American College of Cardiology

EAGLE ET AL., PERIOPERATIVE CARDIOVASCULAR EVALUATION FOR NONCARDIAC SURGERY UPDATE
http://www.acc.org/clinical/guidelines/perio/update/periupdate_index.htm

ACC/AHA Guideline Update for Perioperative Cardiovascular Evaluation for Noncardiac Surgery

A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Update the 1996 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery)

This is a Guideline Update of the 1996 Perioperative Guidelines. To highlight the changes, deleted text is indicated by strikeout, and revised text is presented in red. A clean version of the document, with changes fully incorporated, is available for download and print.

IV. Type of Surgery

Cardiac complications after noncardiac surgery are a reflection of factors specific to the patient, the operation, and the circumstances under which the operation is undertaken. To the extent that preoperative cardiac evaluation reliably predicts postoperative cardiac outcomes, it may lead to interventions that lower perioperative risk, decrease long-term mortality, or alter the surgical decision-making process. Such alterations might include either choosing a lower-risk, less-invasive procedure or opting for nonoperative management (e.g., recommending an endovascular rather than open operative approach for a particular aneurysm or occlusive lesion, electing to follow-up rather than operate on a moderate-sized (4 to 5 cm) infrarenal aortic aneurysm, or choosing nonoperative treatment for the disabled claudicant who has no limb-threatening ischemia).

To the extent that preoperative cardiac evaluation can identify potentially reducible cardiac risks, interventions directed at reducing those risks might improve both short- and long-term cardiac outcomes. The potential for improvement in long-term outcomes is particularly relevant to operative decision making in patients undergoing surgery directed at long-term goals. When, for example, surgery in asymptomatic individuals is undertaken with the objective of prolonging life (e.g., elective repair of aortic aneurysm) or preventing a future stroke (e.g., carotid endarterectomy), the decision to intervene must be made with the expectation that the patient will live long enough to benefit from the prophylactic intervention.

Although different operations are associated with different cardiac risks, these differences are most often a reflection of the context in which the patient undergoes surgery (stability or opportunity for adequate preoperative preparation), surgery-specific factors (e.g., fluid shifts, stress levels, duration of procedure, or blood loss), or patient-specific factors (the incidence of CAD associated with the condition for which the patient is undergoing surgery).

A. Urgency

Mangano ⁽¹⁾ determined that cardiac complications are two to five times more likely to occur with emergency surgical procedures than with elective operations. This finding is not surprising because the necessity for immediate surgical intervention may make it impossible to evaluate and treat such patients optimally. For instance, collected data have confirmed that the composite mortality rate for elective repair of patients with asymptomatic abdominal aortic aneurysms is significantly lower (3.5%) than that for ruptured aneurysms (42%) ⁽⁷⁷⁾. The mortality rate for graft replacements of symptomatic but intact abdominal aortic aneurysms remains relatively high (19%) despite the fact that, like elective cases, they are not associated with antecedent blood loss or hypotension. Unfortunately, most true surgical emergencies (e.g., symptomatic abdominal aortic aneurysms, perforated viscus, or major trauma) do not permit more than a cursory cardiac evaluation.

In addition, some situations do not lend themselves to comprehensive cardiac evaluation, although surgical care may qualify as semielective. In some patients, the impending danger of the disease is greater than the anticipated perioperative risk. Examples include patients who require arterial bypass procedures for limb salvage or mesenteric revascularization to prevent intestinal gangrene. Patients with malignant neoplasms also pose a diagnostic and therapeutic dilemma with respect to preoperative cardiac evaluation, especially when it is difficult to determine whether the malignancy is curable before surgical exploration. Each of these situations illustrates the importance of close communication among consultant, surgeon, and anesthesiologist to plan an approach for cardiac assessment that is appropriate for the individual patient and the underlying disease.

