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.

IX. Perioperative Surveillance

Although much attention has been focused on the preoperative preparation of the high-risk patient, intraoperative and postoperative surveillance for myocardial ischemia and infarction, arrhythmias, and venous thrombosis should also lead to reductions in morbidity. Postoperative myocardial ischemia has been shown to be the strongest predictor of perioperative cardiac morbidity and is rarely accompanied by pain ⁽¹⁾. Therefore, it may go untreated until overt symptoms of cardiac failure develop.

The diagnosis of a perioperative MI has both short- and long-term prognostic value. Traditionally, a perioperative MI has been associated with a 30% to 50% perioperative mortality and has been associated with reduced long-term survival ^{(19,29,214,215)}. Therefore, it is important to identify patients who sustain a perioperative MI and to treat them aggressively since it may reduce both short- and long-term risk.

A. Intraoperative and Postoperative Use of Pulmonary Artery Catheters
1. General Considerations
The pulmonary artery catheter can provide significant information critical to the care of the cardiac patient. Its use, however, must be balanced against the cost and risk of complications from insertion and use of the catheter, which are low when the operators are experienced. Several studies have evaluated the benefit of pulmonary artery catheters in both randomized trials and those using historical controls. In patients with prior MI, when perioperative care included pulmonary artery and intensive care monitoring for 3 days postoperatively, there was a lower incidence of reinfarction than in historical controls ⁽²⁹⁾. Other changes in management occurred during the period under study, however, including the increased use of beta-adrenergic sympathetic blockade. In particular, patients with signs and symptoms of HF preoperatively, who have a very high (35%) postoperative incidence of HF, might benefit from invasive hemodynamic monitoring ⁽⁶⁷⁾.

2. Summary of Evidence
Although a great deal of literature has evaluated the utility of a pulmonary artery catheter during the perioperative period in noncardiac surgery, relatively few only six controlled studies have evaluated pulmonary artery catheterization in relation to clinical outcomes. Three recent rRandomized trials have evaluated the routine use of pulmonary artery catheters vs. central venous pressure catheters or selective use of monitoring in abdominal aortic surgery and in elective vascular surgery. In studies using appropriate patient selection, no differences in cardiac morbidity (MI, cardiac death) were detected ^{(216,217,319,335,336)}. An additional study demonstrated no difference in cardiac morbidity in infrainguinal surgery patients when monitored by a pulmonary artery catheter either from the evening before surgery, 3 hours before surgery, or only if clinically indicated; however, the groups with the pulmonary artery catheter had fewer intraoperative hemodynamic disorders ⁽¹⁹³⁾. Polanczyk et al performed a prospective cohort study of 4,059 patients aged 50 years or older who underwent major elective noncardiac procedures with an expected length of stay of two or more days ⁽³³⁷⁾. Major cardiac events occurred in 171 patients, and those who underwent perioperative pulmonary artery catheterization had a three-fold increased incidence of major postoperative cardiac events (34 [15.4%] vs. 137 [3.6%]; p less than 0.001). In a case-control analysis of a subset of 215 matched pairs of patients who did and did not undergo pulmonary artery catheterization, adjusted for propensity of pulmonary artery catheterization and type of procedure, patients who underwent perioperative pulmonary artery catheterization also had increased risk of postoperative congestive HF (odds ratio, 2.9; 95% CI, 1.4 to 6.2) and major noncardiac events (odds ratio, 2.2; 95% CI, 1.4 to 4.9) ⁽³³⁷⁾. Iberti et al demonstrated in a multicenter survey that physicians' understanding of pulmonary artery catheterization data is extremely variable, which may account for the higher rate of postoperative congestive HF and greater perioperative net fluid intake ⁽³³⁸⁾.

3. Recommendations
Current evidence does not support routine use of a pulmonary artery catheter perioperatively. Although evidence from controlled trials is scant and a large-scale cohort study demonstrated potential harm, the use of pulmonary artery catheters may benefit high-risk patients. This is in keeping with practice parameters for the intraoperative use of a pulmonary artery catheter have recently been published by the American Society of Anesthesiologists ⁽²¹⁸⁾. These parameters approach the decision to place the pulmonary artery catheter as the interrelationship among three variables: patient disease, surgical procedure, and practice setting. With regard to the surgical procedure, the extent of intraoperative and postoperative fluid shifts is a dominant factor. Physician education on the interpretation of the pulmonary artery catheterization data is critical to achieve optimal benefit without harm.

