RSS Feeds
Clinical Guideline: Part II: Risk Stratification after Myocardial Infarction
- Eric D. Peterson, MD, MPH;
- Leslee J. Shaw, PhD; and
- Robert M. Califf, MD
- From Duke University Medical Center, Durham, North Carolina. Acknowledgment: The authors thank Patricia Williams for editorial assistance. Requests for Reprints: Eric D. Peterson, MD, MPH, Duke Clinical Research Institute, Bay A-1, 2024 West Main Street, Durham, NC 27705. Current Author Addresses: Drs. Peterson and Califf: Duke Clinical Research Institute, Bay A-1, 2024 West Main Street, Durham, NC 27705.
Abstract
Purpose: To review the literature on risk stratification after acute myocardial infarction in the reperfusion era and to propose an algorithm for early and continual risk assessment.
Data Sources: A MEDLINE search of the English-language literature on humans was done using the terms myocardial infarction, prospective studies, and prognosis. This search was supplemented by narrowed searches for subheadings (such as cardiogenic shock, thrombolytic therapy, and stress testing) and surveys of references cited in review articles and book chapters.
Study Selection: Literature on prognosis and myocardial infarction published from 1981 to 1996 was considered. From the literature on stress testing methods, studies that enrolled patients before 1980, enrolled patients for indications other than myocardial infarction, tested patients more than 6 weeks after infarction, were missing outcome data, or had inadequate follow-up were excluded.
Data Extraction: Because too few randomized trials were available to allow the cross-comparison of risk-stratification methods, the available observational data were synthesized and supplemented with clinical judgments to produce recommendations.
Data Synthesis: Risk stratification must begin when acute myocardial infarction is diagnosed. High-risk patients (such as those with cardiogenic shock) and candidates for reperfusion therapy must be identified quickly if ideal emergency care is to be given. At specific points during hospitalization, specialized tests may be useful if they add incremental information to the results of clinical evaluations. High-risk patients who have complications after infarction or significant left ventricular dysfunction probably benefit from early angiography; patients without these conditions are at low risk for recurrent events and should have noninvasive stress testing for further risk stratification.
Conclusions: Physicians should continually reappraise risk throughout hospitalization to optimize both patient outcomes and cost containment.
1. Overview and Objectives
The goal of risk stratification after acute myocardial infarction is to identify patients whose outcomes can be improved through specific medical interventions. For example, patients at high risk for recurrent ischemia may benefit from more aggressive medical therapy or revascularization procedures. Other cardiovascular interventions, such as treatment with β-blockers or aspirin, produce equivalent relative risk reductions (qualitative effect) in all risk subsets. Even with these therapies, however, the absolute risk reduction (quantitative effect) varies with the identified risk because fewer high-risk patients must be treated to save the same number of lives. Thus, a central component of an empirical disease-management strategy is the quantification of a patient's short-term and long-term risk. Patients with the highest risk stand to benefit most from aggressive intervention. Care in low-risk patients can be more selective, which limits the use of resources. Although more high-quality information is needed, the available literature can guide clinicians in risk assessment for and management of patients with acute myocardial infarction throughout hospitalization.
Discussion of risk stratification after myocardial infarction has largely focused on stratification immediately before discharge from the hospital [1]. Proposed strategies have incorporated many non-invasive and invasive technologies to identify patients who have the greatest risk for death after discharge. Although the predischarge evaluation remains an important component of risk stratification, emphasis must also be placed on early and continual risk assessment done using simple bedside observations. The reasons for this are many. First, approximately 25% of deaths during the first year after infarction occur within the first 48 hours of hospitalization (Figure 1) [2]. Reperfusion of the infarction-related artery can save lives and reduce long-term complications, but this requires rapid diagnosis and triage in the emergency department. Risk-stratification algorithms that concentrate only on persons who survive the first 5 days of hospitalization miss the opportunity to prevent early deaths.
Focusing on patient assessment solely during the predischarge phase also prevents the resource savings that can result from early and continual risk stratification. The early identification of high-risk patients allows rapid triage to cardiac catheterization, reducing both inpatient stays and overall cost. Low-risk patients, when properly identified by history and clinical course, may be candidates for shorter stays in the intensive care unit, less intensive diagnostic evaluations, and early hospital discharge without impairment in outcomes.
