INTRODUCTION — Cardiovascular complications pose one of the most significant risks to patients undergoing major noncardiac surgery. A prospective study published in 1977 evaluated 1001 such patients who were over age 40; the overall risk of postoperative cardiac death or major cardiac complications (including nonfatal myocardial infarction [MI], pulmonary edema, or ventricular tachycardia) was 5.8 percent [1].
A 1995 review of major published series found that the pooled average rates of the more selective outcome of MI and cardiac death varied with the population studied [2]:
- Among unselected patients over age 40 — MI in 1.4 percent and cardiac death in 1.0 percent.
- Among consecutive surgical patients with some selection criteria — MI in 3.2 percent and cardiac death in 1.7 percent.
- Among patients selected to undergo preoperative thallium scintigraphy — MI in 6.9 percent and cardiac death in 3.2 percent. (See 'Dipyridamole rMPI' below.)
Patients with underlying cardiovascular disease (eg, peripheral artery disease, stroke) have an increased risk of perioperative cardiac complications for two reasons:
- They constitute a selected population with a high incidence of significant coronary artery disease [3,4]. In addition, left ventricular systolic dysfunction (left ventricular ejection fraction ≤40 percent) is five times more common in patients with cerebrovascular disease or peripheral artery disease compared with matched controls [5].
- Physiologic factors associated with surgery predisposed to myocardial ischemia, which is more pronounced in patients with underlying coronary disease. These include volume shifts and blood loss, enhanced myocardial oxygen demand from elevations in heart rate and blood pressure secondary to stress from surgery, and an increase in postoperative platelet reactivity [6].
The optimal approach is to identify patients at high risk so that appropriate testing and therapeutic measures can be undertaken to minimize this risk [7]. The range of risk was illustrated in a retrospective study of 663,665 adults with no contraindications to beta blockers who underwent major noncardiac surgery in 2000 and 2001 at 329 hospitals in the United States [8]. Orthopedic and abdominal surgery accounted for 70 percent of cases. In-hospital mortality in patients not treated with beta blockers increased progressively from 1.4 to 7.4 percent according to a preoperative assessment of risk using the revised cardiac risk index described below. (See 'Revised cardiac risk index' below.)
This topic will review the initial preoperative evaluation, the surgery-specific risk, and the major indices used for risk stratification in patients undergoing noncardiac surgery. The management of cardiac risk in an attempt to reduce morbidity and mortality and issues related to the perioperative evaluation and management of heart failure are discussed separately. (See "Management of cardiac risk for noncardiac surgery" and "Perioperative heart failure in noncardiac surgery".)
INITIAL PREOPERATIVE EVALUATION — The clinician must integrate information from the history, physical examination, and electrocardiogram in order to develop an initial estimate of perioperative cardiac risk. The 2007 American College of Cardiology/American Heart Association (ACC/AHA) guidelines on perioperative cardiovascular evaluation for noncardiac surgery (not changed in the 2009 focused update) concluded that three elements must be assessed to determine the risk of cardiac events [9]:
- Patient specific clinical variables (table 1)
- Exercise capacity
- Surgery-specific risk (table 2)
History — A detailed history of the patient's symptoms, clinical course, and exercise tolerance yields valuable information for risk assessment. In particular, the initial preoperative evaluation should include a history inquiring about previous coronary heart disease, symptoms of angina, heart failure (HF), aortic stenosis, severe hypertension, and peripheral artery disease (PAD). (See 'Clinical predictors of perioperative risk' below.)
Functional capacity — An assessment of cardiac functional status should also be performed. The assessment of functional ability provides valuable prognostic information, since patients with good functional status have a lower risk of complications [2]. Functional status can be expressed in metabolic equivalents (1 MET is defined as 3.5 mL O2 uptake/kg per min, which is the resting oxygen uptake in a sitting position). Perioperative cardiac and long-term risk is increased in patients unable to meet a 4-MET demand during most normal daily activities.
Various activity scales provide the clinician with a set of questions to determine a patient's functional capacity [10]. Indicators of functional status include the following:
- Can take care of self, such as eat, dress or use the toilet (1 MET)
- Can walk up a flight of steps or a hill (4 METs)
- Can do heavy work around the house such as scrubbing floors or lifting or moving heavy furniture (between 4 and 10 METs)
- Can participate in strenuous sports such as swimming, singles tennis, football, basketball, and skiing (>10 METs)
One important indicator of poor functional status and an increased risk of postoperative cardiopulmonary complications after major noncardiac surgery is the inability to climb two flights of stairs or walk four blocks [11,12]. However, the presence of peripheral artery disease may be an important limitation to assessment of functional capacity since such patients often cannot exercise because of claudication.
The triad of good functional status, the absence of known cardiovascular disease, and a low score on one of the multifactorial risk indices described below is associated with a very low rate of major complications, even in patients undergoing major vascular surgery [2].
The role of cardiopulmonary exercise testing to determine functional capacity in this setting is discussed below.
Physical examination — The physical examination should include blood pressure measurements in both arms, an analysis of carotid artery and jugular venous pulsations for the quality of the pulse contour and the presence of bruits, auscultation of the lungs, precordial palpation and auscultation, abdominal palpation, and examination of the extremities for edema and vascular integrity. Important findings include evidence of HF or a murmur suspicious for aortic stenosis (AS), since poorly controlled HF and significant AS increase perioperative risk. (See 'Clinical predictors of perioperative risk' below.)
Resting electrocardiogram — The presence of Q waves or significant ST segment elevation or depression has been associated with an increased incidence of perioperative cardiac complications [13,14].
The additional value of the preoperative ECG compared to clinical characteristics was evaluated in a retrospective study of 2422 noncardiac surgery patients (54 percent referred for high risk surgery) [15]. Although right or left bundle branch block was associated with perioperative MI and left bundle branch block was associated with perioperative mortality, no single ECG abnormality provided additional predictive value for perioperative myocardial infarction or death in multivariate analysis.
Guideline recommendations — The 2007 ACC/AHA guidelines on perioperative cardiovascular evaluation (not changed in the 2009 focused update) recommended a preoperative resting 12-lead ECG in patients with at least one clinical risk factor scheduled to undergo vascular surgery (table 1) OR those patients scheduled to undergo intermediate-risk surgery with known cardiovascular disease, peripheral artery disease or cerebrovascular disease (table 3) [9]. A less strong recommendation was given for those scheduled to undergo vascular surgery with no clinical risk factors OR those scheduled to undergo intermediate-risk surgery with at least one clinical risk factor.
CLINICAL PREDICTORS OF PERIOPERATIVE RISK — The 2007 ACC/AHA guidelines (not changed in the 2009 focused update) summarized clinical predictors of increased perioperative risk for myocardial infarction, HF, and cardiac death (table 1) [9]. These predictors, which are derived from the history, physical examination, and resting ECG, help the clinician decide which patients may benefit most from further evaluation and aggressive management (medical or coronary revascularization).