Recommendations for Intraoperative Use of Pulmonary Artery Catheters ⁽²¹⁸⁾
Class IClass IIa
1. Patients at risk for major hemodynamic disturbances that are most easily detected by a pulmonary artery catheter who are undergoing a procedure that is likely to cause these hemodynamic changes in a setting with experience in interpreting the results (e.g., suprarenal aortic aneurysm repair in a patient with angina).

Class IIb
1. Either the patient's condition or the surgical procedure (but not both) places the patient at risk for hemodynamic disturbances (e.g., total hip replacement in a patient with chronic renal insufficiencysupraceliac aortic aneurysm repair in a patient with a negative stress test).

Class III
No risk of hemodynamic disturbances.

B. Intraoperative and Postoperative Use of ST-Segment Monitoring
1. General Considerations
Some Many contemporary operating rooms and intensive care unit monitors incorporate algorithms that perform real-time analysis of the ST segment. In addition, real-time ST-segment monitoring via telemetry or ambulatory ECG (Holter) monitors with alarms is being developed. Numerous studies have demonstrated the limited poor ability of physicians to detect significant ST-segment changes compared with computerized or off-line analysis. If available, computerized ST-segment trending is superior to visual interpretation in the identification of ST-segment changes. Although proprietary, many of these algorithms have been validated for their ability to accurately detect ST-segment shifts. Because the algorithms used to measure ST-segment shifts are proprietary, variability in accuracy between the different monitors has been evaluated in several studies compared with off-line analysis of standard Holter recordings ^(339-341). ST-trending monitors were found to have an average sensitivity and specificity of 74% (range 60% to 78%) and 73% (range 69% to 89%), respectively ⁽³⁴⁰⁾. Several factors have been identified that decreased the accuracy of the monitors, which have been discussed in detail elsewhere. Additionally, the lead system used affects the incidence of ischemia detected, with leads II and V5 detecting only 80% of all episodes detected by 12-lead ECG ⁽³⁴²⁾.

2. Summary of Evidence
Virtually all studies examining the predictive value of intraoperative and postoperative ST-segment changes have been performed using ambulatory ECG recorders. Using retrospective analysis, investigators have found postoperative ST-segment changes indicative of myocardial ischemia to be the strongest an independent predictor of perioperative cardiac events in high-risk noncardiac surgery patients in multiple studies, with changes of prolonged duration being particularly associated with increased risk ^{(19,51,219,220)}. Additionally, postoperative ST-segment changes have been shown to predict worse long-term survival in high-risk patients ⁽²¹⁴⁾.

In patients at moderate risk for CAD (age less than 45 years without known CAD and only one risk factor), the presence of intraoperative and postoperative ST-segment changes was not associated with either ischemia on an exercise stress test or cardiac events within 1 year ⁽³⁴³⁾. The total cohort of patients was small, which may limit generalizability of these findings.

Intraoperative ST-segment changes may also occur in low-risk populations. ST-segment depression has been shown to occur during elective cesarean sections in healthy patients ^(221,344). Because these changes were not associated with regional wall-motion abnormalities on precordial echocardiography, in this low-risk population such ST-segment changes may not be indicative of myocardial ischemia and CAD.

Thus, although there are data to support the contention that ST-segment monitoring detects ischemia, no studies have addressed the issue of the effect on outcome when therapy is based on the results of ST-segment monitoring.

Recommendations for Perioperative ST-Segment Monitoring
Class IIa
When available, proper use of computerized ST-segment analysis in patients with known CAD or undergoing vascular surgery appropriate high-risk patients may provide increased sensitivity to detect myocardial ischemia during the perioperative period and may identify patients who might would benefit from further postoperative and long-term interventions.

Class IIb
Patients with single or multiple risk factors for CAD.

Class III
Patients at low risk for CAD. Therefore, computerized monitoring of the ST segment can provide useful information in appropriate high-risk patients, if available. The cost-effectiveness of computerized ST-segment analysis for reducing perioperative morbidity, however, has not bee documented. Further research is required.