Finally, diagnostic test results should be considered to add incrementally to a provider's pretest impression of risk, which is first formed from the history taken in the emergency department, the physical examination, and the results of electrocardiography. The early occurrence of adverse clinical events, such as recurrent ischemia, arrhythmias, or congestive heart failure, then modifies this assessment. By day 4 to 5 of hospitalization, prognosis can be estimated fairly accurately for many patients before traditional predischarge testing has begun.
We have created an evidence-based risk-stratification algorithm for care after myocardial infarction; this algorithm is based on clinical predictors of short- and long-term outcomes. This information was derived predominantly from subgroup analyses of clinical trials and registries. Such an approach has important limitations. Patients enrolled in clinical trials and registries often do not mirror the general population with myocardial infarction. For example, women, members of ethnic minority groups, and elderly persons have been underrepresented in most study samples [3-5]. Clinical trials have primarily enrolled patients who have not previously had infarction or revascularization and those whose diagnoses were apparent at presentation (for example, those with ST-segment elevation). As noted recently [6], infarction is not diagnosed by initial electrocardiography or enzyme measurements in 25% to 35% of cases. The care received by patients in clinical trials may also differ in type and quality from that provided in general practice. Finally, early studies of new technologies and therapeutic interventions often lack sufficiently large sample sizes and study designs rigorous enough to provide conclusive evidence about utility. If a diagnostic technique is found to lack efficacy on the basis of small studies, this should encourage further investigation. Similarly, the results of small studies purporting to show that a new technology is dramatically better than established techniques must be replicated before the new technology is generally accepted and incorporated into routine clinical use.
We examined the assessment of risk after infarction with regard to current clinical practice. We did not duplicate the efforts of DeBusk [1] that resulted in his review of specialized testing after infarction; instead, we focused on changes in patient risk over time, beginning with the patient's presentation to the medical system and ending with appropriate discharge plans. We emphasized the ways in which various diagnostic methods can assist the clinician with specific therapeutic decisions. Our risk-stratification algorithm has four phases (Figure 2 and Figure 3)-the acute evaluation phase, the cardiac care unit phase, the hospital phase, and the predischarge phase-that reflect important junctures in clinical decision making for patients after infarction.
2. Literature Review
We synthesized the published data on acute myocardial infarction [7]. Our general review identified 267 reports from a combined search done using the following main Medical Subject Heading (MeSH) terms: myocardial infarction, prognosis, and prospective studies. This search was limited to studies done in humans, English-language studies, and updated patient series published from 1981 to 1994. We supplemented the results of this search by scanning the references cited in relevant review articles, meta-analyses, and pooled analyses published from 1991 to 1996 [8-10]. Finally, we completed several selective searches for myocardial infarction combined with more specific subheadings (such as cardiogenic shock).
We also did a specific meta-analysis of the published literature on predischarge noninvasive testing. We identified 115 reports published between 1980 and 1994 from a combined search of 1) exercise test, radionuclide imaging, or echocardiography; 2) myocardial infarction; or 3) prognosis. This search was limited to published reports on humans, English-language studies, and updated patient series. We excluded studies that tested patients more than 6 weeks after infarction (n = 12), enrolled patients before 1980 (n = 10), enrolled mixed populations (such as patients with unstable angina and patients after infarction; n = 3), had a follow-up rate less than 80% (n = 1), or lacked data on cardiac death or reinfarction end points (n = 4).
For analysis of the predischarge stage, we retained 54 of these 115 reports: 28 on exercise electrocardiography, 8 on exercise myocardial perfusion imaging, 10 on exercise ventricular function imaging (8 of these were on radionuclide angiography and 2 were on echocardiography), and 8 on pharmacologic stress imaging (4 of these were on perfusion, 3 were on echocardiography, and 1 was on both). For each article, the quality of patient selection, study administration, handling of withdrawals, outcome measurements, and confounder measurements were independently assessed. From each report, we abstracted event rates for patients with or without a specific criterion for test positivity, forming 2 × 2 frequency tables (for example, for 1-year mortality rates in patients with or without ST-segment depression). Pooled sensitivity, specificity, and positive and negative predictive values were calculated from the summated frequency data [11]. Finally, we compared the positive predictive values of selected test variables for patients who had received thrombolytic therapy compared with patients who had not received initial thrombolytic therapy.
