Preoperative Cardiac Risk Assessment: Title and subTitle BreakAn Updated Approach FREE

CARDIOVASCULAR disease is a major public health concern affecting 25% of Americans and is the number one killer in the United States, Canada, Europe, and Japan.¹ One third of noncardiac surgical patients may harbor coronary artery disease (CAD), a significant cause of perioperative morbidity and mortality after noncardiac surgery.² In a series of 1000 patients undergoing coronary angiography prior to peripheral vascular surgery, 25% of patients were found to have surgically correctable CAD.³ Thus, the importance of performing cardiac risk assessment prior to surgery to identify high-risk patients in whom adverse cardiac events may be prevented becomes clear. In most patients, however, an extensive cardiac workup is unnecessary and costly and delays definitive patient care. The purpose of this article is to provide a general framework for approaching cardiac risk assessment in patients undergoing noncardiac surgery. Risk assessment needs to be individualized and should be based on the urgency of the operation, the type of procedure planned (operation-specific risk), and the patient's clinical risk factors (patient-specific risk). We will discuss the rationale and guidelines for preoperative cardiac risk assessment, describe various risk-scoring systems and available noninvasive tests, and propose an updated treatment algorithm.

Cardiac complications are the leading cause of perioperative morbidity and mortality following noncardiac surgery. These complications are defined as unstable angina or myocardial ischemia, myocardial infarction (MI), dysrhythmias, congestive heart failure (CHF), and cardiac death. The likelihood of developing a perioperative cardiac event is related to the urgency and magnitude of the procedure and the degree of hemodynamic stress experienced by the patient. Intraoperative myocardial ischemia is often noted as ST segment changes in the lateral leads on an electrocardiogram (ECG) or as segmental wall motion abnormalities on transesophageal echocardiography. Postoperative ischemia is often silent and typically begins between postoperative days 3 and 5. Perioperative ischemic changes are precipitated by factors that increase myocardial oxygen demand such as tachycardia, hypertension, anemia, stress, or the discontinuation of β-blockers.⁴

Based on the previous factors, operations have been classified as low, intermediate, or high risk (Table 1). High-risk operations carry a perioperative adverse cardiac event or death rate of 5% or greater. These include major emergent operations, especially in elderly patients, aortic and lower-extremity vascular surgery, and anticipated prolonged procedures with expected fluid shifts or high estimated blood loss. Intermediate-risk operations are reported to carry a cardiac risk of less than 5%. These include carotid endarterectomy, uncomplicated head and neck procedures, and intraperitoneal, intrathoracic, orthopedic, and prostate surgery. For low-risk procedures such as endoscopy, superficial procedures, cataract surgery, and breast surgery, the associated cardiac risk is reported to be less than 1%.⁴ Thus, with respect to surgery-specific risk, a detailed cardiac assessment should focus on patients undergoing high-risk operations. In most instances, these operations are major vascular in nature because patients undergoing emergency surgery do not have time for an in-depth assessment. Selected patients undergoing intermediate- or low-risk operations may warrant careful evaluation based on patient-specific risk factors.

In addition to the proposed operation, patients' clinical markers of cardiac disease are important in assessing cardiac risk. Major predictors of cardiac risk include recent MI, unstable or severe angina,⁵ decompensated CHF, significant dysrhythmias, and severe valvular disease (Table 2). Reports from the 1960s and 1970s demonstrated that a prior MI within 3 months of surgery increased the perioperative rate of reinfarction to 30%. If the MI was within 3 to 6 months of surgery, the reinfarction rate was 15%; if the MI occurred more than 6 months prior to surgery, the reinfarction rate was 6%.⁶ Recently, however, Rao et al⁷ have shown that although a history of MI does increase the risk of reinfarction, the overall rate is only 1.9%, and only 5.7% for an MI within 3 months of surgery. Patients with symptomatic aortic stenosis manifested by angina, syncope, or heart failure often experience severe perioperative CHF, shock, and even sudden death after undergoing noncardiac surgery.⁸