C. Surveillance for Perioperative MI

Multiple studies have evaluated predictive factors for a perioperative MI. The presence of clinical evidence of coronary artery or peripheral vascular disease has been associated with an increased incidence of perioperative MI. Factors that increase the risk of a perioperative MI have been discussed previously. Because of the increased risk of both short- and long-term mortality from a perioperative MI, accurate diagnosis is important.

1. General Considerations
Perioperative MI can be documented by assessing clinical symptoms, serial electrocardiography, cardiac-specific biomarkersenzyme analyses, comparative ventriculographic studies before and after surgery, and radioisotopic studies specific for myocardial necrosis, and autopsy studies. The criteria used to diagnose infarction in various studies differ not only in the level of cardiac biomarkers that determine abnormality but also the frequency with which they are sampled following noncardiac surgery. The cardiac biomarker profile after infarction exhibits a typical rise and fall that differs among different biomarkers. Daily sampling may miss detection of a cardiac biomarker rise (such as MB isoenzyme of creatine kinase [CK-MB], thus underestimating the incidence of perioperative infarction. The ECG criteria used to define infarction may also differ not only in the definition of a Q wave but also with respect to the magnitude of ST-T wave shifts that determine an abnormal response. Published ECG reporting criteria for MI vary, however, with some authors requiring Minnesota code criteria, others simply reporting new Q waves, and others not specifying the ECG criteria. Nonspecific ST and T wave changes are common in the immediate postoperative period but have not been associated with increase morbidity ⁽²²²⁾. In the analysis of cardiac biomarker enzyme criteria, numerous assays techniques are available to measure CK-MB, cardiac troponin I, and to a lesser extent, cardiac troponin T.and the threshold at which an enzyme rise is considered abnormal is variable. CK-MB may be released from noncardiac sources in patients with ischemic limbs or those undergoing aortic surgery, In addition, CK-MB has been shown to be released from noncardiac sources in patients with ischemic limbs or undergoing aortic surgery, the group at highest risk for a perioperative MI. The use of cardiac troponin I or T offers the potential of enhanced specificity ^{(223,345-350)}.Finally, new myocardial-specific enzymes that are currently being developed and evaluated in the perioperative patient, such as cardiac troponin-I or troponin-T or CK-MB isoforms, require further study before determining their usefulness in this setting ⁽²²³.

2. Summary of Evidence
Very few studies have examined the long-term outcome using protocol-specific criteria for perioperative MI after noncardiac surgery.optimal protocol for diagnosing a perioperative MI. Charlson et al ⁽²²⁴⁾ reported on 232 mostly hypertensive or diabetic patients undergoing elective noncardiac surgery. Serial ECGs and CK-MB were collected for 6 days postoperatively. The incidence of perioperative MI varied greatly depending on the diagnostic criteria used. A strategy using an ECG immediately after the surgical procedure and on the first and second days postoperatively had the highest sensitivity. Strategies including the serial measurement of CK-MB had higher false-positive rates without higher sensitivities. In contrast, Rettke et al ⁽²²⁵⁾ reported that overall survival was associated with the degree of CK-MB elevation in 348 patients undergoing abdominal aortic aneurysm repair, with higher levels associated with worse survival. Yeager et al ⁽²¹⁵⁾ evaluated the prognostic implications of a perioperative MI in a series of 1,561 major vascular procedures. These authors found that the incidence of subsequent MI and coronary artery revascularization was significantly higher in patients who suffered a perioperative MI, except in the subset who only demonstrated an elevated CK-MB without ECG changes or cardiovascular symptoms.

