Patients with Positive Preoperative Stress Tests Undergoing Vascular Surgery

Patients With Positive Preoperative Stress Tests Undergoing Vascular Surgery

Kyung W. Tim Park, MD, Kathirvel Subramaniam, MD, Feroze Mahmood, MD, Fred Shapiro, MD, Selina Long, MD, and David Napoli, MD

Objective: To examine the perioperative cardiac morbidity Cardiology and 7 had a left ventricular ejection fraction and mortality in patients undergoing major vascular surgery <40%. Twenty-three patients had been on a ␤-blocker and with ␤-blockade after a positive stress test or cardiac cath- continued on it, while the remainder started on it de novo eterization. perioperatively. None of the patients suffered from myocar- Design: Retrospective review of a quality assurance data- dial infarction, congestive heart failure, or cardiac death base. perioperatively. Setting: A university teaching hospital. Conclusions: This case series reports on the authors’ ex- Participants: A consecutive series of 31 patients undergo- perience with patients undergoing high-risk vascular sur- ing peripheral vascular or aortic surgery after a positive gery after a positive stress test or catheterization, but with- stress test or catheterization between November 2001 and out an intervening coronary intervention. All patients received perioperative ␤-blockade and had a very low ad- September 2003. verse cardiac event rate. With reduction of adverse events Intervention: None. by ␤-blockade, the likelihood of a positive event may be Measurements and Main Results: All 31 patients had a reduced and the utility of the test in risk stratification may preoperative positive stress test and/or cardiac catheteriza- be questioned. tion, with 12 having multiple areas at risk for myocardial © 2005 Elsevier Inc. All rights reserved. ischemia. None had an intervening coronary revascularization. Twenty-seven had at least one of the intermediate KEY WORDS: vascular surgery, stress test, risk stratification, clinical predictors as defined by the American College of ␤-blockade

HE AMERICAN COLLEGE of Cardiology (ACC)/Amer- operative ␤-blockade with bisoprolol or placebo and showed Tican Heart Association (AHA) guidelines on preoperative that even in such high-risk patients perioperative ␤-blockade cardiac evaluation recommend that for patients undergoing reduced 30-day mortality from 17% to 3.4% and nonfatal MI high-risk surgery such as aortic or peripheral vascular bypass from 17% to 0%. Such a study, especially if reproduced, may surgeries, a noninvasive cardiac test be performed, if the patient provide justification for proceeding with high-risk noncardiac has a major clinical predictor such as unstable angina or evolv- surgery despite the known presence of coronary artery disease ing myocardial infarction (MI) or if the patient has either an (CAD), without surgical or percutaneous coronary revascular- 3 intermediate clinical predictor (eg, history of MI or congestive ization. In this report, the authors’ experience with vascular heart failure [CHF]) or poor functional status.1 If the result of surgical patients who were known to have CAD either by a such noninvasive cardiac testing is positive, it is implied that preoperative stress test or cardiac catheterization and then un- further cardiac workup and/or treatment may then be indicated derwent vascular surgery without an intervening surgical or percutaneous coronary intervention is presented. The data pre- before the proposed noncardiac surgery. However, at present, sented are based on screening of the authors’ quality assurance there are no clear data from randomized trials whether such database for identification of such patients (all of whom were patients should be triaged to myocardial revascularization be- anesthetized by the authors), review of the anesthetic records, fore the noncardiac surgery or an aggressive perioperative nursing records in the recovery room and on the ward, and regimen of ␤-adrenergic blockade during noncardiac surgery medical records. This experience mirrors the study result of without prior revascularization. Poldermans et al2 and may further bolster triage of such pa- 2 Poldermans et al studied vascular surgical patients with new tients to surgery without coronary intervention. reversible wall motion abnormalities on preoperative dobutamine stress echocardiography (DSE) but with normal left CASE