The annual worldwide surgical volume may exceed 300 million cases.¹ Because intraoperative techniques and postoperative care and monitoring have improved, the surgical population has expanded to include patients who once would have been considered too sick to undergo surgery because of their cardiovascular disease (CVD). However, although these patients are now receiving surgery, they remain at increased risk for perioperative cardiac complications. In the Perioperative Ischemia Evaluation (POISE) and POISE 2 trials examining over 18,000 patients with established CVD or other high-risk features undergoing major surgery, the incidence of cardiac-related death or nonfatal myocardial infarction (MI) was approximately 7%.²,³

The risk of perioperative cardiac complications depends on both patient-related risk factors and the risk of the operation. The presence of end-organ CVD or its equivalents is the main patient-related risk factor. This includes coronary artery disease (CAD), heart failure (HF), severe valvular heart disease, and serious arrhythmias. In addition, disease in other organ systems resulting from atherosclerotic vascular disease or chronic hypertension, such as chronic kidney disease or stroke, is also associated with higher risk of perioperative cardiac complications.⁴ Diabetes has also been shown to be an independent predictor, especially if it requires insulin to treat, is long standing, or has been poorly controlled. In contrast, many traditional atherosclerotic risk factors such as hyperlipidemia, tobacco smoking, family history of premature heart disease, and mild-moderate hypertension add little to risk prediction in the absence of identifiable end-organ disease. Advanced age likely increases risk independently of comorbidity. Severe systemic disease, even if unrelated to the cardiovascular system, also increases cardiac risk.

High-risk noncardiac operations are typically prolonged cases involving extensive blood loss, large-volume fluid shifts, or abrupt changes in afterload. Major vascular surgery involving the aorta (eg, abdominal aortic aneurysm repair) is often considered the epitome of a high-risk operation. A major operation performed on an emergency basis can have a much higher risk of cardiac complications than does the same surgery done electively. In contrast, minimally invasive operations, endoscopic cases, or procedures performed on the skin or eye have a very low risk of cardiac complications. Patients undergoing low-risk procedures generally do not require an in-depth preoperative cardiac evaluation.

Death from cardiac causes (cardiac arrest) and nonfatal MI are the most important complications, and are often combined as a composite endpoint termed major adverse cardiac events or MACE.⁵ The definition sometimes also includes other events such as stroke, need for urgent coronary artery revascularization, or overall mortality. In some cases, the definition of perioperative cardiac complication is expanded to include pulmonary edema, angina, and non-life-threatening arrhythmia. Although the risks of these less serious complications cannot be ignored, their impact on the decision on whether to perform surgery is substantially lower. Thus, when contemplating or communicating risk estimates, it is important to understand what outcomes are included.

Owing to the complex interaction between patient-related risk factors and risk that is intrinsic to the operation, clinical risk assessment has long relied on formal multivariate prediction tools. The most widely used tool is the Revised Cardiac Risk Index (RCRI).⁹ The RCRI uses six predictive variables: history of ischemic heart disease; HF; stroke or transient ischemic attack; serum creatinine greater than 2.0 mg/dL (177 µmol/L); diabetes treated with insulin; and high-risk surgery defined as major vascular surgery or an open abdominal or thoracic operation (see Table 25.1). The risk for MACE (defined as MI, primary cardiac arrest, pulmonary edema, and complete heart block) increases with the number of predictors present. More recently, a cardiac risk calculator was derived from the American College of Surgeons’ National Surgical Quality Improvement Program (NSQIP) database.¹⁰ The NSQIP calculator (available online or through a mobile device application at https://qxmd.com/calculate/calculator_245/gupta-perioperative-cardiac-risk) uses five variables: type or location of surgery; age; serum creatinine; ability to perform activities of daily living; and the American Society of Anesthesiology (ASA) physical status classification. The ASA classification is an overall subjective assessment of the severity of systemic disease in a surgical patient (see Table 25.2). The outcome predicted by the NSQIP tool is
the 30-day incidence of myocardial infarction or cardiac arrest, leading to it being called the MICA calculator.