Major predictors — When a major risk factor as described below is present, intensive management is mandated, which may result in delay or cancellation of surgery unless it is emergent.
Recent MI or severe angina — There are two unstable coronary syndromes which are associated with a major perioperative risk: recent myocardial infarction (MI), and unstable or severe angina.
The definition of recent MI has been modified with significant clinical implications. In the past, patients post-MI have been considered as being either three or six months status post-MI, and perioperative cardiac complication rates were assessed for these intervals. Older studies found as much as a 36 percent risk of reinfarction or cardiac death when patients underwent surgery within three months of a previous MI; the risk fell to 15 to 25 percent at three to six months; and 5 percent when surgery was performed after more than six months [16].
However, the risk was lower in later studies in the era of preoperative optimization and intensive perioperative monitoring. As an example, a prospective study performed between 1977 and 1982 evaluated the incidence of recurrent perioperative MI in 733 patients who had had a previous MI [16]. The reinfarction rate was 1.9 percent overall, 2.3 percent for those operated upon three to six months after a previous MI, and 5.7 percent for those operated on within three months of the previous MI. Current acute reperfusion therapies and coronary revascularization after MI presumably reduce the risk further, but data are insufficient.
Current management of MI or unstable angina provides for risk-stratification during convalescence in patients who have not been fully revascularized. If a stress test does not indicate ischemic myocardium at risk, the likelihood of reinfarction after noncardiac surgery is low. A positive stress test is usually an indication for revascularization. (See "Risk stratification after non-ST elevation acute coronary syndrome", section on 'Stress testing' and "Risk stratification after acute ST-elevation myocardial infarction", section on 'Stress testing'.)
The American College of Cardiology National Database Library defines recent MI as greater than seven days but ≤30 days [10]. Although there are no adequate clinical trials on which to base firm recommendations, it appears reasonable to wait at least four to six weeks after MI to perform elective surgery in such patients [10].
Recent PCI — Determining the optimal timing for elective surgery is more difficult in patients who have undergone a recent percutaneous coronary intervention (PCI), particularly with a drug-eluting stent. The risk of both postoperative MI and death is substantially elevated in patients treated with a stent who subsequently discontinue antiplatelet therapy at the time of surgery. For patients treated with a bare metal stent, it may be safe to stop antiplatelet therapy four to six weeks after implantation, but for those treated with a drug-eluting stent, where coronary endothelialization is delayed, it is recommended that dual antiplatelet therapy is continued for at least three to six months and preferably longer, even if surgery is performed. (See "Management of cardiac risk for noncardiac surgery", section on 'Preoperative PCI' and "Antiplatelet therapy after coronary artery stenting", section on 'Early noncardiac surgery or gastrointestinal endoscopy'.)
Severe angina (Canadian Cardiovascular Society [CCS]) class III or IV; (table 4) is also associated with a major risk of perioperative complications.
Decompensated heart failure or significant valvular heart disease — Decompensated (preoperative) HF or severe valvular disease predicts an increase in risk of mortality at the time of noncardiac surgery. These issues are discussed separately. (See "Perioperative heart failure in noncardiac surgery", section on 'HF as a risk factor for postoperative cardiac complications' and "Noncardiac surgery in patients with aortic stenosis".)
Significant arrhythmias — Significant arrhythmias are defined as high-grade atrioventricular block, sustained ventricular tachycardia, nonsustained ventricular tachycardia in the presence of underlying heart disease, and supraventricular arrhythmias with an uncontrolled ventricular rate.
Other clinical predictors — The 2007 ACC/AHA guidelines (not changed in the 2009 focused update) chose to replace intermediate predictors with clinical risk factors in the revised cardiac risk index (table 1 and table 5). When present, they justify careful assessment of the patient's current status, which may necessitate noninvasive testing [9,17]:
- Ischemic heart disease including mild angina pectoris (CCS class I or II) (table 4) or prior MI by history or pathologic Q wave
- Compensated or prior heart failure (HF)
- Diabetes mellitus
- Renal insufficiency (preoperative creatinine >2.0 mg/dL [177 µmol/L])
- Cerebrovascular disease
While not included in the ACC/AHA list of clinical risk factors, a single, retrospective, administrative database study (Alberta, Canada) demonstrated an association between a history of prior admission for atrial fibrillation (AF) and postoperative complications, as well as confirming that compensated or prior HF is a risk factor [18]. This study, which included patients undergoing both major and minor surgery, was used to assess the risk of postoperative mortality in patients with stable coronary artery disease (n=13,783), ischemic HF (n=12,249), nonischemic HF (n=7700) or AF (n=4312) [18]. Patients were assigned a diagnosis of CAD, HF, or AF only if previously hospitalized for that diagnosis. The following findings were noted:
- The unadjusted 30-day mortality was 2.9, 9.2, 9.3, and 6.4 percent respectively in the four groups.
- After multivariable adjustment, the rates for both HF groups and the AF group were significantly higher than the rate for the CAD group.
- Patients with HF had a significantly worse outcome than those with CAD even after a minor surgical procedure.
Obese patients — The prevalence of coronary heart disease (CHD) and its risk factors is increased in obese patients. (See "Obesity, weight reduction, and cardiovascular disease", section on 'Coronary disease'.)
A 2009 American Heart Association (AHA) scientific advisory on obesity noted an association with cardiac and pulmonary diseases, which may negatively affect the outcome of surgery [19]:
For these reasons obese patients are at increased risk for adverse cardiovascular events at the time of non-cardiac surgery. The estimation of risk is more difficult in obese patients due to the uncertain significance of certain components of the history (dyspnea) and physical examination (lower-extremity edema). The generally poor exercise capacity of the obese further complicates risk assessment.
The issue of whether the preoperative approach to obese patients should differ from that in the general population is uncertain. The 2009 scientific advisory on cardiovascular evaluation and management of severely obese patients undergoing surgery from the AHA states that specific tests should be performed only if the results will change management [19]. We agree with that approach and with the following specific suggestions:
- A 12 lead electrocardiogram (ECG) is reasonable in all obese patients with at least one risk factor for CHD (diabetes, smoking, hypertension, or hyperlipidemia) or poor exercise tolerance.
- A chest radiograph (posteroanterior and lateral) should be obtained on all severely obese patients prior to surgery, as the sensitivity and specificity (for the diagnosis of cardiopulmonary disease) of the physical examination is decreased.
Minor predictors — Minor risk factors are recognized markers for cardiovascular disease that have not been definitively proven to independently increase perioperative risk and were not included in the recommendations for treatment by the ACC/AHA perioperative guidelines [9]:
- Advanced age (greater than 70 years of age)
- Abnormal ECG (left ventricular hypertrophy, left bundle branch block, ST-T wave abnormalities)
- Rhythm other than sinus (such as atrial fibrillation)
- Uncontrolled systolic hypertension
The presence of more than one of these may lead to a higher suspicion of coronary artery disease.