SERIES ventricular function. The authors excluded patients with exten- sive wall motion abnormalities, asthma, or stress testing sug- Between November 2001 and September 2003, the authors have provided anesthetic care for 31 vascular surgical patients, having either gestive of left main or 3-vessel coronary artery disease. With- aortic or peripheral vascular bypass surgery, who preoperatively had out triaging the patients to further diagnostic workup or positive noninvasive or invasive cardiac testing and did not have an treatment, the authors randomized the patients to either peri- intervening surgical or percutaneous coronary revascularization. This number represents all such patients in the quality assurance database, which captures all patients undergoing vascular surgery at the authors’ From the Beth Israel Deaconess Medical Center, Harvard Medical institution. The decision on whether to do a stress test and whether to School, Boston, MA. treat the patient medically rather than interventionally after a positive Presented in part at the 2003 annual meeting of the International stress test was up to each patient’s cardiologist. The stress test took Anesthesia Research Society, New Orleans, LA, March 21-25, 2003. place 68 Ϯ 117 (median 13) days before the surgery. Demographic and Address reprint requests to Kyung W. Tim Park, MD, Department of comorbidity profile of the patients is shown in Table 1. The types of Anesthesiology, Santa Clara Valley Medical Center, 751 S Bascom surgeries undertaken are shown in Table 2. Avenue, San Jose, CA 95128. E-mail: [email protected]. Twenty-seven of the 31 patients had at least one of the intermediate ca.us clinical predictors, as defined by the ACC/AHA1 (Table 1). Fifteen of © 2005 Elsevier Inc. All rights reserved. the patients had multiple intermediate clinical predictors. History of 1053-0770/05/1904-0014$30.00/0 angina, documented in only 4 of 31 patients, might have been under- doi:10.1053/j.jvca.2005.05.008 documented because most of the patients had diabetes and had limited

494 Journal of Cardiothoracic and Vascular Anesthesia, Vol 19, No 4 (August), 2005: pp 494-498 PREOPERATIVE STRESS TESTS 495

Table 1. Demographic and Comorbidity Proﬁle of the Patients

Comorbidity or Demographic Variable N

Sex (male:female) 20:11 Age (mean Ϯ SD) 72 Ϯ 12 Ն65 24/31 Ն75 13/31 Intermediate clinical predictors Diabetes mellitus 17/31 History of myocardial infarction 15/31 History of angina 4/31 History of congestive heart failure 6/31 Serum creatinine Ն2 mg/dL 5/31 At least 1 of the intermediate clinical predictors 27/31 Two of the intermediate clinical predictors 10/31 Three or more of the intermediate clinical predictors 5/31 Preoperative medications ␤-adrenergic blocker 23/31 Statin 23/31 Angiotensin-converting enzyme inhibitor 20/31 Fig 1. Flow diagram depicting the preoperative workup of the patients. Calcium channel blocker 8/31 Oral nitrate 7/31 Aspirin 13/31 Other comorbidities test with thallium nuclear imaging to detect perfusion defects. In 3 Hypertension 25/31 patients, a positive stress test was followed by a cardiac catheterization, Hypercholesterolemia 22/31 which revealed diffuse 3-vessel coronary artery disease in 1 patient, History of smoking 18/31 occlusion of a previously grafted artery in the second patient, and a Cerebrovascular accident or transient ischemic 10/31 minimal 1-vessel disease (40% obstruction of the left circumflex artery) attacks in the third. The patient with the 3-vessel disease was the only one Asthma or chronic obstructive pulmonary disease 5/31 referred to cardiac surgery but was turned down for coronary artery surgery because of poor risk profile and yet presented for vascular surgery. All patients were thus managed medically preoperatively, without any surgical or percutaneous coronary intervention. A sche- activity from peripheral vascular disease. Functional status could not be matic of the preoperative workup is shown in Figure 1. Results of the easily assessed in most patients because of the limited activity level stress tests are summarized in Table 3. Twelve of the 31 patients had from peripheral vascular insufficiency. In the 13 patients whose func- multiple areas at risk for myocardial ischemia. The inferior (16/31) and tional status could be determined, the New York Heart Association anterior walls (14/31) were the most commonly