B. Surgical Risk

For elective surgery, cardiac risk can be stratified according to a number of factors, including the magnitude of the surgical procedure. Some operations are simply more dangerous than others. Backer et al ⁽⁷⁸⁾ encountered no cardiac complications after 288 ophthalmologic procedures in 195 patients with a prior history of MI compared with a reinfarction rate of 6.1% for a number of nonophthalmologic surgeries at the same center. A recent large-scale study supported the low morbidity and mortality rates in superficial procedures performed on an ambulatory basis. Warner et al ⁽⁷⁹⁾ determined the perioperative (30-day) incidence of MI and cardiac death in 38,500 patients who underwent 45,090 consecutive anesthesias. Fourteen (0.03% anesthesia) perioperative MIs occurred, of which two resulted in death on postoperative day 7 after the infarction. Two MIs occurred either intraoperatively or within the first 8 hours, one of which was fatal. Using age- and gender-adjusted annual incidence rates for MIs and sudden death, the authors predicted that 17.8 MIs should have occurred among this population during the study period, suggesting that these events may have occurred independent of the procedure. Several large surveys have demonstrated that perioperative cardiac morbidity is particularly concentrated among patients who undergo major thoracic, abdominal, or vascular surgery, especially when they are 70 years or older ^(1,78,80-82). Ashton et al ⁽¹⁵⁾ prospectively studied the incidence of perioperative MI associated with thoracic, abdominal, urologic, orthopedic, and vascular surgery in a cohort of 1487 men older than 40 years. The highest infarction rate (4.1%; odds ratio, 10.39; 95% confidence interval [CI], 2.3 to 47.5) occurred in the subset of patients with an established diagnosis of CAD. Nevertheless, independent significant risk factors for infarction also included age greater than 75 years (odds ratio, 4.77; 95% CI, 1.17 to 19.41) and the need for elective vascular surgery even in the absence of suspected CAD (adjusted odds ratio, 3.72; 95% CI, 1.12 to 12.37).

Few procedure-specific data are available regarding perioperative cardiac morbidity in most surgical specialties, perhaps because advanced age and serious, incidental CAD are assumed to be distributed randomly within groups of patients who undergo noncardiac operations in such fields as general surgery, thoracic surgery, orthopedics, urology, gynecology, and neurosurgery. Pedersen et al ⁽⁸³⁾ found by logistic regression that age greater than or equal to 70 years, MI within the preceding 12 months, and HF were associated with an increased incidence of postoperative cardiac complications in a series of 7,300 patients who underwent a mix of both "major" and "minor" gastrointestinal, urologic, gynecologic, and orthopedic procedures. Marsch et al ⁽⁸⁴⁾ reached similar conclusions in a much smaller series of 52 patients who required elective hip arthroplasty; the 11 patients in this study who had previous clinical indications of CAD sustained significantly higher rates of monitored ischemia or MI during the perioperative period (adjusted odds ratio, 1.9; 95% CI, 0.7 to 5.2) and late cardiac events during 4 years of follow-up (adjusted odds ratio, 3.5; 95% CI, 1.3 to 9.2) than did the remaining 41 patients.

As shown by Ashton et al ⁽¹⁵⁾ and many others, however, patients who require vascular surgery appear to have an increased risk for cardiac complications because:

• Many of the risk factors contributing to peripheral vascular disease (e.g., diabetes mellitus, tobacco use, hyperlipidemia) are also risk factors for CAD.
• The usual symptomatic presentation for CAD in these patients may be obscured by exercise limitations imposed by advanced age or intermittent claudication, or both.
• Major arterial operations often are time-consuming and may be associated with substantial fluctuations in intra-extravascular fluid volumes, cardiac filling pressures, systemic blood pressure, heart rate, and thrombogenicity ⁽¹⁾.

Several Two recent studies have attempted to stratify the incidence of perioperative and intermediate-term MI according to the original type of vascular surgery performed. In a prospective series of 53 aortic procedures and 87 infrainguinal bypass grafts for which operative mortality rates were nearly identical (9% and 7%, respectively), Krupski et al ⁽⁸⁵⁾ found that the risk for fatal/nonfatal MI within a 2-year follow-up period was 3.5 times higher (21% vs. 6%) among patients who received infrainguinal bypass grafts. This difference probably is related to the fact that diabetes mellitus (44% vs. 11%) and history of previous MI (43% vs. 28%), angina (36% vs. 15%), or HF (29% vs. 9%) also were significantly more prevalent in the infrainguinal bypass group. L'Italien et al ⁽⁸⁶⁾ have presented comparable data regarding the perioperative incidence of fatal/nonfatal MI and the 4-year event-free survival rate after 321 aortic procedures, 177 infrainguinal bypass grafts, and 49 carotid endarterectomies. Slight differences in the overall incidence of MI among the three surgical groups, which may have been related to the prevalence of diabetes mellitus, were exceeded almost entirely in significance by the influence of discrete cardiac risk factors (previous MI, angina, HF, fixed or reversible thallium defects, and ST-T depression during stress testing) ⁽⁸⁶⁾. These and other studies ⁽⁵⁾ suggest that the clinical evidence of CAD in a patient who has peripheral vascular disease appears to be a better predictor of subsequent cardiac events than the particular type of peripheral vascular operation to be performed.