Another way of assessing patient-specific risk is to determine functional capacity. Functional capacity is an important predictor of future cardiac events.⁹ The metabolic equivalent is a standardized way of measuring functional capacity. An assessment of the patient's functional capacity has been shown to correlate with the maximum uptake on treadmill tests and reliably predicts future cardiac events.¹⁰ Patients with poor functional capacity (ie, an inability to walk up a flight of stairs or to climb a hill) are at an increased cardiac perioperative and long-term risk.¹¹

In recent decades, several classification systems have been developed to assess cardiac risk. The Dripps index of the American Society of Anesthesiologists (ASA) is used to predict cardiac death within 48 hours of surgery.¹² Patients are divided into 5 classes of physical status according to cardiac history, with ASA 1 being a normal, healthy patient and ASA 5 being a moribund patient. Perioperative mortality rates range from 0% for ASA 1 to 9.4% for ASA 5. However, the ASA classification system has limited utility; it is very subjective and has not been uniformly reproducible. The Goldman Cardiac Risk Index and the Detsky index are multifactorial approaches to risk assessment that were developed to overcome the limitations of the ASA classification system.^{7 ,13} Unfortunately these systems are cumbersome to use and lack general applicability; their findings have been supported or refuted by an equal number of studies.^{14 - 18}

Because most procedures with high cardiac risk are vascular in nature, Eagle et al¹⁹ identified risk factors specific for these patients (Table 3). Multivariate analysis revealed 6 clinical factors that were predictive of adverse perioperative cardiac events: age older than 70 years, history of MI or Q waves on an ECG, diabetes mellitus, angina, CHF, and ventricular ectopy requiring therapy. Adverse cardiac event rates were 3.1%, 15.5%, and 50% for patients with 0, 1 or 2, and 3 or more risk factors, respectively, according to Eagle and colleagues. Several studies by de Virgilio et al^{20 - 22} have confirmed the utility of these risk factors as predictors of adverse perioperative cardiac events in patients undergoing major vascular surgery. In patients undergoing abdominal aortic aneurysm (AAA) repair, they noted a 20% adverse event rate in patients with 2 or more risk factors vs no cardiac events in patients with 1 or no risk factors.²² However, the risk factors developed by Eagle and colleagues do not account for the severity of the angina, the length of time since the last MI, or whether the CHF is decompensated.

Lee et al²³ prospectively produced the Revised Cardiac Risk Index stratification system in an attempt to simplify the Goldman index. They identified 6 independent predictors of cardiac complications in patients undergoing noncardiac surgery: high-risk surgery (intraperitoneal, intrathoracic, or suprainguinal vascular procedures), history of ischemic heart disease, history of CHF, history of cerebrovascular disease, preoperative treatment with insulin, and a preoperative serum creatinine level greater than 2.0 mg/dL (>177 µmol/L). Rates of major cardiac complications (including MI, CHF, ventricular fibrillation, primary cardiac arrest, and complete heart block) with 0, 1, 2, or 3 or more of these risk factors were 0.5%, 1.3%, 4%, and 9%, respectively.

Exercise stress testing in ambulatory patients provides a good estimate of functional capacity and can detect myocardial ischemia. The sensitivity of exercise stress testing for detecting obstructive CAD depends on the severity of the stenosis, the extent of disease, and the clinical history. Gianrossi et al²⁴ performed a meta-analysis of 147 consecutive studies involving 24 074 patients who underwent both exercise stress testing and coronary angiography. The mean sensitivity of an exercise ECG for multivessel CAD was 81% (range, 40%-100%) with a mean specificity of 66% (range, 17%-100%). Sensitivity has been shown to decrease in patients who cannot reach maximum levels of exercise. This is a major limitation of the exercise ECG in vascular patients; 50% of these patients are unable to reach maximum exercise levels. The specificity of exercise ECG is poor in the presence of ST segment/T wave abnormalities on a resting ECG or with the use of digoxin.^{25 - 26} Other limitations of the exercise ECG are that the results are difficult to interpret in the presence of left bundle branch block, the extent of myocardial ischemia cannot be determined, and it provides no information on left ventricular function.²⁷ In patients who cannot exercise, stress testing can be performed by chemical means using either dobutamine, dipyridamole thallium, or adenosine.