affected. Seven of 31 status averaged 2.4, with a median of 2. In the 4 patients without any patients had a left ventricular ejection fraction Ͻ40%. of the intermediate clinical predictors, activity level was severely Twenty-three of the 31 patients had been on a ␤-adrenergic blocking limited by peripheral vascular disease or severe degenerative arthritis. agent and continued on the medication. The remainder were started on The incidence of other commonly examined comorbidities is listed in a ␤-adrenergic blocking agent de novo in the perioperative period. For Table 1. In 3 patients, cardiac catheterization was performed without a preceding stress test and showed 3-vessel disease in 1 patient and occlu- Table 3. Results of Preoperative Cardiac Workup sion of a previously revascularized coronary artery in the other 2 patients; in all 3, no surgical or percutaneous coronary intervention was N done and the patients were elected to be treated medically. Because of Left ventricular ejection fraction the peripheral vascular disease, all but 4 of the remaining 28 patients Ͻ40 % 7/31 had pharmacologic stress testing (dipyridamole, adenosine, or dobut- Ն40 % 18/31 amine) with various nuclear-imaging modalities or echocardiography Not reported 6/31 to detect regions of reversible ischemia. The remaining 4 had an Areas at risk* exercise stress test with thallium imaging. A large majority (21/24) of On the stress test (n ϭ 28) those who had a pharmacologic stress test had a dipyridamole stress One 17/28 Two 5/28 Three or more 6/28 Table 2. Types of Surgeries Undertaken On cardiac catheterization after the stress test (3/28) Type Number One 2/3 Aortic Three 1/3 Open repair of infrarenal aortic aneurysm 3 On cardiac catheterization without preceding Endovascular repair of infrarenal aortic aneurysm 3 stress tests (n ϭ 3) Infrainguinal 23 One 2/3 Other Three 1/3 Right renal artery bypass 1 *Areas at risk are areas with reversible ischemia on a stress test or Right axillofemoral bypass 1 areas with an occluded coronary artery on cardiac catheterization. 496 PARK ET AL these patients, metoprolol was started a mean of 2 days (range 0-6 days, median 0) before surgery and was increased up to 50 mg twice a day as tolerated to bring the heart rate down to 70 beats/min. On the day of surgery, the patient’s usual dose of the ␤-adrenergic blocking agent was continued. All patients were monitored with an arterial catheter and a pulmonary artery catheter (30/31) or a central venous catheter (1/31), in addition to the usual American Society of Anesthesiologists’ standard monitors. In 5 patients, transesophageal echocardiography was used as well. All patients were given general anesthesia, although in 4 of the 6 patients having aortic surgery, it was supplemented with thoracic epidural analgesia with a catheter placed between T7-8 and T11-12. Thoracic epidural analgesia, when used, was started toward the end of the case, with a mixture of 0.1% bupivacaine and 10 ␮g/mL of hydromorphone running at 6 to 8 mL/h. For the induction of anesthesia, the patient was given 100 to 150 ␮g of fentanyl, followed by metoprolol in 1-mg increments as needed to bring the heart rate Յ70 beats/min, and thiopental, propofol, or etomidate in doses just sufficient to obliterate the eyelid reflex. These amounted to 50 to 200 mg of thiopental, 30 to 100 mg of propofol, or 0.5 to 1.5 mg/kg of etomidate in the present patients. Endotracheal intubation was facilitated with succinylcholine, 1 to 1.5 mg/kg, or vecuronium, 0.1 mg/kg. Muscle relaxation was maintained with additional doses of vecuronium as needed; muscle relaxants such as pan- curonium with vagolytic properties were avoided. Anesthesia was maintained with nitrous oxide and a choice of potent inhalation anesthetic (usually isoflurane) and 25 to 50 ␮g increments of fentanyl as Fig 2. Change in heart rates, expressed as mean ؎ standard needed. Additional doses of metoprolol were given as needed during deviation. PACU, postanesthesia care unit; POD, postoperative day. Although the heart rates averaged consistently below 70 beats/min the case to maintain the heart rate Յ70 beats/min. Hypertension was intraoperatively, they were in the 70 to 80 beats/min range in the treated with deepening the anesthetic level or nitroglycerin. Hypoten- postoperative period. sion was treated with adjusting the anesthetic level, fluids, and/or