SURGERY-SPECIFIC RISK — The type and timing of surgery significantly affects the patient's risk of perioperative cardiac complications. The 2007 ACC/AHA guidelines (not changed in the 2009 focused update) stratified surgical risk by procedure (table 2) [9]. By definition, the reported rate of cardiac death or nonfatal MI is more than 5 percent in high-risk procedures, between 1 and 5 percent in intermediate-risk procedures, and less than 1 percent in low-risk procedures. Institutional and/or individual surgeon experience with the procedure may increase or lower the risk. Emergency surgery is associated with particularly high risk, as cardiac complications are two to five times more likely than with elective procedures [20].
Invasive dental procedures — Chronic periodontitis has been associated with an increased risk of coronary heart disease events in observational studies. (See "Epidemiology, pathogenesis, and clinical manifestations of odontogenic infections", section on 'Association with cardiovascular risk'.)
Although invasive dental procedures are associated with bacteremia and are often performed in patients with chronic periodontitis, dental procedures are considered to be of low cardiovascular risk [21] and were not mentioned in the ACC/AHA guidelines (table 2) [9]. The relationship between cardiovascular risk and invasive dental procedures (defined as those that could possibly result in bacteremia and induce an inflammatory response) was evaluated in a self-controlled case series in 1152 individuals who sustained either an ischemic stroke or myocardial infarction [22]. The rate of vascular events significantly increased in the first four weeks after invasive dental treatment (incidence ratio 1.50, 95% CI 1.09-2.06, reflecting an increase in absolute risk of approximately 1.2 percent) and gradually returned to the baseline rate within six months. The observational nature of this study precludes drawing definitive conclusions about the observed relationship.
RISK MODELS — Multivariable analyses, initially developed by Goldman, Detsky, and Eagle, identified combinations of factors, generally based upon routine clinical information and laboratory tests, that could be used to estimate the risk of cardiac complications. These risk indices were derived from patients who were candidates for surgery; they did not include patients with high-risk conditions that were considered major predictors of risk in the ACC/AHA guidelines that require intensive management and often lead to delay in or cancellation of surgery (table 1) [3]. (See 'Major predictors' above.)
More recent models such as the Revised cardiac risk index and the National Surgical Quality Improvement Program (NSQIP)risk model, have included patients with high risk characteristics and who have been managed with more current standards of care. For these reasons, we suggest using one of these two models.
Goldman risk index — In 1977, Goldman and colleagues assigned a point value to each of nine clinical risk factors and divided 1001 surgical patients into four risk classes on the basis of their total point scores (table 6) [1]. Due to multiple limitations, including complexity and changes in the perioperative care of patient, we do not recommend the use of this model.
Revised cardiac risk index — To simplify the prediction of risk, Goldman monitored 2893 patients (mean age 66) undergoing elective major noncardiac procedures and identified six independent predictors of major cardiac complications (defined as myocardial infarction, pulmonary edema, ventricular fibrillation or primary cardiac arrest, and complete heart block; all cause mortality was not included) (table 5) [17]:
- High-risk type of surgery (intraperitoneal, intrathoracic, or suprainguinal vascular surgery)
- History of ischemic heart disease (history of MI or a positive exercise test, current complaint of chest pain considered to be secondary to myocardial ischemia, use of nitrate therapy, or ECG with pathological Q waves; do not count prior coronary revascularization procedure unless one of the other criteria for ischemic heart disease is present)
- History of HF
- History of cerebrovascular disease
- Diabetes mellitus requiring treatment with insulin
- Preoperative serum creatinine >2.0 mg/dL (177 µmol/L)
These predictors were then validated in a cohort of 1422 patients. The predictive value was significant in all types of elective major noncardiac surgery except for abdominal aortic aneurysm surgery (figure 1). A 2009 systematic review evaluated the ability of the revised cardiac risk index (RCRI) to predict cardiac complications and mortality after major non-cardiac surgery in various populations and settings [23]. The RCRI performed moderately well in distinguishing patients at low compared to high risk for all types of non-cardiac surgery, but was somewhat less accurate in patients undergoing only vascular non-cardiac surgery. In addition, RCRI did not predict all-cause mortality well, but this is expected, as it does not capture risk factors for noncardiac causes of perioperative mortality. Only one third of perioperative deaths are due to cardiac causes.
The risk of major cardiac complications varied according to the number of risk factors. If one includes only cardiac death, nonfatal MI, and nonfatal cardiac arrest as major cardiac events in the Goldman cohort, the following rates of adverse outcomes were seen [7]:
- No risk factors — 0.4 percent (95% CI 0.1-0.8 percent)
- One risk factor — 1.0 percent (95% CI 0.5-1.4 percent)
- Two risk factors — 2.4 percent (95% CI 1.3-3.5 percent)
- Three or more risk factors — 5.4 percent (95% CI 2.8-7.9 percent)
The RCRI has better predictive value than the original Goldman index or the Detsky modified risk index [17].
Later studies have found a higher rate of events at the same RCRI as in the Goldman cohort in which the patients were seen between 1989 and 1994:
- In a large retrospective analysis on mortality, the perioperative risk was evaluated in 663,665 adults with no contraindications to beta blockers who underwent major noncardiac surgery in 2000 and 2001 at 329 hospitals in the United States [8]. In-hospital mortality in patients not treated with beta blockers increased progressively from 1.4 percent at a score of 0, to 7.4 percent at a score ≥4. The rate of mortality in this study was higher at the same RCRI than the combined end point of cardiac death, nonfatal MI, and nonfatal cardiac arrest in the earlier Goldman population [7].
- In the POISE randomized trial of over 8000 patients undergoing noncardiac surgery between 2002 and 2007, the combined rate of cardiovascular death, nonfatal MI, and nonfatal cardiac arrest was 6.9 percent in the placebo group [24]. The majority of these individuals were RCRI 1 or 2. (See "Management of cardiac risk for noncardiac surgery", section on 'Beta blockers'.)
There are at least four factors that probably contribute to the higher event rate in these two later studies:
- Patients were sicker at baseline in the later studies, an effect that may be even larger now.
- The original RCRI risk prediction model did not take mortality into account [17].
- In earlier studies CK-MB was used to diagnose MI, rather than troponins, which are more sensitive.
- The original RCRI study did not include patients requiring emergent surgery [17].
Perioperative myocardial infarction after noncardiac surgery is discussed in detail elsewhere. (See "Perioperative myocardial infarction after noncardiac surgery".)
NSQIP database risk model — The American College of Surgeons’ National Surgical Quality Improvement Program database was used to determine risk factors associated with intraoperative/postoperative myocardial infarction or cardiac arrest (MICA) [25]. Among over 200,000 patients who underwent surgery in 2007, 0.65 percent developed perioperative MICA. On multivariate logistic regression analysis, type of surgery, dependent functional status, abnormal creatinine, American Society of Anesthesiologists’ class (table 7), and increased age were identified as predictors of MICA. A risk model was developed using these five risks factors and subsequently validated on a 2008 data set (n = 257,385). The risk model had a relatively high predictive accuracy (C statistic of 0.874), and outperformed the RCRI (C statistic of 0.747). An easy to use calculator was developed from this model [26].