3. Acute Evaluation
3.1 Baseline Clinical and Demographic Risk Factors
Risk stratification of patients who have had acute myocardial infarction begins with the initial history and physical examination done in the emergency department. Table 1 shows the major demographic and clinical predictors of worse short- and long-term prognoses in these patients [12-17]. The most important predictors of death within 30 days are age, systolic blood pressure and heart rate at presentation, evidence of congestive heart failure on physical examination, location of infarction, and previous infarction [12, 18]. The importance of these risk factors has been confirmed in many trials and by expert consensus [14, 17, 19-22].
The importance of age as a risk factor, a finding that has been reported in various community and academic settings [6, 23], cannot be overestimated. For example, one community-based registry [6] showed in-hospital mortality rates ranging from 3% for persons younger than 55 years of age to 28% for persons 85 years of age or older. Advanced age is also a marker for depressed left ventricular function (from previous cardiac events) and more severe underlying coronary disease. Elderly patients are more likely to develop congestive heart failure and cardiogenic shock after infarction [24].
Other factors also increase the likelihood of death. Women have a higher short-term risk for death, but this effect is partly explained by clinical risk factors other than sex, such as advanced age and comorbidity [25-32]. Patients who have previously had an infarction are at increased risk because of residual damage to the left ventricle. The presence of diabetes mellitus, hypertension, or noncardiac vascular disease also increases risk, reflecting the greater burden of severe coronary disease in these patients. Smokers, paradoxically, have a lower short-term risk for death, primarily because they develop coronary disease (and thus have infarctions) at a younger age [33, 34].
The physical examination provides additional information that is crucial to initial risk assessment. Clinical signs of left ventricular dysfunction, including hypotension, tachycardia, S3 gallop, peripheral hypoperfusion, and pulmonary congestion, indicate patients with a high risk for death. The Killip classification, which is based on the presence of pulmonary congestion, peripheral hypoperfusion, or both, is a powerful predictor of outcome [17, 35, 36]. Finally, certain catastrophic complications of myocardial infarction, including acute mitral valve dysfunction and myocardial rupture, can be suspected on the basis of the initial physical examination.
3.2 Baseline Risk Factors Found on Electrocardiography
Findings on initial electrocardiography not only aid in the diagnosis of acute infarction but also contain critical prognostic information. Most important is the presence or absence of ST-segment elevation, which determines whether the patient can benefit from acute reperfusion therapy. The location of the infarction is also important: Of patients with ST-segment elevation, those who have had anterior infarction have twice the risk for death of patients who have had inferior infarction [17, 37, 38]. Quantification of the distribution and extent of ST-segment deviation (elevation or depression measured 60 milliseconds after the J point) makes more accurate risk assessment possible. The benefit of thrombolytic therapy is directly related to the number of electrocardiographic leads with ST-segment elevation and to the overall sum of ST-segment deviations [37, 39]. For example, patients who died within 30 days had a greater absolute sum of ST-segment elevation and depression (median, 16 mm; interquartile range, 10 mm to 23 mm) than did patients who survived for 30 days (median, 12 mm; interquartile range, 8 mm to 18 mm). Similarly, the median number of leads with ST-segment elevations of at least 1 mm was 4 (interquartile range, 3 to 5) for patients who died within 30 days and 3 (inter-quartile range, 3 to 5) for patients who survived for 30 days.
Other electrocardiographic findings, such as conduction abnormalities, high-degree atrioventricular block [40-43], and atrial fibrillation [44], also indicate patients at higher risk for death. Finally, electrocardiography can help distinguish among subgroups of patients at higher risk. For instance, although inferior infarction generally confers a better prognosis than anterior infarction, the incidence rates of congestive heart failure, reinfarction, and death are as much as four times greater in patients who have inferior infarction with significant precordial ST-segment depression or evidence of right ventricular involvement than in patients without these findings [45-48].