The use of cardiac troponin I to examine the diagnosis of perioperative MI was assessed in a series of 96 subjects undergoing vascular surgery and 12 undergoing spinal surgery. Blood samples were obtained every 6 hours for 36 hours postoperatively, and ECGs were acquired daily. The appearance of a new segmental wall-motion abnormality on a postoperative day 3 echocardiogram was used to diagnose perioperative infarction. All 8 patients who underwent vascular surgery and had segmental wall-motion abnormalities had elevated cardiac troponin I levels; 6 had elevated CK-MB. Of 100 patients without new segmental wall-motion abnormalities, 19 had CK-MB elevations; 1 had cardiac troponin I elevation ⁽²²²⁾. Several studies have examined cardiac troponin T as a marker for perioperative necrosis after noncardiac surgery. Of 772 patients who underwent major noncardiac procedures without major cardiovascular complications during the index hospital admission, 12% and 27%, respectively, had elevated cardiac troponin T and CK-MB values. During six-month follow-up, 19 subjects had major cardiac complications (14 cardiac deaths, 3 nonfatal MIs, and 2 admissions for unstable angina). The relative risk of cardiac events was 5.4 when cardiac troponin T was elevated, whereas CK-MB did not predict late postdischarge cardiac events ⁽³⁴⁹⁾. In another report ⁽³⁴⁶⁾, the diagnosis of perioperative MI was defined prospectively as total CK-MB greater than 174 units per liter and 2 of the following: 1) CK-2/CK (mass or activity) greater than 5%, 2) Q waves greater than 40 ms and 1 mm deep in 2 contiguous leads, 3) troponin T greater than 0.2 mcg per liter, or 4) a positive pyrophosphate scan. Of 323 patients undergoing noncardiac surgery (13.6% vascular), 18 (5.6%) had a perioperative MI. The incident rate of perioperative MI was 5.3% when the diagnosis included autopsy data, new Q waves, or CK-2 elevation greater than 5% of total CK associated with new ECG changes. The incidence increased to 11.2% when the definition included autopsy data, new Q waves, cardiac troponin T greater than 0.2 mcg per liter, and ECG changes. The MI rate increased to 20.7% when the definition of perioperative MI included autopsy data, new Q waves, or cardiac troponin T greater than 0.2 mcg per liter.

3. Recommendations
Further evaluation regarding the optimal strategy for surveillance and diagnosis of perioperative MI is required. On the basis of current evidence, before one method is advocated. Iin patients without documented evidence of CAD, surveillance should be restricted to patients who develop perioperative signs of cardiovascular dysfunction. In patients with high or intermediate clinical risk who have known or suspected CAD and who are undergoing high- or intermediate-risk surgical procedures associated with a high incidence of cardiovascular morbidity, the procurement of ECGs at baseline, immediately after the surgical procedure, and daily on the first 2 days after surgery appears to be the most cost-effective strategy. Measurements of cardiac enzymes are best reserved for patients at high risk or those who demonstrate ECG or hemodynamic evidence of cardiovascular dysfunction. Cardiac troponin measurements 24 hours postoperatively and on day 4 or hospital discharge (whichever comes first) should be part of the diagnostic strategy for perioperative MI detection ⁽³⁵⁰⁾. The majority of perioperative MI events will be non-Q wave. Additional research is needed to correlate long-term outcome results to magnitude of isolated cardiac troponin elevations. The diagnosis of MI should be entertained when the typical cardiac biomarker profile is manifest in the immediate postoperative phase. A risk gradient can be based on the magnitude of biomarker elevation and presence or absence of concomitant new ECG abnormalities, hemodynamic instability, and quality and intensity of chest pain syndrome, if present. The ACC and the European Society of Cardiology have provided a redefinition of acute MI based on studies examining cardiac troponins and clinical presentation/outcomes ⁽³⁵¹⁾. Patients who sustain a perioperative MI should have evaluation of left ventricular function performed before hospital discharge, and standard postinfarction therapeutic medical therapy should be prescribed as defined in the ACC/AHA Acute Myocardial Infarction guidelines ⁽³⁷⁰⁾. Perioperative surveillance for acute coronary syndromes using routine ECG and cardiac serum biomarkers is unnecessary in clinically low-risk patients undergoing low-risk operative procedures.