Detsky modified risk index — Detsky and coworkers added angina and pulmonary edema to the original Goldman variables: unstable angina within three months prior to surgery, stable angina occurring with minimal physical activity, and recent pulmonary edema were assigned a high number of points, thereby contributing to an increased risk (table 8) [27,28].
The Detsky modification may not have sufficient discriminant power to identify significant coronary artery disease in patients at the lower end of the spectrum of clinical risk. Both indices may also underestimate cardiac risk in patients with vascular disease [29].
Eagle criteria — Eagle and colleagues retrospectively studied 254 consecutive patients presenting to the nuclear cardiology laboratory for dipyridamole-thallium imaging before proposed vascular surgery [13]. Logistic regression analysis identified five clinical predictors of postoperative cardiac events in patients undergoing major vascular surgery:
- Q waves on the ECG
- History of angina
- History of ventricular ectopy requiring treatment
- Diabetes mellitus requiring therapy other than diet
- Age above 70 years
A subsequent validation study added three more clinical predictors to the Eagle criteria [30]:
- Ischemic ST segment abnormalities on resting ECG
- Hypertension with severe left ventricular hypertrophy
- History of heart failure
Logistic regression further identified two independent dipyridamole-thallium test predictors of ischemic events: thallium redistribution; and ischemic ECG changes during or after dipyridamole infusion [13]. (See "Basic properties of myocardial perfusion agents", section on 'Redistribution' and "Vasodilator stress radionuclide myocardial perfusion imaging in the diagnosis and prognosis of coronary heart disease".)
The original report and the prospective validation study demonstrated that noninvasive testing can be used to stratify intermediate to high risk vascular surgery patients [13,30]. The original report suggested that lower risk patients with none or one of the five risk factors are likely to receive little benefit from noninvasive testing [13]. Patients at low risk are also highly unlikely to have severe and extensive coronary disease or critical stenoses on angiography [31].
Fleisher-Eagle criteria — A subsequent review by Fleisher and Eagle emphasized six factors, the first five of which are also in the revised cardiac risk index, associated with increased cardiac risk in patients undergoing noncardiac surgery, including vascular surgery [32]:
- Ischemic heart disease (angina or prior MI)
- Heart failure
- High-risk surgery (including intraperitoneal, intrathoracic, and suprainguinal vascular procedures)
- Diabetes mellitus (especially insulin-requiring)
- Renal insufficiency
- Poor functional status (defined as the inability to walk four blocks or climb two flights of stairs)
The authors recommended that further evaluation was required in patients with one or more of these risk factors. (See 'Algorithmic approaches' below.)
Aortic stenosis — Most of the risk indices did not include aortic stenosis as a risk factor, probably because their sample populations did not include enough patients with the condition. However, moderate-to-severe aortic stenosis appears to be a significant risk factor for adverse cardiac outcomes after noncardiac surgery. In the original 1977 study of the Goldman cardiac risk index, for example, there were 23 patients with aortic stenosis (2.3 percent) [1]. These patients had a 17.3 percent risk of cardiac complications and a 13 percent cardiac mortality. As a result, aortic stenosis was given a risk of 3 points.
The approach to patients with aortic stenosis who need to undergo noncardiac surgery is discussed in detail elsewhere. (See "Noncardiac surgery in patients with aortic stenosis".)
Summary — When assessing preoperative cardiac risk, we prefer either the RCRI (table 5) or the NSQIP risk model. The advantage of the RCRI is that it has been in use for over 10 years and many practitioners are familiar with it. The strengths of the NSQIP model include validation in a very large cohort, somewhat higher predictive accuracy than the RCRI, and development from a cohort of coronary artery disease patients treated with more recent preventative and revascularization strategies. Practitioners should become familiar with one model and use it regularly.
(See 'Revised cardiac risk index' above.)
The use of estimated risk to determine whether further testing is indicated is discussed in the following sections. The use of this information for deciding whether preoperative interventions are indicated is discussed elsewhere. (See "Management of cardiac risk for noncardiac surgery".)
NONINVASIVE CARDIAC TESTING — The above risk indices provide useful guidelines but are not sufficiently precise to estimate risk, particularly in patients at intermediate risk. Noninvasive testing with either stress testing, resting echocardiography, or both permits further stratification in patients deemed to be at intermediate risk after clinical evaluation.
General principles — The following discussion supports two general conclusions about noninvasive testing with exercise or pharmacologic stress [7]:
- For the purpose of predicting perioperative cardiac events, stress testing has a high negative predictive value (90 to 100 percent) but low positive predictive value (6 to 67 percent) [33]. A 1995 review of 1410 selected patients referred for dipyridamole-thallium MPI prior to vascular surgery found a negative predictive value of 98 percent but a positive predictive value of only 18 percent [2].
The positive predictive value will be lower when positive tests are used to initiate therapies that prevent adverse cardiac outcomes, such as coronary revascularization, adjusting perioperative medical treatment and monitoring, or selecting different surgical procedures [13]. By comparison, the negative predictive value remains uniformly high (eg, as high as 98 to 99 percent with pharmacologic stress testing) [13]. - Exercise is usually the preferred stress since, as noted below, exercise tolerance is an important predictor of outcome that appears to be more important than the ECG response. Pharmacologic stress testing can be used in patients who cannot exercise. Test performance and, in patients who cannot exercise, the choice between dipyridamole-thallium radionuclide myocardial perfusion imaging and dobutamine echocardiography are discussed below. (See 'Choice of stress test' below.)
The role of preoperative cardiac testing in intermediate risk patients (one to two RCRI risk factors) has been further elucidated by the DECREASE-II trial of 770 patients scheduled to undergo major vascular surgery (mostly for abdominal aortic aneurysm or peripheral artery disease) who were randomly assigned to pharmacologic stress testing or no testing [34]. Patients with positive tests and extensive ischemia could be managed with preoperative revascularization. All patients received perioperative beta blocker therapy, with dose adjustment with a goal resting heart rate between 60 to 65 beats/min.
The incidence of the primary end point of cardiac death or nonfatal myocardial infarction at 30 days after surgery was similar between the non-testing and stress testing groups (1.8 versus 2.3 percent; odds ratio 0.78; 95% CI 0.3-2.1). The authors concluded that preoperative stress testing can be safely omitted in intermediate risk patients who have stable or no clinical coronary disease in whom beta blockers are given aiming at tight heart rate control. However, only 34 patients randomized to stress testing had extensive ischemia and few (six patients) had complete revascularization.
Algorithmic approaches — In an attempt to improve risk stratification, a number of algorithmic approaches have been proposed, all of which include stress testing in selected patients at increased risk [9,32,35-37]. Perioperative risk stratification and management can characterize and possibly reduce risk, but will not eliminate it. Furthermore, no studies have been performed showing that use of these algorithms improves patient outcomes.