The utility of dipyridamole thallium scintigraphy (DTHAL) as a predictor of perioperative adverse cardiac events has been studied extensively.^{19 ,28 - 37} Early reports found a strong correlation between redistribution on DTHAL and perioperative adverse cardiac events.^{19 ,28 - 31} Cutler and Leppo²⁸ prospectively studied 116 patients undergoing aortic surgery. They found that all postoperative MIs and deaths occurred in the patients with redistribution on DTHAL. Eagle and colleagues reported retrospectively that DTHAL results were best used when combined with clinical risk factors. Patients with 1 or 2 risk factors and redistribution on DTHAL had a 29% cardiac event rate, vs 3.2% for those without redistribution.¹⁹

More recent reports question the value of DTHAL as a preoperative cardiac screening device.^{32 - 37} In a study of 457 vascular patients undergoing AAA repair, Baron et al³³ demonstrated that a history of documented CAD and an age of 65 years or older were better predictors of adverse cardiac events than perfusion imaging. The lack of correlation between abnormal DTHAL and adverse cardiac events was confirmed in 2 retrospective studies by de Virgilio et al^{21 ,32} and a prospective blinded study by Mangano et al.³⁴ More recently, a prospective blinded study by de Virgilio et al³⁷ confirmed a lack of association between reversible defects on DTHAL and adverse cardiac events in 82 patients undergoing elective vascular surgery. Unlike the previous studies, this prospective study excluded low-risk patients (ie, those with no risk factors according to the classification by Eagle and colleagues). The adverse cardiac event rate was 13.8% for patients with a reversible defect vs 9.8% for those without (P = .70). The adverse event rate in patients with 2 or more reversible defects was 12.5%, vs 11.1% in patients with fewer than 2 reversible defects. The sensitivity of DTHAL with 2 or more reversible defects was only 11%, with a specificity of 90%. The positive and negative predictive values were 12.5% and 89%, respectively. Because at least 7 studies in the literature demonstrate no correlation between DTHAL and perioperative adverse cardiac events, we currently do not recommend DTHAL as a screening tool prior to vascular surgery.

Dobutamine stress echocardiography is another noninvasive test used to detect hemodynamically significant CAD.^{38 - 43} The development of new wall motion abnormalities (NWMAs) following dobutamine administration is considered an indication of significant CAD.^{44 - 47} Unfortunately, there are few prospective blinded studies available on dobutamine echocardiography as a preoperative test. Davila-Roman et al⁴⁴ reported a 20% adverse event rate in patients with NWMAs in a retrospective study of 98 patients undergoing elective vascular surgery. There were no cardiac events in patients with normal stress test results. In a retrospective study of 81 patients undergoing AAA repair, Langan et al⁴⁵ noted no perioperative cardiac complications in the 56 patients with normal findings on an echocardiogram. Of the 25 patients with NMWAs, 9 had their noncardiac surgery deferred. Of these, 4 underwent a preoperative coronary artery bypass graft (CABG), 1 of whom died of a stroke and 1 of a ruptured AAA. Sixteen patients with NWMAs proceeded to surgery without a CABG, 3 (19%) of whom experienced a postoperative MI. Lalka et al⁴⁶ reported a 29% cardiac event rate after aortic surgery in patients with NMWAs vs a 4.6% event rate in patients with normal study results. In one prospective blinded study of 75 patients undergoing major vascular surgery, the sensitivity and negative predictive value of NWMAs on the echocardiogram were 100%, but the positive predictive value was only 19%.⁴⁷ Of note, 21% of patients had an adverse reaction to the dobutamine infusion. In the largest retrospective study to date, the Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography Study Group found that the adverse event rate was 10.6% in patients with 3 or more cardiac risk factors who had 5 or more segments of NWMAs, vs a 2% adverse event rate in patients without NWMAs.⁴⁸ This 10.6% cardiac event rate occurred despite the use of β-blockade. Four patients with severe NWMAs who were excluded from the study underwent a CABG, and 2 of them died as a result of this operation. The authors suggest that coronary angiography should be considered in patients with multiple cardiac risk factors and a positive stress echocardiogram result.