In a selective review of several thousand vascular surgical procedures (carotid endarterectomy, aortic aneurysm resection, and lower-extremity revascularization) reported in the English literature from 1970 to 1987, Hertzer ⁽⁶⁾ found that cardiac complications were responsible for about half of all perioperative deaths and that fatal events were nearly five times more likely to occur in the presence of standard preoperative indications of CAD. Furthermore, the late (5-year) mortality rate for patients who were suspected to have CAD was twice that for patients who were not (approximately 40% vs. 20%). It is noteworthy that both the perioperative and 5-year mortality rates for the small groups of patients who previously had coronary bypass surgery were similar to the results reported for larger series of patients who had no clinical indications of CAD at the time of peripheral vascular surgery.

In a study based on the 24,959 participants with known CAD in the Coronary Artery Surgery Study (CASS) database, Eagle et al found that the cardiac risk associated with noncardiac operations involving the thorax, abdomen, vasculature, and head and neck was reduced significantly in those patients who had undergone prior coronary artery bypass graft (CABG) (postoperative deaths 1.7% vs. 3.3%, MI's 0.8% vs. 2.7%) ⁽²⁶⁰⁾. In a recent randomized, multicenter trial, Poldermans et al documented the cardioprotective effect of perioperative beta-blockade in substantially and significantly reducing the cardiac morbidity and mortality in high-risk patients undergoing major vascular surgery ⁽²⁵²⁾.

Published mortality rates from large referral centers may not reflect the results at thousands of other hospitals throughout the United States in which, collectively, most vascular surgeries are performed on an individual, low-volume basis. Hsia et al ⁽⁸⁷⁾ have calculated that fewer than 10 carotid endarterectomies were performed annually at 45% of all hospitals in which Medicare beneficiaries received this procedure from 1985 to 1989, and Fisher et al ⁽⁸⁸⁾ demonstrated that the perioperative mortality rate (1.1% to 3.2%) had an inverse relation to the low volume of carotid endarterectomies in 2,089 Medicare patients at 139 New England hospitals. Similar trends (high volume/low risk, low volume/high risk) have been confirmed by statewide audits of aortic aneurysm resection in Vermont, Kentucky, and New York ^(89-91). In New York, for example, Hannan et al ⁽⁹¹⁾ reviewed 3570 elective aneurysm resections from 1985-1987 and found a linear, inverse relation between case volume and mortality rates for surgeons who annually performed two or fewer operations (11% mortality), three to nine operations (7.3% mortality), or 10 or more operations (5.6% mortality). No comparable data are available for lower-extremity bypass procedures, but according to the National Center for Health Statistics, the potential magnitude of this problem is illustrated by the fact that each year approximately 100,000 patients are discharged from U.S. hospitals after lower-extremity revascularization ⁽⁹²⁾.

Chassin et al ⁽⁹³⁾ collected 1984 data for the 30 most common diagnosis-related groups for which charges were submitted from nearly 5,000,000 admissions to over 5,000 hospitals. Of 48 homogeneous medical and surgical conditions developed from a statistical model, only four had adjusted mortality rates that clearly could be correlated from one condition to another; three (carotid endarterectomy, aortic reconstruction, and lower-extremity revascularization) involved vascular surgery, and the fourth (total hip replacement), orthopedic surgery. Thus, if a hospital did well or poorly with one of these operations, it tended to do equally well or poorly with the rest of them. Considering the fact that the prevalence of CAD contributes substantially to the perioperative risk of vascular surgery, at least some of the differences in surgical outcome from one hospital to another may be accounted for by variations in the degree to which it is recognized and appropriately treated. The level of this awareness also has implications regarding survival. In the prospectively randomized Veterans Administration trial of carotid endarterectomy vs. nonoperative management for asymptomatic carotid stenosis, for example, more than 20% of both randomized cohorts died of cardiac-related complications within a follow-up period of 4 years ⁽⁹⁴⁾.