Three major algorithms have been proposed by the ACC/AHA, Fleisher and Eagle, and the American College of Physicians (ACP). The ACC/AHA guidelines and the Fleisher-Eagle approach are largely based upon the criteria in the revised cardiac risk index (RCRI) (table 5) [17]. (See 'Revised cardiac risk index' above.)
- The ACC/AHA algorithm utilizes the presence or absence of active clinical conditions (eg, acute coronary syndrome or heart failure), the number of clinical risk factors (table 1), functional capacity, and type of surgery (table 2) [9].
- The Fleisher-Eagle algorithm uses five of the six criteria in the RCRI, and substitutes a measure of functional status for a history of cerebrovascular disease [32].
- The ACP guidelines are based upon the Detsky modified risk index (table 8) [35,36].
The ACP and Fleisher-Eagle approaches were derived from patients who were candidates for surgery; they did not include patients with high-risk conditions that were considered major predictors of risk in the ACC/AHA guidelines; such patients require intensive management and often lead to delay in or cancellation of surgery (table 1) [36]. (See 'Major predictors' above.)
The importance of underlying risk status as a determinant of the value of stress testing can be illustrated by the following calculations with dobutamine echocardiography [33], which has a reported sensitivity and specificity of 85 and 70 percent, respectively, for predicting perioperative death or nonfatal MI [38].
- A patient with two RCRI risk factors may have a 2 percent pretest probability of a cardiac complication (table 5). A positive dobutamine stress test results in a modest increase in posttest probability to 5 percent.
- A patient with four RCRI risk factors may have a 9 percent pretest probability of a cardiac complication (table 5). A positive dobutamine stress test increases the posttest probability to about 20 percent, while a negative test reduces the posttest probability to about 2 percent.
ACC/AHA guidelines — The 2007 ACC/AHA (not changed in the 2009 focused update) recommended that the estimation of perioperative risk should integrate major, intermediate, and minor predictors of cardiac risk (table 1), functional capacity (ie, metabolic equivalent [MET] or exercise duration), the surgery-specific risk (table 2), and, when indicated, the results of noninvasive studies, including stress testing (table 9) and (see 'Choice of stress test' below) and evaluation of left ventricular function (table 10) [9]. Using this approach, only a minority of patients undergoing elective, as opposed to urgent or emergent surgery, will undergo preoperative noninvasive or invasive testing.
The following points summarize the algorithm as it applies to STABLE patients. Patients with ACTIVE cardiac conditions (eg, unstable coronary syndromes, decompensated heart failure, significant arrhythmias, or severe valvular stenosis) not undergoing emergency surgery should first be managed according to ACC/AHA guidelines:
- Patients without active cardiac symptoms undergoing low risk surgery (table 2) OR those with good functional capacity (MET level ≥4) require no preoperative testing aside from an ECG (table 3).
- Patients with poor or unknown functional capacity, or potential cardiac symptoms who are scheduled to undergo intermediate risk or vascular surgery (table 2) are managed according to the number of clinical risk factors (table 1). For such patients, heart rate control with beta blockade may be of benefit. (See "Management of cardiac risk for noncardiac surgery", section on 'Beta blockers'.)
- In patients with one or two clinical risk factors (table 1), it is reasonable to proceed to surgery, unless testing will alter management
- In patients with three of more clinical risk factors scheduled to undergo intermediate risk surgery (table 2), it is reasonable to proceed without noninvasive testing, unless that testing will alter management
- In patients with three or more clinical risk factors scheduled to undergo vascular surgery, the guidelines suggest noninvasive testing if it will alter management.
ACP guidelines — The ACP guidelines for assessing and managing perioperative risk are based upon the Detsky modified risk index (table 8) [28] and the results of noninvasive testing (dipyridamole thallium imaging or dobutamine stress echocardiography) [35,36]. (See 'Detsky modified risk index' above and 'Eagle criteria' above.)
Fleisher-Eagle algorithm — The Fleisher-Eagle criteria emphasized six factors associated with increased cardiac risk in patients undergoing noncardiac surgery, including vascular surgery [32]. Five of these factors were also in the revised cardiac risk index. The Fleisher-Eagle algorithm is virtually identical to a subsequent algorithm by Auerbach and Goldman that used the revised cardiac risk index [33]. (See 'Fleisher-Eagle criteria' above.)
The authors recommended that patients with none of the risk factors required no further testing or treatment, while further evaluation was required in other patients. Those with one or two risk factors should be further evaluated to determine if a history consistent with coronary heart disease is present. If such a history is absent, perioperative beta blocker therapy is prudent, but no further testing is recommended. For patients with a history consistent with coronary disease, and for those with three or more of the above risk factors, noninvasive testing should be performed, followed by cardiac catheterization if the results are positive.
In this and other approaches, the algorithm can also identify patients who might benefit from beta blocker therapy [32,33,37]. (See "Management of cardiac risk for noncardiac surgery", section on 'Beta blockers'.)
Choice of stress test — Exercise ECG testing is usually the preferred stress test since, as noted below, exercise tolerance is an important predictor of outcome that appears to be more important than the ECG response [2,39,40]. Exercise ECG testing is usually performed with perfusion imaging or echocardiography since imaging can better identify high-risk features that would warrant referral for angiography (eg, reversible large anterior wall defect, multiple reversible defects, ischemia occurring at a low heart rate, extensive stress-induced wall motion abnormalities, transient ischemic dilatation).
In addition, concurrent imaging is essential if the resting ECG has abnormalities that can interfere with the detection of ischemia during exercise stress. These include preexcitation (Wolff-Parkinson-White) syndrome, a paced ventricular rhythm, more than 1 mm of ST depression at rest, complete left bundle branch block, and patients taking digoxin or with ECG criteria for left ventricular hypertrophy, even if they have less than 1 mm of baseline ST depression. (See "Exercise ECG testing: Performing the test and interpreting the ECG results", section on 'Limitations to exercise ECG testing'.)
A possible exception to the preference for exercise stress occurs in patients with an aortic aneurysm, since exercise increases systolic pressure and heart rate [41]. Despite this theoretical concern, exercise stress appears to be safe in patients with abdominal aortic aneurysms, particularly those less than 6.0 cm in diameter [41]. Pharmacologic stress testing is preferred in patients with abdominal aortic aneurysms ≥6.0 cm in diameter or aneurysms that are symptomatic. It is appropriate to control hypertension prior to stress testing in patients with aneurysms.
Limited data exist upon which to make recommendations for stress testing in patients with thoracic aortic aneurysms, but a similar approach to that for abdominal aortic aneurysms seems reasonable.
Among patients who cannot exercise sufficiently to reach the target heart rate (eg, those who are not limited by claudication, orthopedic degenerative joint disease, or severe deconditioning), the two most widely used alternatives are dipyridamole-thallium radionuclide myocardial perfusion imaging (rMPI) and dobutamine echocardiography.