Left ventricular systolic or diastolic dysfunction in the perioperative period has been shown to be a good predictor of postoperative CHF and of death in critically ill patients.⁴ Several retrospective and prospective studies have demonstrated that a decreased preoperative ejection fraction is associated with increased perioperative morbidity and mortality.^{49 - 54} However, in a more recent study, McCann and Wolfe⁵⁵ found no significant difference in perioperative mortality or cumulative life-table survival rates between patients with a normal and low left ventricular ejection fraction.

Most physicians would agree that preoperative coronary angiography should be performed in patients with overt cardiac symptoms such as severe or unstable angina.⁵ In the absence of severe symptoms, the indications for angiography are controversial and must be individualized. Coronary angiography with possible revascularization should be considered for patients undergoing elective high-risk operations (particularly aortic surgery) who have a positive stress test result and multiple cardiac risk factors.

To date, no prospective randomized studies show a lower perioperative cardiac event rate in patients undergoing cardiac evaluation and a prophylactic preoperative CABG. Retrospective studies, however, do demonstrate a decreased perioperative cardiac event rate and an improved 5-year survival rate in patients with a preoperative CABG.^{56 - 59} In 1000 consecutive patients undergoing preoperative coronary arteriography, Hertzer⁶⁰ and colleagues found that 251 patients had severe correctable CAD. Of these, 216 patients underwent a CABG, with a 5.5% mortality rate. There was an additional 0.9% mortality rate due to AAA rupture after the CABG. The overall mortality rate was 3.9% for patients undergoing AAA repair and 2.4% for lower extremity revascularization. The late mortality rate was the same for patients undergoing a preoperative CABG as those without clinically evident CAD. Follow-up from the Coronary Artery Surgery Study revealed a 0.9% mortality rate after noncardiac surgery in patients who had previously undergone a CABG, vs a 2.3% mortality rate for medically treated patients with CAD.⁶¹ In patients undergoing high-risk surgery (abdominal, thoracic, or head and neck surgery), a prior CABG reduced the postoperative MI rate from 2.7% to 0.8% (P = .002) and the mortality rate from 3.3% to 1.7% (P = .03), compared with patients treated medically. However, there was no benefit to undergoing a CABG prior to a low-risk procedure. Patients with severe angina and/or multivessel disease were found to benefit the most.⁶²

Although a CABG does appear to afford myocardial protection during subsequent surgery, one must factor in the risk of the intervention, the consequences of delaying definitive care of the vascular disorder (eg, ruptured AAA or limb loss), and the cost. Coronary arteriography has a 0.3% risk of mortality. A coronary artery bypass graft carries an operative mortality rate of 3% overall, and closer to 5% in patients with peripheral vascular disease. Thus, the strategy of coronary angiography with a possible CABG reduces overall mortality only when the estimated mortality rate of the proposed operation is substantially higher than 5%.⁶³ Even then, mortality would be reduced at the expense of greater morbidity and higher cost. Bry et al⁶⁴ estimated that the costs of cardiac screening and intervention were $392 253 per life saved and $181 039 per MI prevented.

In recent years, several authors have recommended against prophylactic coronary revascularization prior to vascular surgery. Massie et al⁶⁵ noted that any benefit from invasive cardiac evaluation was offset by several deaths and MIs that complicated this evaluation. Seeger et al³⁶ reported that stress thallium imaging confirmed a high incidence of CAD in aortic surgery patients but that prophylactic cardiac intervention did not lower operative or long-term mortality rates. Taylor et al⁶⁶ limited cardiac evaluation to patients with severe symptomatic CAD (unstable angina, severe CHF, or uncontrolled dysrrhythmias), which represented only 5.8% of their subjects. Of these, only 0.5% had a prophylactic CABG. Despite the paucity of cardiac evaluation and intervention, the perioperative MI rate for the overall group was just 3.9%, with no deaths in patients undergoing elective vascular surgery. This morbidity and mortality rate is no different from that of studies using routine cardiac screening. Our vascular group espouses a similar selective approach for the cardiac assessment of patients undergoing elective aortic surgery.⁶⁷

