Etiology of Acute Coronary Syndrome after Noncardiac Surgery

MYOCARDIAL infarction (MI) is a common and serious complication after noncardiac surgery.^1–3  Perioperative MI is associated with significant mortality that often exceeds mortality after acute MI in a nonsurgical setting.^4–6  Acute myocardial infarction often presents as acute coronary syndrome, a syndrome that encompasses several distinct entities, namely MI and unstable angina, but with similar clinical signs and symptoms (e.g., chest pain). The etiology of acute coronary syndrome in a nonsurgical setting is generally well understood and involves either a thrombotic cause (plaque rupture, type 1 MI, or stent thrombosis, type 4B) or demand ischemia (type 2 MI) or a combination of the two factors.^7,8  In contrast, etiology and pathophysiology of acute coronary syndrome in the context of noncardiac surgery are incompletely understood, and there is ongoing controversy regarding whether thrombosis or demand ischemia is the dominant cause.^9–13 

The objective of this investigation was to determine the etiology of perioperative acute coronary syndrome with a particular emphasis on thrombosis versus demand ischemia. We studied a cohort of surgical patients who underwent diagnostic coronary angiography after presenting with acute coronary syndrome within 30 days after noncardiac surgery.

This was a retrospective cohort study of adult patients undergoing noncardiac surgery between January 2008 and July 2015 at Barnes–Jewish Hospital (St. Louis, Missouri) who developed acute coronary syndrome and received a coronary angiogram within 30 days of surgery. We screened the Barnes–Jewish Hospital billing database for patient records meeting inclusion criteria, reviewed the tentative list of patients case by case for potential exclusion criteria, and then matched the 146 identified patients with the local CathPCI Registry of the National Cardiovascular Data Registry, which captures all patients who undergo diagnostic angiography and percutaneous coronary intervention at our institution.¹⁴  The study was approved by the Washington University School of Medicine Institutional Review Board (St. Louis, Missouri) and granted a waiver of informed consent, and it was not registered at a registry.

Inclusion criteria were age more than 18 yr; noncardiac surgery requiring monitored anesthesia care, general, or regional anesthesia; and the diagnosis of acute coronary syndrome. Exclusion criteria were cardiac surgery or related procedures (such as pacemaker insertion), surgery under local anesthesia, age less than 18 yr, and indications for cardiac catheterization other than acute coronary syndrome.

Records were reviewed for all subjects who underwent a coronary angiogram within 30 days of noncardiac surgery and used to identify demographics, medical history, symptoms, electrocardiographic changes, cardiac biomarkers concentrations, echocardiography findings, medications, and transfusion requirements. Noncardiac surgeries were categorized as vascular versus nonvascular surgery. Acute coronary syndrome events were categorized as ST–elevation MI, non–ST–elevation MI, or unstable angina. The 30-day mortality and 1-yr mortality were obtained reviewing the social security registry and from the https://www.familysearch.org website (accessed July 1, 2017).

Coronary angiograms were independently reviewed by two blinded interventional cardiologists (P.L. and A.A.). In cases of discordant angiogram interpretation, angiogram findings were jointly reviewed by the interventionalists and a consensus was reached. Myocardial infarction was classified by type according to the Third Universal Definition.¹⁵  All lesions resulting in more than 50% stenosis were analyzed and characterized by angiographic features of complexity adapted from Goldstein et al.¹⁶  A lesion was considered complex if it exhibited any of the following features: (1) an intraluminal filling defect suggestive of thrombus as defined by abrupt vessel occlusion with contrast persistence; (2) plaque ulceration defined by the extension of contrast beyond but contiguous to the vessel lumen; or (3) two or more of the following: fissuring, evidenced by intraplaque dye penetration not fulfilling the definition of ulceration; irregular margins or overhanging edges; or hazy appearance. A spontaneous MI (type 1) was considered to have occurred by the presence of at least one complex lesion.

Primary study outcomes included types of MI: type 1 (plaque rupture, coronary occlusion), type 2 MI (demand ischemia), and type 4B (stent thrombosis); acute coronary syndrome was also classified per electrocardiogram and biomarker presentation as ST–elevation MI, non–ST–elevation MI, or unstable angina. Secondary outcomes included hospital length of stay, as well as 30-day and 1-yr mortality rates. Significant bleeding was defined as more than 1 l of blood loss.

We included a statistical analysis plan in the study protocol that was submitted and approved by the institutional review board. Primary outcomes were established a priori, but several additional analyses were performed post hoc and have been identified as such in this article. The data are presented as means and SD or medians and interquartile range for nonnormally distributed data. Univariate comparisons were made with a Mann–Whitney U test, chi-square test, or Fisher exact test. We calculated unadjusted odds ratios from 2 × 2 tables. Kaplan–Meier survival analyses were performed to compare survival probabilities between types of MI and MI presentation as ST–elevation MI, non–ST–elevation MI, or unstable angina. A log-rank test was used to determine statistical significance. There was no formal sample size estimation, because the goal was to capture all patients who developed acute coronary syndrome after noncardiac surgery in our institution. Analyses were performed using SPSS 23.0 (IBM, USA) or JMP 13.2 (SAS Institute, USA). All analyses were two-sided, and P < 0.05 was considered statistically significant.