A meta-analysis evaluated the sensitivity and specificity of the different noninvasive tests for predicting perioperative death or nonfatal MI in patients undergoing major vascular surgery [38]. The following values were obtained:
- Exercise ECG — sensitivity 74 percent, specificity 69 percent
- rMPI, mostly dipyridamole stress — sensitivity 83 percent, specificity 49 percent
- Dobutamine echocardiography — sensitivity 85 percent, specificity 70 percent
However, these findings should be interpreted with caution because the majority of studies in the meta-analysis were not well designed, there was significant heterogeneity in the results for individual tests, and the results were analyzed as either negative or positive [7].
When pharmacologic stress testing is indicated, the choice between tests should be based upon local experience and availability and the relative safety of the different procedures in the individual patient [42]:
- Dipyridamole-thallium imaging is usually preferred in patients with known cardiac arrhythmias, since dobutamine can induce atrial or ventricular arrhythmias.
- Dobutamine echocardiography is preferred in patients with bronchospastic lung disease and in those with severe carotid stenosis, because dipyridamole can induce bronchospasm and a decrease in blood pressure. It is also preferred when information about left ventricular function or valvular heart disease is desired [33]. Dobutamine appears to be safe in patients with abdominal aortic aneurysms [43].
Stress testing has a very high negative predictive value for postoperative cardiovascular events (between 90 and 100 percent) but a low positive predictive value (between 6 and 67 percent, with a value of 18 percent in a review of five large studies of thallium perfusion imaging) [2,33]. Thus, stress testing is more useful for reducing estimated risk if negative (or normal) than for identifying patients at very high risk if positive [33].
Exercise stress testing — Exercise stress testing without myocardial imaging is the standard method for determining functional capacity and for the detection of myocardial ischemia. Exercise tolerance appears to be more important than the ECG response to exercise [2,39,40]. Inability to perform moderate exercise or to achieve greater than 85 percent of predicted maximal heart rate during exercise treadmill testing is associated with a high risk of a postoperative cardiac event, even in the absence of diagnostic ischemic ECG changes [39,40].
There are, however, a number of limitations to exercise stress testing in assessing cardiac risk prior to noncardiac surgery:
- Many patients are unable to exercise or unable to exercise to their target heart rate due to leg claudication, arthritis, deconditioning, or associated pulmonary disease.
- Patients with resting ECG abnormalities may require imaging to detect myocardial ischemia.
- The sensitivity of exercise testing in detecting coronary artery disease is lower with submaximal than with maximal exercise.
Exercise stress testing can be coupled with respiratory gas analysis (see "Functional exercise testing: Ventilatory gas analysis") and this is often referred to as cardiopulmonary exercise testing. Although cardiopulmonary exercise testing has been advocated to assess functional capacity and to guide prognosis for patients undergoing major noncardiac vascular surgery, a 2012 meta-analysis of six observational studies with significant limitations found insufficient evidence to support its use in this setting [44].
Pharmacologic stress testing — Pharmacologically-induced cardiac stress testing is warranted in patients who cannot exercise and have abnormalities on the baseline ECG that interfere with interpretation. The two most common used tests are dipyridamole-thallium radionuclide myocardial perfusion imaging and dobutamine echocardiography. (See "Vasodilator stress radionuclide myocardial perfusion imaging in the diagnosis and prognosis of coronary heart disease" and "Stress echocardiography in the diagnosis and prognosis of coronary heart disease".)
Dipyridamole rMPI — A number of studies have evaluated the predictive value of dipyridamole-thallium imaging in patients undergoing major noncardiac surgery [2,13,30,45-55]. In a 1995 review of the five largest series of 1410 selected patients referred for dipyridamole-thallium rMPI prior to vascular surgery, the sensitivity and specificity for a major cardiac event was 85 and 60 percent, respectively, while the relative risk for a cardiac event in patients with a positive test was 9.0 [2]. The negative predictive value of dipyridamole-thallium imaging was 98 percent, but the positive predictive value was only 18 percent. Thus, a positive test is a relatively weak predictor of a perioperative cardiac event.
The ability to identify patients at risk may be increased when taking into account the extent of ischemia rather than only its presence. This was illustrated in a meta-analysis of nine studies that evaluated the probability of MI or cardiac death after vascular surgery according to the extent of reversible myocardial defects [56]. The probability of MI or cardiac death without reperfusion ranged from 3 to 4 percent in patients with no or only fixed defects to 9 percent with reversible defects involving less than 20 percent of the left ventricle to 18 and 45 percent with reversible defects involving 30 to 49 percent and ≥50 percent of the left ventricle, respectively. However, most of the studies were deemed to be of low quality.
Other studies showed that the specificity can be increased by testing in subgroups at intermediate risk [13,51] or by measuring the amount of myocardium at risk [50,54]:
- The potential value of stress testing was illustrated in a report in which 18 of 116 patients with one or two Eagle predictors who were undergoing major vascular surgery had cardiac ischemic events in the early postoperative period [13]. Among these 116 patients, the rate of events was much higher in the 54 patients with thallium redistribution compared to the 62 without thallium redistribution (30 versus 3 percent). Use of both clinical and thallium variables had significantly higher specificity and equivalent sensitivity compared to predictions from clinical or thallium variables alone.
A similar increase in risk of a perioperative event with a reversible thallium defect in one or more segments was noted in a series of 161 patients with an intermediate to high risk of coronary disease (27 versus 6 percent) [51].
- The importance of the extent of jeopardized myocardium was illustrated in a study of 355 patients undergoing vascular and major general surgery [54]. The rate of postoperative events varied from 1.3 percent in the 225 patients with a normal dipyridamole-thallium study to 52 percent in the 29 patients with reversible defects in all three segments, at least one severe reversible defect, or transient left ventricular dilation. Among the 101 patients with one or two reversible defects, the rate of cardiac events ranged from 5 to 36 percent depending upon the number of defects, age greater than 70, and the presence of diabetes.
The predictive value of dipyridamole-thallium rMPI may also be improved by concurrent echocardiography to detect a reduced left ventricular ejection fraction [53]. (See 'Resting echocardiography' below.)
In contrast to these findings, a prospective study of 457 consecutive (unselected) patients undergoing elective abdominal aortic surgery did not confirm the usefulness of dipyridamole-thallium scintigraphy in stratifying unselected patients with definitive evidence of coronary artery disease [14]. However, a number of methodological issues may make these results less applicable for many patients undergoing abdominal aortic surgery. High-risk patients (such as those with severe coronary disease) and patients undergoing revascularization were excluded, and the population of patients selected appeared to represent a low-risk group. Furthermore, invasive perioperative monitoring was used in all patients, a regimen that differs from general practice.
- Use in risk model — A risk prediction model was developed for patients considered for vascular surgery that included the results of radionuclide myocardial perfusion imaging [57]. The model was retrospectively developed based upon 567 patients and validated prospectively in 514 patients. A risk score was derived using clinical variables including:
- Age >70
- Angina
- History of MI
- Diabetes mellitus
- History of HF
- Prior coronary revascularization within five years, which was given a negative weight since it is protective
A second model was based upon dipyridamole-thallium predictors of myocardial infarction, ie, fixed and reversible myocardial defects. The following findings were noted:
- The prognostic accuracy was 74 percent for the clinical and 81 percent for the clinical and dipyridamole model.