Several studies have analyzed the role of β-blockers in reducing the rate of perioperative myocardial ischemia, infarction, and death. Stone et al⁶⁸ reported a 28% frequency of ST segment depression in patients treated with a placebo vs a 2% rate in patients treated with oral β-blockade. Similarly, Pasternack et al⁶⁹ reported an 18% MI rate in controls vs a 3% perioperative MI rate in patients treated with oral metoprolol immediately following the operation and with intravenous metoprolol during AAA repair. Mangano et al⁷⁰ noted that with the use of atenolol, the postoperative mortality rate was reduced from 14% to 3% during the first year and from 21% to 10% the second year after noncardiac surgery. The study by Poldermans et al⁷¹ demonstrated that the 30-day complication rate in patients with NWMAs who were given perioperative β-blockade was similar to that in patients without NWMAs. More recently, in a follow-up retrospective review, they confirmed the benefit of β-blockade in intermediate-risk patients. However, β-blockade failed to lower the cardiac event rate in patients with very high risk (≥3 clinical risk factors and ≥5 NWMAs).⁴⁸

Every patient undergoing nonemergent surgery should undergo cardiac risk assessment (Figure 1). This assessment may be limited to a careful history and physical examination. The first goal is to identify the patient with high cardiac risk. For patients with unstable angina, acute MI, symptomatic aortic stenosis, or decompensated CHF, the operation poses an inordinate cardiac risk and should be postponed if the condition is not immediately life threatening. In most instances these patients should proceed to coronary angiography if surgery is being contemplated. If no severe cardiac symptoms are present, the number of cardiac risk factors should be determined. If the cardiac history is positive, it is important to establish whether the cardiac disease is active, its severity, and whether coronary revascularization has been performed. If a patient has a cardiac history but has undergone coronary revascularization within the last 5 years and is now symptom free, such a patient has a low cardiac risk and can proceed to surgery.

If the patient has 1 or no cardiac risk factors, the operation will carry a low predicted risk and can be performed without further cardiac testing. This applies to both low- and high-risk surgery (aortic and major vascular). Perioperative β-blockade should be used in most patients with at least 1 cardiac risk factor. On the other hand, if the patient has multiple cardiac risk factors, one should consider the type of operation to be performed. If the proposed operation is low or moderate risk, the patient can proceed to surgery without additional testing. If the patient with multiple cardiac risk factors is about to undergo a high-risk operation, the urgency of that operation must be considered. If the patient has limb-threatening ischemia or a very large or symptomatic nonruptured AAA, further cardiac testing should probably be omitted. For these patients, one can change the proposed operation to a less stressful procedure or consider perioperative β-blockade. In a purely elective case (eg, an aortobifemoral bypass in a patient with claudication), dobutamine stress echocardiography should be considered. If the results of the stress echocardiogram are normal, one can proceed with surgery. If the echocardiogram shows NWMAs, the number of segments should be determined. If there are 5 or more segments of NWMAs, there are 4 options: cancel the surgery, modify the operation (axillofemoral bypass instead of aortobifemoral bypass), proceed with surgery using β-blockade, or perform coronary angiography. The decision of which option to choose must be individualized based on the indications for surgery, age, expected life span, and the estimated risk of coronary intervention.

In summary, adverse cardiac events remain a significant source of morbidity and mortality after noncardiac surgery. In a small subgroup of patients, major predictors of cardiac risk will be identified during a clinical evaluation. In these patients, coronary angiography will likely be necessary. With intermediate-risk patients, the urgency and type of surgical procedure should be taken into account; however, the use of noninvasive cardiac stress testing should be restricted. As we enter the new millennium, enthusiasm for aggressive cardiac evaluation has been tempered by studies that question the predictive value of these techniques. Cardiac testing may be a double-edged sword. Although it can identify significant CAD, it adds expense and delays definitive treatment. Stress test results may call for invasive cardiac procedures that can lead to the very same cardiac complications they are designed to prevent. We recommend that stress testing be considered only for the small group of patients who are undergoing elective high-risk surgery and have multiple cardiac risk factors, and then only if there are no risks to delaying the operation and if coronary revascularization has a low mortality rate at one's particular institution. In the vast majority of patients, advances in perioperative medical and anesthetic treatment, including the use of β-blockade, have led to acceptably low rates of cardiac complications, rendering cardiac stress testing unnecessary.

Corresponding author and reprints: Christian de Virgilio, MD, Harbor UCLA Medical Center, Department of Surgery, 1000 W Carson St, Torrance, CA 90509.