When assessing the risks predicted by these two tools, only the risk of MI, cardiac death, and nonfatal cardiac arrest outcomes from the RCRI should be used for comparison.¹¹ The NSQIP calculator has greater external validity, because it was derived and validated in a large database that included patients and hospitals across the United States, whereas the RCRI originated from a single tertiary care hospital. Routine postoperative surveillance ECG and cardiac enzyme levels were obtained in many of the patients who were evaluated to derive the RCRI, whereas screening for MI was not part of the NSQIP study.

Although other prediction models and tools exist and may perform better in specific populations, the RCRI and NSQIP MICA calculator are broadly applicable, easy to use, and incorporated into clinical practice guidelines from North America and Europe. The 2014 American College of Cardiology (ACC) and American Heart Association (AHA) guideline on cardiac evaluation for noncardiac surgery endorses both the RCRI and NSQIP MICA tool for clinical risk assessment.¹² Patients with two or more RCRI predictors or an estimated complication risk greater than or equal to 1%, as calculated by the NSQIP tool, are defined as being at elevated risk.

Preoperative noninvasive stress testing is sometimes performed to improve risk stratification in patients with known or suspected ischemic heart disease. Most studies on the predictive utility of stress testing have been conducted in patients undergoing major vascular surgery and utilized radionuclide
myocardial perfusion imaging or echocardiography. It remains unclear how well the conclusions from these studies apply to patients undergoing nonvascular operations, especially where the baseline risk for MACE is much lower than that with vascular surgery. In patients undergoing vascular surgery (and presumably other surgical populations), stress testing has strong negative predictive value. Patients without demonstrated ischemic potential typically have a very low risk of cardiac death or MI and can proceed to surgery without further cardiac testing. The positive predictive value of a stress test is limited. Evidence of ischemic potential on a stress test predicts increased risk in major vascular surgery; however, the large majority of patients with a positive stress test will not suffer a serious perioperative cardiac complication.¹³,¹⁴ The complication rate correlates with the amount of myocardium at risk. In a meta-analysis, stress tests showing only fixed abnormalities or very limited ischemic potential (<20% of myocardium at risk) had no predictive value.¹⁵ Tests results showing 20% to 29% reversibility were associated with a borderline significant increase in pretest probability of adverse events (positive likelihood ratio of 1.6). In patients with 30% to 49% of myocardium at risk, the posttest probability of complications increased by roughly threefold (positive likelihood ratio of 2.9), and those showing greater than or equal to 50% reversibility had an 11-fold increase in risk.

Because a positive test result generally does not increase the probability of complications by a large margin, stress testing should only be reserved for patients determined to have a high pretest probability of MACE based on clinical assessment or who have indications for the test independent of the surgery. Guidelines from the ACC/AHA and European Society of Cardiology and European Society of Anesthesiology (ESC/ESA) identify patients who have both multiple RCRI predictors and poor functional capacity (<4 METs) who are undergoing major surgery as the most appropriate population for noninvasive stress testing.¹⁶ A pharmacologic stress test (pharmacologic myocardial perfusion imaging or dobutamine stress echocardiography) is generally preferred because these modalities have been better studied, and many stress testing candidates will have limited exercise capacity.

The indications for preoperative resting transthoracic echocardiography (TTE) are similar to those for nonsurgical patients. Patients with a new or suspected HF diagnosis without a prior TTE should undergo the study before surgery, if possible. Similarly, patients with worsening HF symptoms should have a reassessment of left ventricular function with TTE. Patients with a murmur or other findings that raise concern for serious valvular heart disease (particularly aortic stenosis) should be evaluated with TTE. Patients with established moderate to severe valvular disease should undergo surveillance echocardiography if not performed in the prior year or if cardiovascular symptoms have worsened. Similarly, TTE should be considered when there is concern for other undiagnosed structural heart disease that can assessed by echocardiography (eg, hypertrophic obstructive cardiomyopathy, severe pulmonary hypertension).