3.3 Risk Factors Found through Blood Sampling
Myocardial damage causes the release of enzymes into the circulation; these enzymes can be detected by rapid laboratory assays. Serial testing strategies that use quantitative measures of total creatine kinase and its myocardial band isoenzyme as well as lactate dehydrogenase isoenzyme ratios can confirm or exclude a diagnosis of myocardial infarction within 12 hours of the onset of symptoms [49]. However, these markers are initially elevated in only 34% of patients who have an infarction (poor sensitivity) and are elevated in 12% of patients who do not have an infarction (substantial nonspecificity) [50]. Thus, a single sample value should not be used to exclude the diagnosis of infarction. Recently, investigators have examined the usefulness of other serum markers of myocardial damage for risk prediction [51-56]. In patients without ST-segment elevation, a positive result on a troponin T assay was a strong predictor of in-hospital death (9% for elevated levels compared with 1% for nonelevated levels), shock (6% for elevated levels compared with 2% for nonelevated levels), and congestive heart failure (16% for elevated levels compared with 7% for nonelevated levels).
3.4 Risk Factors Found on Imaging
Imaging methods, including nuclear perfusion studies and echocardiography, have been proposed for use in the emergency department for diagnosis and acute risk assessment in patients who have myocardial infarction [57-79]. In patients receiving thrombolytic therapy, these tests can provide information about reperfusion, infarction size, and myocardial salvage [60, 74, 80]. They can also aid in the diagnosis of acute infarction in subsets of patients for whom diagnosis is unclear, such as those who lack ST-segment elevation [80]. Preliminary investigations [81-86] have also assessed the use of two-dimensional echocardiography for early risk stratification. In one study [86], quantitative echocardiographic assessments of left ventricular function and mitral regurgitation were found to add independent prognostic information to that obtained from the clinical history and electrocardiography in patients who had ischemic syndromes. Another study [83] found that an area of hypokinesis (wall-motion score ≥ 2) detected within 12 hours of the onset of symptoms identified patients at high risk for pump failure, malignant arrhythmias, or death. Left ventricular ejection fraction less than 40% determined within 72 hours of the onset of symptoms was also associated with an increased risk for death [84].
Acute imaging is limited, however, because the images produced cannot be used to distinguish between areas damaged by previous events and those damaged by new ischemia. In addition, the incremental value of the information derived from acute imaging (in addition to that derived from the baseline history, physical examination, and electrocardiographic data) has not been clearly shown. A final practical limitation is that many hospitals do not have trained personnel available to obtain and interpret images at all times.
By using data derived from the initial clinical history, physical examination, and electrocardiography, it is generally possible to categorize patients with acute myocardial infarction as 1) those in cardiogenic shock, 2) those who may be candidates for reperfusion therapy, and 3) those who lack ST-segment elevation or whose presentation is delayed beyond 12 hours after symptom onset and who are thus unlikely to benefit from thrombolysis.
3.5 Cardiogenic Shock
Cardiogenic shock is defined hemodynamically by a combination of features: systolic blood pressure less than 90 mm Hg for 30 minutes or more, an arteriovenous oxygen difference greater than 5.5 mL/dL, a cardiac index less than 2.2 L/min per m2 body surface area, and a pulmonary capillary wedge pressure greater than 18 mm Hg [87, 88]. Because hemodynamic data are not usually available at initial assessment, the triad of hypotension (systolic blood pressure < 90 mm Hg or at least 30 mm Hg less than the usual pressure), evidence of hypoperfusion (oliguria, cyanosis, or altered mental status), and pulmonary edema, all in the absence of other correctable factors (such as volume depletion or sepsis), has been used to establish the clinical diagnosis. The reported incidence of cardiogenic shock among patients with myocardial infarction ranges from 5% to 15% and is associated with mortality rates of 50% to 80% [36, 89-93].
The patient's desires with regard to aggressive care should be ascertained during the initial assessment because prolonged intensive care and high costs are likely for those who survive the first hours of cardiogenic shock. For appropriate candidates, interventions should include urgent consultation with a cardiologist, transfer to an intensive care setting capable of providing hemodynamic monitoring (for example, with a pulmonary artery catheter), and consideration for immediate angiography, intra-aortic balloon-pump placement, and revascularization. Although the routine use of pulmonary artery catheterization is not indicated in myocardial infarction [94-96], invasive hemodynamic monitoring in patients with cardiogenic shock can help maximize the effect of therapeutic interventions [88]. In a nonrandomized comparison [97], the use of intra-aortic balloon-pump therapy decreased in-hospital mortality rates by 33% in 85 patients with cardiogenic shock.