After screening 215,077 patients who underwent noncardiac surgery in our hospital during the study period, the final study population of patients who underwent coronary angiography due to acute coronary syndrome within 30 days after noncardiac surgery was 146 patients (0.07%; fig. 1; table 1). More than 50% of the patients had a diagnosis of coronary artery disease, and 19% had a diagnosis of heart failure. Acute coronary syndrome after emergent surgery accounted for 8.2% of cases in this cohort (n = 12). Median estimated intraoperative blood loss was modest (150 ml, interquartile range of 0 to 405 ml), although 11.0% (n = 16) experienced significant intraoperative hemorrhage.

More than 80% of acute coronary syndrome events were classified as non–ST–elevation MI (n = 117); smaller fractions were classified as ST–elevation MI (n = 21, 14.4%) and unstable angina (n = 8, 5.5%, table 2). Approximately half of events presented with clinical symptoms such as chest pain or shortness of breath; 22% of events were clinically silent. Electrocardiogram changes consistent with myocardial ischemia were observed in more than 90% of patients. Eight percent of patients presented in cardiogenic shock (n = 11). The majority of patients developed acute coronary syndrome within the first 3 postoperative days (63%, n = 92); 16% of patients (n = 23) developed acute coronary syndrome in the intraoperative period. On average, a median of 3 days passed between acute coronary syndrome presentation and coronary angiography, except for patients with acute stent thrombosis (type 4B MI; n = 3), who were taken to the catheterization laboratory on the same day.

Patients with type 1 MI received medical management in 68% of cases (n = 25), underwent coronary artery bypass grafting surgery surgery in 30% of cases (n = 11), and received a percutaneous coronary intervention (coronary stent) in 3% of cases (n = 1). Patients with type 2 MI were predominantly managed medically (94%; n = 100).

Among all acute coronary syndrome events, 73% were adjudicated as type 2 MI (demand ischemia; n = 106,) after review of angiography films. Acute coronary thrombosis (type 1 MI) was adjudicated in 25% of patients (n = 37), and acute coronary stent thrombosis (type 4B) was adjudicated in 2% of patients (n = 3; table 3). There was no association between type 1 and type 2 MI and clinical presentation as ST–elevation MI or non–ST–elevation MI (P = 0.97). Absent or only mild (nonobstructive) coronary artery disease was found in 27% of patients (n = 39), 5% of with type 1 MI (n = 2), and 35% of patients with type 2 MI (n = 37). In 10% of events (n = 14), acute coronary syndrome was likely caused by stress-induced cardiomyopathy and not by an intracoronary event. In a post hoc analysis, stress-induced cardiomyopathy was observed in two patients on chronic β-blocker therapy (n = 67; 3%) compared to 12 patients who did not take β-blockers (n = 79; 15%; odds ratio, 0.17; 95% CI, 0.03 to 0.78; P = 0.02).

Unadjusted 30-day and 1-yr mortality rates for all patients were 7 and 14%. Survival rates were not statistically significant between type 1 MI, type 2 MI, and type 4B MI (fig. 2; P = 0.52) but were statistically significant between patients who experienced an ST–elevation MI versus non–ST–elevation MI (fig. 3; P = 0.02). Additional results are available as Supplemental Digital Content (https://links.lww.com/ALN/B628).

Findings from this investigation in 146 surgical patients who developed perioperative acute coronary syndrome within 30 days after surgery and who underwent coronary angiography show that: (1) three of four events were due to demand ischemia (type 2 MI) and presented as non–ST–elevation MI, (2) more than 90% had electrocardiogram changes consistent with myocardial ischemia, (3) one in four patients had no evidence of obstructive coronary artery disease, (4) 10% of events were likely due to stress-induced cardiomyopathy, and (5) acute coronary syndrome was associated with marked short- and long-term mortality (7% 30-day mortality and 14% 1-yr mortality).

The results of our study suggest that most cases of perioperative acute coronary syndrome are triggered by demand ischemia, and only a small proportion are due to an acute thrombotic event, such as coronary plaque rupture. Acute coronary syndrome events captured in our cohort probably represent one of the most severe phenotypes of perioperative cardiac events, because nearly all of them were clinically readily apparent and triggered a cardiology consultation before coronary angiography. It was a priori unclear whether the dominant cause was acute thrombosis or demand ischemia. Mechanistically both mechanisms seemed equally probable: surgery causes acute inflammation and a prothrombotic state, whereas hypoxemia, surgical stress, hyper- or hypotension, and anemia increase demand and reduce supply. Autopsy studies of fatal MI after noncardiac surgery report about half the cases as caused by plaque rupture and acute thrombosis.^17,18  A larger proportion of thrombotic events was also reported in two separate angiography studies of nonfatal MI after noncardiac surgery: Berger et al.¹⁹  found a 63% rate (n = 48), and Gualandro et al.²⁰  found a 50% rate (n = 120). Conversely, a recent angiography study reported a 22% percutaneous coronary intervention rate due to a thrombotic event among 1,093 patients with acute coronary syndrome after noncardiac surgery.²¹  Duvall et al.¹¹  found a 55% rate of type 2 MI versus a 26% type 1 MI rate (n = 66). In a recent study among 103 veterans with coronary artery disease and previously implanted stents who developed perioperative MI, 48% were found to have demand ischemia (stable obstructive coronary artery disease), and 46% were caused by a thrombotic event.²²  It is important to note that several independent lines of evidence strongly suggest that most events of myocardial ischemia after noncardiac surgery are silent and caused by demand ischemia, as detected by Holter electrocardiogram monitoring or postoperative cardiac troponin elevation.^23–28 