a vasopressor such as phenylephrine. Two patients required dobutamine infusion for low cardiac index associated with low mixed venous oxygen saturation (Ͻ50%-55%). Inotropic agents were otherwise diagnosis of CHF was confirmed by signs of pulmonary edema on chest avoided. At the end of the case, the muscle relaxant was reversed with radiograph in conjunction with the appropriate clinical symptoms/signs neostigmine (up to 70 ␮g/kg) and glycopyrrolate (up to 7-10 ␮g/kg). such as orthopnea and pulmonary rales. This represented one half to two thirds of the usual dose of glycopyr- Perioperatively, none of the patients suffered from MI, CHF, or rolate in an effort to avoid tachycardia on emergence. Tachycardia cardiac death. One patient, who preoperatively had a history of non- and/or hypertension on emergence was treated with incremental doses sustained ventricular tachycardia, had a recurrence of the same problem of metoprolol (1 mg at a time) or labetalol (5 mg at a time) to maintain postoperatively and was given an automatic implantable cardioverter the heart rate Յ70 beats/min with an acceptable perfusion pressure. All defibrillator. Another patient developed angina pectoris and ECG patients were extubated either in the operating room or shortly after changes suggestive of ischemia while she was getting hemodialyzed for leaving the operating room. The patients requiring inotropic support her preexisting end-stage renal disease on the third postoperative day. with dobutamine spent a day in the intensive care unit, until they were Cardiac enzymes were negative; but in view of her preoperative pos- weaned from the medication. All others spent 1 to 2 days in the itive dipyridamole-thallium test, she was taken to cardiac catheterization. She was found to have 80% occlusion of the right coronary artery, step-down unit for hemodynamic monitoring, before being sent to the which was stented. There were no other significant cardiovascular regular ward. ␤-Adrenergic blockade was continued postoperatively in events. There were no heart blocks, clinically significant bradycardia, an effort to maintain the heart rate Յ70 beats/min. The actual heart or bronchospastic episodes related to the use of ␤-adrenergic blockers. rates obtained both intraoperatively and postoperatively are depicted in Figure 2. Patients who had been chronically on a ␤-adrenergic blocking DISCUSSION agent were discharged on the same agent. Patients who were started de novo on metoprolol in the perioperative period were discharged on oral This case series reports on 31 patients who presented for metoprolol (dose range 50-150 mg daily in divided doses), except 1 either repair of an abdominal aortic aneurysm or peripheral patient who was discharged without a ␤-adrenergic blocker for unclear vascular bypass surgery between November 2001 and Septem- reasons. Duration of hospitalization after surgery was 6.5 Ϯ 4.8 days ber 2003, after a positive stress test or cardiac catheterization (median, 5 days). but without an intervening coronary revascularization. All pa- Postoperatively, an electrocardiogram (ECG) was obtained in all tients were given strict perioperative ␤-adrenergic blockade patients immediately after surgery. The ECG was initially read by the and had no adverse cardiac events, defined as MI, CHF, or anesthesia and surgical team and, if any new changes were found, cardiac death. confirmed by a cardiologist. New ECG changes suggestive of myocardial ischemia were defined as horizontal or downsloping ST-segment For 27 of 31 patients, there was a clear indication for a depression of at least 0.1 mV or ST elevation of at least 0.15 mV4 and preoperative stress test as dictated by the ACC/AHA guidelines led to 3 serial cardiac isoenzymes and ECGs every 8 hours. Postoper- because they were scheduled for a high-risk surgery and had at ative MI was defined by creatine kinase–MB fraction Ͼ5% and/or least one of the intermediate clinical predictors. For the other 4, troponin-I Ͼ2.0 ng/mL in the presence of appropriate ECG changes. A in the absence of an intermediate clinical predictor, poor func- PREOPERATIVE STRESS TESTS 497 tional status would have to be present to warrant a preoperative myocardial ischemia. In the study by Poldermans et al,2 the stress test. Their functional status could not be adequately study group patients were given titrated ␤-adrenergic