The limited ability of thrombolytic therapy to improve the outcomes of patients who have cardiogenic shock has prompted the investigation of more aggressive revascularization strategies [37, 98, 99]. In uncontrolled studies [100-104], acute reperfusion with emergency percutaneous transluminal coronary angioplasty has been reported to reverse cardiogenic shock and reduce mortality rates by as much as 50% compared with conventional care. The results of the ongoing SHOCK (SHould we revascularize Occluded Coronaries for cardiogenic shocK?) study, a randomized trial of direct angioplasty in patients with cardiogenic shock, should clarify the role of reperfusion in this population.
3.6 Identify Candidates for Acute Reperfusion
With regard to the remaining two risk strata, acute evaluation should next be oriented toward identifying candidates for acute reperfusion therapy. A meta-analysis of randomized thrombolytic trials [98] has found that patients presenting within 12 hours of the onset of chest pain who have bundle-branch block or ST-segment elevation in two contiguous leads clearly benefit from acute thrombolysis. Patients eligible for thrombolysis must be screened quickly for contraindications to this therapy: suspected aortic dissection, pregnancy, bleeding diathesis, recent major bleeding, surgery or trauma, prolonged cardiopulmonary resuscitation, recent stroke (within 6 months), intracranial vascular malformation or neoplasm, and poorly controlled hypertension (systolic pressure ≥ 200 mm Hg or diastolic pressure ≥ 120 mm Hg) [105]. Several recent randomized trials [106-109] have also found that acute angioplasty of the infarction-related artery results in clinical outcomes equivalent to or better than those seen with acute thrombolysis. Patients who have contraindications to thrombolysis may still be candidates for primary angioplasty.
The use of reperfusion therapy in certain patient populations, including elderly persons, persons who have inferior infarction, and persons who present long after the onset of symptoms, has been controversial. The elderly have often been excluded from thrombolytic trials, partly because of concerns about increased risk for intracranial bleeding [110]. However, analyses have shown similar relative risk reductions in older and younger patients [98, 111, 112], a finding that translates to a much greater absolute benefit (or more lives saved per 1000 patients treated) in the elderly.
As noted, inferior infarction generally confers a lower risk for death than anterior infarction; even so, thrombolysis remains associated, in elderly patients, with reduced mortality and improvement in left ventricular function [98, 113]. Finally, although the benefits of reperfusion decrease with increasing delays in therapy, a recent overview of 13 000 patients reported that thrombolysis reduced mortality by 14%, even in patients who presented 7 to 12 hours after the onset of symptoms [98].
3.7 Patients Who Are Not Candidates for Reperfusion
Patients who lack ST-segment elevation (that is, patients who have ST-segment depression or non-specific changes) and those who present more than 12 hours after the onset of symptoms form the last stratum of the population with infarction. Persons in this stratum may represent as many as 50% to 75% of patients with acute infarction, although this representation varies within given populations [6, 114]. Overall in-hospital mortality rates for these patients are slightly less than those for patients with ST-segment elevation; however, the 1-year prognoses of the two groups are similar. Fewer data are available for specific clinical markers of risk in patients who lack ST-segment elevation, partly because such patients have not been enrolled in as many large randomized trials as have patients with ST-segment elevation. The recent Global Use of Strategies to Open Occluded Arteries in Acute Coronary Syndromes (GUSTO)-IIb [115] and Thrombolysis and Thrombin In Myocardial Infarction (TIMI)-9B [116] studies should provide a wealth of information on these subsets of patients.
In terms of acute therapeutic interventions, patients who lack ST-segment elevation do not benefit from and may be harmed by acute thrombolysis [98, 117]. An exception to this is the group of patients with acute isolated posterior infarction, who sometimes present with precordial (leads V1 to V3) ST-segment depression only. The use of primary angioplasty in patients who lack ST-segment elevation or who present late in their clinical course has not been studied in randomized trials and should be reserved for patients with the greatest baseline risk.