The electrocardiogram presentation of acute coronary syndromes typically falls into one of two main categories: ST–elevation MI or non–ST–elevation MI (some patients have unspecific electrocardiogram changes). Transmural myocardial ischemia caused by a type 1 MI often presents as ST–elevation MI. Demand ischemia (type 2 MI) causing subendocardial ischemia presents mostly as non–ST–elevation MI. Our data showed that there was no association between type 1 and 2 MI and presentation as ST–elevation MI versus non–ST–elevation MI. Only acute stent thrombosis appeared to present mostly as ST–elevation MI. Therefore, it may clinically not be feasible to speculate about the underlying etiology of a perioperative acute coronary syndrome event based on its electrocardiogram presentation.

A surprising finding of this investigation was that in 27% of coronary angiographies after perioperative acute coronary syndrome, neither a culprit lesion nor obstructive coronary artery disease was found, with some patients having a normal coronary anatomy. In a small angiography study, Ellis et al.²⁹  also found evidence for postoperative acute coronary syndromes occurring in the absence of obstructive coronary artery disease. In 10% of our patients, the angiographic findings were most consistent with stress-induced cardiomyopathy (Takotsubo syndrome) and not a coronary process.

This study had several limitations. First, the study population was highly selective and does not represent the full spectrum of postoperative MI or acute coronary syndrome. On the one hand, the most severe cases of postoperative acute coronary syndromes are associated with sudden cardiac death and fatal MI—patients we did not capture. On the other hand, the majority of ischemic events in the postoperative period are silent and often missed clinically. Second, the initial identification of patients was by the hospital billing database, a method with significant limitations. Although we reviewed each case and matched it with the prospectively designed National Cardiovascular Data Registry, there is a possibility we missed patients. For example, if patients presented with acute coronary syndrome to an outside hospital that was not part of our hospital network, we would have been unable to capture the event and thus missed the patient. Third, although the sample size of 146 patients represents one of the largest thus far in the literature and has been collected from more than 215,000 noncardiac surgeries at our hospital, it is still rather small and prone to bias. In addition, we cannot exclude temporal trends that may have influenced procedural and medical management of postoperative acute coronary syndrome. Lastly, we could not capture data derived from novel adjunct techniques now sometimes used during coronary angiography, such as intravascular ultrasound or data on left ventricular hypertrophy.

Evidence from this investigation suggests that the dominant mechanism of acute coronary syndrome after noncardiac surgery is demand ischemia and that a sizable proportion of patients develop acute coronary syndrome without evidence for obstructive coronary artery disease, with some of these due to stress-induced cardiomyopathy. Thus, it appears plausible that strategies that reduce myocardial oxygen demand (e.g., β-blocker therapy, adequate pain control) and increase myocardial oxygen supply (e.g., adequate oxygenation and coronary blood flow, correction of anemia [if hemoglobin was less than 8 g/dl], adequate intravascular volume status, correction of hypotension) will likely have the biggest impact in preventing and treating acute coronary syndrome in patients undergoing noncardiac surgery compared to antithrombotic therapy.

The authors thank Tom C. Bailey, M.D., Division of Infectious Diseases, Washington University, St. Louis, Missouri, and John Lasala, M.D., Ph.D., Cardiovascular Division, Department of Internal Medicine, Washington University, for help with initial study design and database query.

This work was supported by departmental funds only. Dr. Nagele is currently supported by National Institutes of Health (Bethesda, Maryland) grant No. R01HL126892. Dr. Amin is funded via a KM1 career development award from the Clinical and Translational Science Award program of the National Center for Advancing Translational Sciences of the National Institutes of Health (grant Nos. UL1TR000448, KL2TR000450, and TL1TR000449), National Cancer Institute of the National Institutes of Health (grant No. 1KM1CA156708-01), an R18 grant from the Agency for Healthcare Research and Quality (Rockville, Maryland; grant No. R18HS0224181), and two additional grants by the Barnes–Jewish Hospital Foundation (St. Louis, Missouri).

Dr. Nagele reports receiving research grants and other research support from Roche Diagnostics (Indianapolis, Indiana) and research grants and other research support from Abbott Diagnostics (Lake Forest, Illinois). Dr. Amin has received research funding from Volcano (San Diego, California) and is a consultant for Terumo (Somerset, New Jersey), The Medicines Company (Parsippany, New Jersey), and AstraZeneca (Wilmington, Delaware). The other authors declare no competing interests.