blockers assessed because of severe degenerative arthritis or peripheral to maintain the heart rate Յ80 beats/min and their actual heart vascular disease and a pharmacologic stress test (dipyridamole- rate averaged about 70 beats/min. In the absence of information thallium imaging) was performed for each. Despite 11 of the on the ischemic threshold, the authors’ institutional approach stress tests suggesting multiple areas at risk, only 3 of the has been to maintain the heart rate Յ70 beats/min. When the positive tests were followed by cardiac catheterization. One of ischemic threshold is known, the authors aim to keep the heart the 3 revealed 3-vessel CAD and led to a cardiac surgical rate below it. consultation, but the patient was refused for cardiac surgery and The beneficial effect of perioperative ␤-adrenergic blockade presented to have his vascular surgery without coronary revas- has been shown in many previous studies,8-11 in addition to that cularization. of Poldermans et al.2 The effect appears to be a class effect The authors’ findings of a low cardiac complication rate in because the studies have been performed with various ␤- these high-risk patients are consistent with the report of Polder- adrenergic blockers including bisoprolol, atenolol, esmolol, mans et al2 on the benefit of perioperative ␤-adrenergic block- labetalol, oxprenolol, and metoprolol. The authors’ case series ade. Yet, the current patients differed from those of Poldermans adds to the mounting evidence for the benefit of perioperative et al in several respects. Although the patients of Poldermans et ␤-adrenergic blockade in those at high risk for perioperative al were all screened by DSE, most of the present patients cardiac complication. underwent a pharmacologic stress test with dipyridamole. The Reduction of perioperative cardiac complications with ␤- predictive value of a positive test on either DSE or a dipyrid- adrenergic blockade means that the post-test probability of amole test depends on the pretest cardiac risk of the patient an adverse cardiac event after a positive preoperative stress population tested. When DSE and dipyridamole-thallium test- test is decreased as well. The likelihood ratio of a positive ing are compared in the same population, they have comparable test, which indicates how much the odds of a cardiac event specificity and sensitivity.5 The incidence of perioperative MI increase when the test is positive, should be Ͼ10 in order for or cardiac death was about 9% for patients with a reversible a test to indicate a substantial change in risk from the pretest defect in 1 or more regions on dipyridamole-thallium testing in level and thus be considered useful in risk stratification.12 a meta-analysis.6 Second, the present series included patients Current data on the stress tests show that the likelihood ratio who might have been excluded in the study by Poldermans et of a positive test, whether the test is DSE or dipyridamole- al. Patients with multiple regions at risk for ischemia (12/31), thallium imaging, is somewhat less than 10.12 With reduc- with left ventricular ejection fraction Ͻ40% (7/31), with tion in the likelihood ratio of a positive stress test because of chronic obstructive pulmonary disease or asthma (5/31), or on perioperative use of ␤-adrenergic blockade, the usefulness a ␤-adrenergic blocker preoperatively (23/31) would have been and need of a preoperative stress test may need to be excluded in their study. In addition, the present series included questioned. a patient who was turned down for cardiac surgery. Finally, Boersma et al13 reanalyzed the patient population (N ϭ 1,351) many of the patients (15/31) had multiple intermediate clinical screened for the study of Poldermans et al.2 Important clinical predictors of the ACC/AHA guidelines. Lee et al7 identified 6 predictors of adverse cardiac outcome in this population were age variables predictive of cardiac risk in 2,893 patients undergoing Ն70 years, current or prior angina pectoris, history of MI, history major noncardiac surgeries and validated them in another co- of CHF, or history of cerebrovascular accident. With use of hort of 1,422 patients. The 6 variables were high-risk surgery, perioperative ␤-adrenergic blockade, a positive DSE had a signif- history of CAD without revascularization, history of CHF, icant prognostic value only among