Fleisher et al analyzed a 5% sample of Medicare claims from 1992 to 1993 of patients undergoing major vascular surgery. A total cohort of 2865 individuals underwent aortic surgery with a 7.3% 30-day mortality rate and a 11.3% one-year mortality rate. A total cohort of 4030 individuals underwent infrainguinal surgery with a 5.8% 30-day mortality rate and 16.3% one-year mortality rate. This work further confirms that aortic and infrainguinal surgery continues to be associated with high 30-day and one-year mortality, with aortic surgery being associated with the highest short-term and infrainguinal surgery being associated with the highest long-term mortality rates ⁽²⁶¹⁾.

Patients undergoing major vascular surgery constitute a particular challenge (i.e., high-risk operations in a patient population with a high prevalence of significant CAD). There are, however, other surgical procedures for which the interaction of patient-specific and surgery-specific factors have been examined. Transplantation surgery generally represents a high-risk procedure in a patient with multiple comorbidities. Significant CAD is common in diabetic patients with end-stage renal disease. In a study of 176 consecutive patients undergoing either kidney or kidney-pancreas transplants, there was a high correlation between adverse postoperative cardiac events and preoperative documentation of reversible defects on intravenous dipyridamole thallium-201 myocardial imaging in combination with significant CAD on coronary angiograms: 3 (11.1%) of 27 vs. 1 (0.9%) of 111 patients with a normal dipyridamole thallium-201 scan ⁽²⁶²⁾.

Although the prevalence of CAD is relatively low in patients with end-stage liver disease undergoing liver transplantation, 2 studies ^(263,264) have documented the reliability of dobutamine stress echocardiography in predicting posttransplant cardiac events. Stress echocardiography has also been shown to be useful in predicting cardiac outcomes in patients with advanced obstructive pulmonary disease undergoing lung volume reduction surgery ^(265,266).

As Fleisher and Barash ⁽⁹⁵⁾ have emphasized, the specific surgical setting must be considered within any algorithm regarding preoperative cardiac evaluation. The term noncardiac operation is exceedingly broad in its definition; it embraces aging patients with complex technical problems as well as younger patients scheduled for straightforward surgical procedures. As described above, cardiovascular morbidity and mortality vary not only among procedures but also among institutions for the same procedure. Therefore, in assessing the risks and benefits of perioperative intervention strategy, risks associated with noncardiac surgery must be individualized. It is important to remember, however that the indications for coronary intervention should not be redefined simply because a patient who has CAD of marginal significance also happens to require a major noncardiac procedure. Conversely, the long-term implications of severe left main or triple-vessel disease and diminished left ventricular function are no less ominous after a minor noncardiac operation than they are in any other patient situation. In the final analysis, one of the ultimate objectives of the preoperative cardiac assessment is to exclude the presence of such serious CAD that some form of direct intervention would be warranted even if no noncardiac operation were necessary. In this regard, the presentation for noncardiac surgery may simply represent the first time that a patient with overt or suspected CHD has had an opportunity for cardiovascular assessment.

In summary, the surgical procedures have been classified as low, intermediate, and high risk as shown in Table 3. Although coronary disease is the overwhelming risk factor for perioperative morbidity, procedures of different levels of stress are associated with different levels of morbidity and mortality. Superficial and ophthalmologic procedures represent the lowest risk and are rarely associated with excess morbidity and mortality. Major vascular procedures represent the highest-risk procedures. Within the intermediate-risk category, morbidity and mortality vary, depending on the surgical location and extent of the procedure. Some procedures may be short, with minimal fluid shifts, while others may be associated with prolonged duration, large fluid shifts, and greater potential for postoperative myocardial ischemia and respiratory depression. Therefore, the physician must exercise judgment to correctly assess perioperative surgical risks and the need for further evaluation.