D. Arrhythmia/Conduction Disease Disorders

Postoperative arrhythmias are often due to remedial noncardiac problems such as infection, hypotension, metabolic derangements, and hypoxia. The approach taken to the acute management of postoperative tachycardias varies depending on the likely mechanism. If the patient develops a sustained regular narrow-complex tachycardia, which is likely due to atrioventricular nodal re-entrant tachycardia or atrioventricular reciprocating tachycardia, the tachycardia can almost always be terminated with vagal maneuvers (Valsalva maneuver or carotid sinus massage) or with intravenous adenosine. Most antiarrhythmic agents (especially beta blockers, calcium channel blockers, and type 1a or 1c antiarrhythmic agents) can be used to prevent further recurrences in the postoperative setting. A somewhat different approach is generally recommended for atrial fibrillation and atrial flutter. The initial approach to management generally involves the use of intravenous digoxin, diltiazem or a beta blocker in an attempt to slow the ventricular response. Among these three types of medications, digitalis is least effective and beta blockers most effective for controlling the ventricular response during atrial fibrillation ⁽³¹³⁾. An additional benefit of beta blockers is that they have been shown to accelerate the conversion of postoperative supraventricular arrhythmias to sinus compared with diltiazem ⁽³¹⁴⁾. Cardioversion of atrial fibrillation/flutter is generally not recommended for asymptomatic or minimally symptomatic arrhythmias until correction of the underlying problems has occurred, which frequently leads to a return to normal sinus rhythm. Also, cardioversion is unlikely to result in long-term normal sinus rhythm if the underlying problem is not corrected. The avoidance of an electrolyte abnormality, especially hypokalemia and hypomagnesemia, may reduce the perioperative incidence and risk of arrhythmias, although acute preoperative repletion of potassium in an asymptomatic individual may be associated with greater risk than benefits ^{(226-228,352)}. Unifocal or multifocal premature ventricular contractions do not merit therapy. Very frequent ventricular ectopy or prolonged runs of nonsustained ventricular tachycardia may require antiarrhythmic therapy if they are symptomatic or result in hemodynamic compromise. Patients with an ischemic cardiomyopathy who have nonsustained ventricular tachycardia in the perioperative period may benefit from referral for electrophysiologic testing to determine the need for an ICD ^(353,354). Ventricular arrhythmias may respond to intravenous beta blockers, lidocaine, procainamide, or amiodarone ^{(186,355-357)}. Electrical cardioversion should be used for sustained supraventricular or ventricular arrhythmias that cause hemodynamic compromise.

Bradyarrhythmias that occur in the postoperative period are usually secondary to some other cause, such as certain medications, an electrolyte disturbance, hypoxemia, or ischemia. On an acute basis, many will respond to intravenous medication such as atropine, and some will respond to intravenous aminophylline. Bradyarrhythmias due to sinus node dysfunction and advanced conduction abnormalities such as complete heart block will respond to temporary or permanent transvenous pacing or permanent pacing. The indications are the same as those for elective permanent pacemaker implantations.

Cardioversion of supraventricular arrhythmias is generally not recommended until correction of the underlying problems has occurred, which frequently leads to a return to normal sinus rhythm. Also, cardioversion is unlikely to result in long-term normal sinus rhythm if the underlying problem is not corrected. The avoidance of an electrolyte abnormality, especially hypokalemia and hypomagnesemia, may reduce the perioperative incidence and risk of arrhythmias, although acute preoperative repletion of potassium in an asymptomatic individual may be associated with greater risk than benefits (226-228). Bradarrhythmias occurring in the postoperative period are usually secondary to some other cause, such as an electrolyte disturbance, hypoxemia, or ischemia. On an acute basis, many will respond to intravenous medication such as atropine, and some will respond to intravenous aminophylline. Those bradarrhythmias due to sinus node dysfunction and advanced conduction abnormalities such as complete heart block will respond to temporary or permanent transvenous pacing or permanent pacing with indications the same as those for elective pacemaker implantations.

Supraventricular arrhythmias may respond to this digitalis, calcium channel blockers, or blockers with slowing of heart rate and cardioversion. Unifocal or multifocal premature ventricular contractions do not merit vigorous therapy. Complex ventricular ectopy such as nonsustained or sustained ventricular tachycardia requires more vigorous therapy, especially in the presence of ongoing or threatened myocardial ischemia, left ventricular dysfunction, or valvular heart disease. Ventricular arrhythmias may respond to intravenous blockers, lidocaine, procainamide, or amiodarone (186). Electrical cardioversion should be used for supraventricular or ventricular arrhythmias causing hemodynamic compromise.