those with more than 3 clinical history of cerebrovascular disease, serum creatinine Ͼ2 mg/dL, predictors and even then only if the DSE showed Ն5 ischemic and the use of insulin. Rates of cardiac complications with 0, 1, segments. Among those with more than 3 clinical predictors, those 2, Ն3 of these factors were 0.5%, 1.3%, 4%, and 9%. In the with DSE showing 1 to 4 ischemic segments had a 2.8% cardiac present series of patients, 16 of 31 had Ն3 of the risk factors of complication rate, which is not significantly different from 2.0% Lee et al and 11 had 2 of the factors, with the remaining 4 complication rate of those with a negative DSE, when given having 1 risk factor. Clearly, despite some differences, these perioperative ␤-adrenergic blockade. Thus, with anticipated use of patients represented a high-risk group, like the patients of perioperative ␤-adrenergic blockade, a stress test such as DSE Poldermans et al. may be indicated if the patient has more than 3 clinical risk factors Nevertheless, in the current series, there were no periopera- such as identified by Boersma et al13 or Lee et al.7 Even then, a tive nonfatal MI, CHF, or cardiac death with strict adherence to positive result may have prognostic value only if it shows evi- perioperative ␤-adrenergic blockade. The authors’ regimen in- dence of diffuse CAD. cluded close monitoring of the patient’s heart rate, titration of The current patients were aggressively monitored with invasive metoprolol to try to maintain the rate Յ70 beats/min, and monitoring including a PAC in 30 of 31 patients. Recently, avoidance of any vagolytic or chronotropic medications as Sandham et al14 reported on a trial of the use of PAC in high-risk much as possible. There is no clear literature basis for choosing surgical patients and concluded that there was no benefit to its use. any specific target heart rate in the perioperative period. For In an attempt to standardize the treatment of the patients random- those patients in whom the ischemic threshold is known, Raby ized to care with a PAC, the authors imposed strict, supraphysi- et al8 advocated maintenance of the heart rate to 20% below the ologic treatment goals on the patients, such as oxygen delivery ischemic threshold but also noted that those who stayed under index of 550 to 600 mL/min/m2 and a cardiac index of 3.5 to 4.5 the ischemic threshold by any amount did not exhibit signs of L/min/m2, as long as the heart rate was kept less than 120 beats/ 498 PARK ET AL min. Most of the patients in the present study would not have met undertaken with low cardiac event rates with perioperative these goals, unless driven to potential myocardial ischemia with ␤-adrenergic blockade. Second, the authors’ practice of se- inotropic/chronotropic agents and fluid loading. The authors’ use lectively obtaining cardiac enzymes only in the presence of of the PAC did not involve these supraphysiologic targets but was ECG changes might have resulted in underestimation of the as a guide to judicious fluid therapy for the first 2 postoperative incidence of nonfatal MI in the patients. The 12-lead ECG days and then gentle diuresis as needed afterward. Even when the may be indeterminate in about 20% of patients with an acute cardiac index was low, the authors did not routinely resort to an MI.15 Although routine measurement of troponin might have inotropic agent, unless there was evidence of compromised oxy- picked up the missed cases, troponin may also be elevated gen delivery, such as low mixed venous saturation. nonspecifically in patients with renal insufficiency or sys- There are several limitations of this report in drawing temic inflammatory states16,17 and should be measured se- definitive conclusions regarding the care of high-risk vascu- lectively when clinically indicated. lar patients. First, this is merely a report of a series of cases, In summary, the authors’ experience with vascular surgi- not a randomized controlled trial or even a retrospective case cal patients who had positive preoperative stress tests and no match study. The gist of the report is not a comparison of intervening cardiac intervention is reported. With perioper- different treatment strategies, but that even in very-high-risk ative ␤-adrenergic blockade, patients had a very low cardiac patients with positive stress tests, high-risk surgery may be complication rate.