- Using the clinical model, the postoperative event rate was 3, 8, and 18 percent for patients classified as low, moderate, and high risk. Addition of the dipyridamole-thallium scan reclassified the moderate risk patients into low (3 percent) and high (19 percent) risk categories. The scan provided no additional help for patients classified as low or high risk based upon clinical parameters.
Dobutamine echocardiography — Dobutamine stress echocardiography (DSE) has a discriminant ability in vascular and nonvascular surgery patients that is similar to dipyridamole-thallium studies and provides additional information about valvular anatomy and resting and stress systolic function. (See "Stress echocardiography in the diagnosis and prognosis of coronary heart disease".)
- Role in vascular surgery — Dobutamine stress echocardiography (DSE) is an alternative accepted method for preoperative assessment of cardiac risk in patients undergoing vascular surgery [58-62].
The potential efficacy was illustrated in a study of 316 patients undergoing vascular surgery who had preoperative DSE and were followed for 19 months postoperatively [59]. Using a multivariate regression analysis, the occurrence of extensive (three or more segments) or limited (one or two segments) stress-induced new wall motion abnormalities on DSE, and a history of previous MI independently predicted late cardiac events, increasing the risk by 6.5, 2.9, and 3.8 fold, respectively (figure 2). The severity of ischemia during stress and the threshold heart rate for ischemia were not predictive in this study, but have been important in other studies [58,62].
The predictive value of DSE appears to vary with patient risk, having limited utility in patients at very low risk [58] and in lower risk patients taking a beta blocker [62]. This was illustrated in an analysis of 1097 patients in which DSE did not provide additional prognostic information among those estimated to be at low risk (less than three clinical risk factors using the Eagle criteria cited above) who were treated with a beta blocker (figure 3) [62]. In contrast, DSE provided additional prognostic information in patients who had three or more clinical risk factors. In addition, patients with ≥5 segments involved had more cardiac events than those with limited stress-induced ischemia (one to four segments) (36 versus 3 percent). Among such patients taking a beta blocker, those with stress-induced ischemia had a significantly higher risk of a cardiac event at 30 days (10.6 versus 2 percent without ischemia) (figure 3).
Hypotension during DSE is another factor associated with an adverse cardiovascular outcome (cardiac death, infarction, and ischemia) in the perioperative period. In one series of 300 patients, a hypotensive response was observed in 28 percent; with a multivariate analysis, hypotension was a significant predictor of a cardiac event (odds ratio 4.1) [60].
A potential concern with dobutamine echocardiography is the safety of increasing cardiac output and blood pressure in patients with an abdominal aneurysm. In a series of 98 such patients, there were no serious complications with its use [43].
Role in nonvascular surgery — Data are more limited concerning the utility of DSE for risk stratification before nonvascular surgery. The likelihood of coexistent coronary heart disease in this more general population is probably lower than among those undergoing vascular surgery. This issue was addressed in a report of 530 patients who underwent DSE before nonvascular surgery [63]. The following findings were noted:
- A postoperative cardiac event occurred in 6 percent (31 nonfatal MIs and one cardiac death), all of whom had inducible ischemia on dobutamine echocardiography. The sensitivity and specificity of DSE were 100 and 63 percent, respectively.
- DSE identified 60 percent of patients as low risk (no ischemia), 32 percent as intermediate risk (ischemia at ≥60 percent of age predicted maximal heart rate), and 8 percent as high risk (ischemia at <60 percent of maximal heart rate). The postoperative event rates in these three groups were 0, 9, and 43 percent, respectively. In contrast, use of the Eagle criteria identified 21, 68, and 11 percent of patients as low, intermediate, or high risk; the event rates were 3, 6, and 14 percent, respectively. Thus, DSE added incremental value to the Eagle criteria in identifying patients at high risk. (See 'Eagle criteria' above.)
Resting echocardiography — Resting transthoracic echocardiography (TTE) may add predictive information in certain patients undergoing noncardiac surgery [53,64]. TTE can assess left ventricular systolic function, including the effect of a prior MI, and evaluate possible valvular heart disease, such as aortic stenosis. (See 'Aortic stenosis' above.)
Potential utility of resting TTE was suggested in a study of 570 patients who underwent TTE within three months of surgery [64]. In a multivariate model that included both clinical and echocardiographic variables, TTE findings of systolic dysfunction, moderate to severe left ventricular hypertrophy, and a peak aortic valve gradient of ≥40 mmHg were associated with odds ratios for major cardiac events of 2.0, 2.1, and 6.8, respectively.
However, other studies have found that measurement of left ventricular ejection fraction was useful only for the prediction of heart failure [14] and that resting TTE does not add appreciably to the information obtained in the clinical risk models described above [2,65].
The main indications for preoperative resting TTE are similar to those in nonsurgical patients (eg, to evaluate valve function in patients with a murmur or left ventricular systolic function, especially in patients with heart failure of unknown cause) [2].
The 2007 ACC/AHA guidelines on perioperative cardiovascular evaluation and care for noncardiac surgery (not changed in the 2009 focused update) recommended assessment of LV function in patients with current heart failure or prior heart failure with worsening dyspnea, as well as those with dyspnea of unknown origin [9].
Troponin — The potential role of troponin testing in perioperative risk stratification is discussed elsewhere. (See "Perioperative myocardial infarction after noncardiac surgery", section on 'Biomarker testing'.)
Ischemia monitoring — Among patients with known or at high risk for coronary heart disease who undergo preoperative ambulatory ECG monitoring, 20 to 40 percent have frequent and often silent episodes of myocardial ischemia during the 48 hours before surgery [2,66-68]. Such patients appear to be at increased risk for adverse cardiac outcomes [66-68]. On the other hand, the absence of preoperative or immediate postoperative ischemic ST-segment depression on continuous electrocardiographic monitoring for 24 to 48 hours has been associated with a very low postoperative cardiac event rate in vascular surgery patients otherwise determined to be at intermediate to high risk [66].
However, it is not clear if ischemia monitoring is sufficiently sensitive to identify patients at low risk [2,68]. In addition, the usefulness of this test in the general population is limited because the above reports excluded patients with ECG findings that preclude accurate evaluation for ischemia (left bundle branch block, left ventricular hypertrophy with "strain" pattern, or digitalis effect).
SUMMARY AND RECOMMENDATIONS — The process of estimating and reducing the risk of perioperative cardiac events (eg, cardiac death and nonfatal MI), includes the following four components:
- Defining the urgency of surgery. For instance, risk assessment may not alter the management of patients who need emergency surgery. (See 'Urgency of surgery' below.)
- Initial risk assessment.
- Refinement of initial risk assessment with noninvasive testing in selected patients.
