiac surgery within the first 72 hours of hospital admission. In these sensitivity analyses, patients were unlikely to have presented with an acute stroke prior to the primary noncardiac surgery.
There are some notable limitations of this study. First, analyses are based on administrative coding data, which may be subject to reporting bias or coding errors. Race and ethnicity data was missing for 17% of subjects. Second, the analysis was limited to adults age 45 years or older, the population at risk for cardiovascular complications of noncardiac surgery. Rates of perioperative AMI are lower in this analysis than in some previously published studies, likely due to the inclusion of larger numbers of low-risk patients from the NIS data set.9 Third, the timing of nonfatal AMI or ischemic stroke after noncardiac surgery cannot be firmly established from this administrative database. However, because major noncardiac surgery is contraindicated early after AMI or stroke, patients presenting with these acute cardiovascular conditions were unlikely to undergo major noncardiac surgery during the index hospital admission. Furthermore, trends in perioperative outcomes were similar in sensitivity analyses of patients undergoing surgery during elective hospitalization. Fourth, medical therapy was not available from this administrative data set; consequently, the use of β-blockers, antiplatelet agents, and other cardiovascular therapies in the perioperative period could not be evaluated. Fifth, although a history of heart failure was included in modeling, left ventricular function and other important markers of cardiovascular risk were not available for inclusion in this analysis. Finally, results of perioperative laboratory testing, including cardiac biomarkers, were not available from this administrative data set. Myocardial injury after noncardiac surgery, a well-described independent risk factor for short and long-term mortality, could not be ascertained and was not included in the composite outcome.4,23,24
To our knowledge, this is the largest analysis of perioperative MACCE in patients undergoing major noncardiac surgery in the United States. Cardiovascular complications after noncardiac surgery remain a major source of morbidity and mortality. Despite improvements in perioperative outcomes over the past decade, the significant increase in the rate of ischemic stroke in this analysis requires confirmation and further study. Additional efforts are necessary to improve perioperative cardiovascular care of patients undergoing noncardiac surgery.
Corresponding Author: Sripal Bangalore, MD, MHA, Division of Cardiology, Department of Medicine, New York University School of Medicine, 550 First Ave, New York, NY 10016 (email@example.com).
Accepted for Publication: October 14, 2016.
Published Online: December 28, 2016. doi:10.1001/jamacardio.2016.4792
Author Contributions: Dr Smilowitz 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. Drs Berger and Bangalore are cosenior authors and contributed equally to this article.
Concept and design: Smilowitz, Gupta, Ramakrishna, Berger, Bangalore.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Smilowitz, Ramakrishna.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Smilowitz, Gupta, Guo.
Supervision: Ramakrishna, Berger, Bangalore.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.