REFERENCES 1. Eagle KA, Berger PB, Calkins H, et al: ACC/AHA Guideline 9. Mangano DT, Layug EI, Wallace A, et al: Multicenter study of Update for Perioperative Cardiovascular Evaluation for Noncardiac perioperative ischemia research group. Effect of atenolol on mortality Surgery—Executive Summary: A report of the ACC/AHA task force and cardiovascular morbidity after noncardiac surgery. N Engl J Med on practice guidelines (Committee to Update the 1996 Guidelines on 335:1713-1720, 1996 Perioperative Cardiovascular Evaluation for Noncardiac Surgery). 10. Stone JG, Foex P, Sear JW, et al: Myocardial ischemia in J Am Coll Cardiol 39:542-553, 2002 untreated hypertensive patients: Effect of a single small oral dose of a 2. Poldermans D, Boersma E, Bax JJ, et al: Dutch Echocardio- beta-adrenergic blocking agent. Anesthesiology 68:495-500, 1988 graphic Cardiac Risk Evaluation Applying Stress Echocardiography 11. Urban MK, Markowitz SM, Gordon MA, et al: Postoperative Study Group. The effect of bisoprolol on perioperative mortality and prophylactic administration of ␤-adrenergic blockers in patients at risk myocardial infarction in high-risk patients undergoing vascular surgery. N Engl J Med 341:1789-1794, 1999 for myocardial ischemia. Anesth Analg 90:1257-1261, 2000 3. Park KW: Preoperative cardiology consultation. Anesthesiology 12. Grayburn PA, Hillis LD: Cardiac events in patients undergoing 98:754-762, 2003 noncardiac surgery: Shifting the paradigm from noninvasive risk strat- 4. Slogoff S, Keats AS: Does perioperative myocardial ischemia lead to ification to therapy. Ann Intern Med 138:506-511, 2003 postoperative myocardial infarction? Anesthesiology 62:107-114, 1985 13. Boersma E, Poldermans D, Bax JJ, et al, for the DECREASE 5. Kontos MC, Akosah KO, Brath LK, et al: Cardiac complications Study Group: Predictors of cardiac events after major vascular surgery: in noncardiac surgery: Value of dobutamine stress echocardiography Role of clinical characteristics, dobutamine echocardiography, and versus dipyridamole thallium imaging. J Cardiothorac Vasc Anesth beta-blocker therapy. JAMA 285:1865-1873, 2001 10:329-335, 1996 14. Sandham JD, Hull RD, Brant RF, et al: A randomized, con- 6. Shaw LJ, Eagle KA, Gersh BJ, et al: Meta-analysis of intravenous trolled trial of the use of pulmonary artery catheters in high-risk dipyridamole-thallium-201 imaging (1985-1994) and dobutamine surgical patients. N Engl J Med 348:5-14, 2003 echocardiography (1991-1994) for risk stratification before vascular 15. Mangano DT: Beyond CK-MB: Biochemical markers for peri- surgery. J Am Coll Cardiol 27:787-798, 1996 operative myocardial infarction. Anesthesiology 81:1317-1320, 1994 7. Lee TH, Marcantonio ER, Mangione CM, et al: Derivation and prospective validation of a simple index for prediction of cardiac risk 16. Freda BJ, Tang WHW, van Lente F, et al: Cardiac troponin in of major noncardiac surgery. Circulation 100:1043-1049, 1999 renal insufficiency. J Am Coll Cardiol 40:2065-2071, 2002 8. Raby KE, Brull SJ, Timimi F, et al: The effect of heart rate 17. Lucreziotti S, Foroni C, Fiorentini C: Perioperative myocardial control on myocardial ischemia among high-risk patients after vascular infarction in noncardiac surgery: The diagnostic and prognostic role of surgery. Anesth Analg 88:477-482, 1999 cardiac troponins. J Intern Med 252:11-20, 2002