- Efforts to reduce risk in high-risk patients (eg, beta blockers, revascularization). These therapeutic issues and those related to the perioperative evaluation and management of heart failure in relation to noncardiac surgery are discussed separately. (See "Management of cardiac risk for noncardiac surgery" and "Perioperative heart failure in noncardiac surgery".)
Initial risk assessment — The initial risk assessment consists of three steps:
- Does the patient have a high-risk condition that is considered a major predictor of risk in the ACC/AHA guidelines on perioperative cardiovascular evaluation and care for noncardiac surgery (not changed in the 2009 focused update) [9]? Such patients require intensive management and often a delay in or cancellation of surgery. (See 'Major predictors' above.)
- What is the surgery-specific risk of the planned operation?
- What is the patient-specific risk?
The indices described above all combine surgery- and patient-specific risks to provide an initial estimate of operative risk. We and others suggest the revised cardiac risk index (RCRI). It is the best validated risk index, is simple to use, and it appears to have greater predictive value than the original Goldman and Detsky risk indices [7,17,69]. However, as discussed above, it may underestimate a patient's true risk. We believe that the use of the NSQIP database risk model, is also a reasonable alternative. (See 'Revised cardiac risk index' above.)
Refinement of initial risk estimate — The initial risk estimate derived from one of the above indices is often adequate to guide decisions regarding perioperative management. However, it is helpful in some cases to refine this estimate with noninvasive stress testing. In selected patients, the results of noninvasive testing will affect both the timing of surgery and the need for preoperative revascularization.
The algorithmic approaches cited above provide strategies to determine which patients require additional risk stratification prior to surgery, with indications for noninvasive testing and cardiac catheterization [9,32,33,35,36].
We prefer either the ACC/AHA guidelines or the Fleisher-Eagle algorithm, because the predictors used largely overlap with the revised cardiac risk index [9,32]. A subsequent algorithm from Auerbach and Goldman is virtually identical to the Fleisher-Eagle algorithm [33]. (See 'Algorithmic approaches' above.)
The Fleisher-Eagle algorithm is easier to use, but refers only to patients who are candidates for surgery [32]. It does not include patients with high-risk conditions that are considered major predictors of risk [32]. (See 'Major predictors' above.)
Stress testing — Recommendations for preoperative stress testing are based upon clinical risk factors, functional capacity, and the surgery-specific risk. The 2007 ACC/AHA guidelines concluded that the evidence was in favor of benefit of preoperative stress testing for [9]:
- Patients deemed to be at high risk (≥3 revised cardiac index criteria) and with poor functional capacity (<4 METS) who are scheduled for vascular surgery when such testing will change management.
The evidence was considered less well established for the benefit of preoperative stress testing in the following circumstances:
- Patients with at least one clinical risk factor and poor functional capacity (<4 METS) who are scheduled for intermediate-risk surgery when such testing will change management
- Patients with at least one clinical risk factor and good functional capacity (≥4 METS) who are scheduled for vascular surgery
Stress testing has a very high negative predictive value for postoperative cardiovascular events (between 90 and 100 percent) but a low positive predictive value (between 6 and 67 percent, with a value of 18 percent in a review of five large studies of thallium perfusion imaging) [2,33]. Thus, stress testing is more useful for reducing estimated risk if negative (or normal) than for identifying patients at very high risk when positive [33].
When stress testing is performed, patients who are able to exercise sufficiently to reach the target heart rate (eg, those who are not limited by claudication, orthopedic degenerative joint disease, or severe deconditioning) can undergo exercise ECG testing often with concurrent imaging to better identify high-risk features that would warrant referral for angiography (eg, reversible large anterior wall defect, multiple reversible defects, ischemia occurring at a low heart rate, extensive stress-induced wall motion abnormalities, transient ischemic dilatation).
In addition, concurrent imaging is essential if the ECG has abnormalities that interfere with interpretation of exercise stress, exercise perfusion imaging or echocardiography. (See "Exercise ECG testing: Performing the test and interpreting the ECG results", section on 'Limitations to exercise ECG testing'.)
Patients who are unable to ambulate sufficiently to increase their cardiovascular workload should undergo pharmacologic stress with either dipyridamole-thallium rMPI or dobutamine stress echocardiography. The choice of test should be based upon local experience and availability and the relative safety of the different procedures in the individual patient [42]. (See 'Choice of stress test' above.)
Resting echocardiography — We suggest resting echocardiography to quantify valvular dysfunction in patients with a murmur, to evaluate ventricular dysfunction in poorly controlled HF or in dyspnea of uncertain cause, or to evaluate for possible pulmonary hypertension [33]. Routine echocardiography is not recommended [9,33]. (See 'Resting echocardiography' above.)
Preoperative revascularization — The proportion of noncardiac surgery patients who undergo preoperative cardiac catheterization and revascularization appears to be very low [69]. In three studies that used the approach in the ACC/AHA guidelines, 2 to 11 percent of patients underwent coronary angiography and only 0 to 2 percent underwent preoperative revascularization [70-72]. The rate of intervention may be even lower in asymptomatic patients, since these studies included symptomatic patients who may have been more likely to undergo revascularization [69].
These findings indicate that there are few asymptomatic intermediate-risk patients who are candidates for preoperative revascularization. (See "Management of cardiac risk for noncardiac surgery", section on 'Revascularization'.)
Urgency of surgery — The above approach to cardiac risk stratification primarily applies to elective surgery. In this setting there is adequate time to complete recommended tests and, if necessary, revascularization procedures. If preoperative PCI is considered, any potential benefit must be balanced against the requirements for a full course of aggressive antiplatelet therapy with aspirin and clopidogrel. Premature discontinuation of antiplatelet therapy carries a substantial risk of stent thrombosis, a risk that may be increased further by surgery. (See "Antiplatelet therapy after coronary artery stenting" and "Management of cardiac risk for noncardiac surgery", section on 'Preoperative PCI'.)
Emergent surgery — Emergent surgery carries unique, often substantial risks. In these cases, risk indices derived from elective surgery cohorts are not accurate, although they may provide an estimate of the minimal risk. Further testing and interventions are unlikely to be beneficial. Despite the elevated risk, patients are usually best served by proceeding directly to surgery. Beta blockers may be helpful in patients who are hemodynamically stable and in whom benefit has been shown. (See "Management of cardiac risk for noncardiac surgery", section on 'Beta blockers'.)
Urgent surgery — For urgent surgery (eg, required during the same admission but a delay of days may be acceptable), initial risk estimates should be made. However, the value of additional testing and treatment is often limited except for identifying and stabilizing patients with unstable cardiac disease [69].
Such patients are usually not candidates for CABG and the need for prolonged dual antiplatelet therapy markedly limits the use of PCI. The main option for risk reduction is a beta blocker, and recommendations are often made on the basis of initial risk indices. If the surgical approach would be altered on the basis of additional risk stratification (eg, surgery would be delayed, canceled, or performed on dual antiplatelet therapy), additional testing and treatment may be warranted.
Searching for