4. Cardiac Care Unit Phase
After initial diagnosis and stabilization of the patient, a decision must be made about the portion of the hospital to which the patient will be admitted. Patients with definite acute myocardial infarction should be admitted to a facility, such as a cardiac care unit, that has nursing staff trained to interpret arrhythmias and hemodynamic data and to rapidly defibrillate patients who have ventricular fibrillation. Early comparisons between patients treated in such facilities and historic controls indicate that placement in intensive care units can substantially reduce the rate of early death from myocardial infarction [118, 119]. Consensus guidelines suggest that patients who lack initial enzymatic evidence of myocardial necrosis should be stratified into groups with low, moderate, or high short-term risk on the basis of a combination of clinical and electrocardiographic factors [117]. High-risk patients (those who have unrelieved chest pain and electrocardiographic changes, pulmonary edema, hypotension, or new mitral regurgitation) should probably be managed in an intensive care setting. Those with intermediate or low risk (that is, patients who lack these features) can probably be managed safely in a cardiac “step-down” unit. These units may have lower patient-to-nurse ratios than do intensive care units but still should be capable of prompt defibrillation [120-122].
After admission to the cardiac care or step-down unit, patients should be monitored for a standard set of clinical findings that signal increased risk for death or that demand emergency intervention. These findings include recurrent ischemia [123, 124], reinfarction [125, 126], life-threatening arrhythmias (sustained ventricular tachycardia or fibrillation, high-degree atrioventricular block, or major supraventricular arrhythmias), or clinical evidence of pump dysfunction (rales or hypotension). Patients must be monitored for signs of serious complications of infarction (cardiogenic shock, ventricular septal defect, acute mitral regurgitation, and free-wall rupture) that require emergency treatment. Strokes also occur in 1% to 3% of patients who have infarction. Because many patients also receive anticoagulation therapy and thus are at risk for intracranial hemorrhage, a new focal neurologic deficit is a critical finding and may prompt urgent cranial computed tomography.
4.1 Identify Candidates for Rescue Therapy
The first hours in an intensive care unit should be used to assess the effectiveness of acute interventions. Specifically, although newer thrombolytic regimens achieve patency rates between 55% and 81% at 90 minutes [127], patients who do not have reperfusion have two to three times the mortality rate of patients who have reperfusion [127-129]. Thus, the identification of lack of reperfusion after thrombolysis has considerable prognostic significance. The diagnosis can often be made on clinical grounds. For example, a 20% to 50% resolution in the sum of ST-segment elevation within 90 minutes of the start of therapy corresponds to an 88% probability of reperfusion of the infarction-related artery [41, 130-137]. Improved or resolving chest pain is a much less accurate predictor (observed patency, 60% to 84%) [130]. Continuous electrocardiographic monitoring was recently shown to further increase the ability to detect reperfusion and predict patient outcomes [138, 139].
Determination of whether the infarction-related artery has reperfused may also affect acute clinical management. Several nonrandomized studies [15, 38, 140-146] have investigated the use of emergency or “rescue” angioplasty or repeated thrombolysis for patients in whom initial thrombolysis has failed. The largest randomized trial, the RESCUE trial, randomly assigned 150 patients to angioplasty or conservative treatment after thrombolysis of the left anterior descending coronary artery had failed. Although patients randomly assigned to rescue angioplasty had no improvement in resting left ventricular function, they did have a significantly lower combined rate of death or heart failure than did those who received conservative care [147].
4.2 Assessment at 24 Hours
Approximately 24 hours after admission, a directed assessment should consolidate the findings of the initial course. Serial myocardial enzyme measurements can accurately identify myocardial necrosis within 18 hours of admission, allowing patients without necrosis to be discharged from the cardiac care unit at that point [120, 121, 148]. Most patients with infarction who do not have recurrent ischemia, electrical instability, clinical pump dysfunction, or other serious complications can also be safely moved from the intensive care unit [120]. Nonetheless, Lee and colleagues [149] found that providing prognostic information to physicians failed to shorten stays in the intensive care unit.
5. Hospital Phase
The use of electrocardiographic monitoring in the acute phase of infarction is supported by relatively high event rates, but the effectiveness of continued rhythm monitoring after discharge from the cardiac care unit has not been evaluated. For example, ventricular fibrillation occurs within the first 24 hours in 2.5% of patients who have an infarction (predominately in those with larger infarctions), but the percentage of patients decreases to 0.4% after day 1 [150, 151]. Nevertheless, because of the finite risk for sudden arrhythmic events after the first 24 hours, continued rhythm monitoring for 24 additional hours seems reasonable until more definitive empirical data become available.
5.1 Identify Candidates for Aggressive Treatment
Mark and colleagues [152] developed a risk score based on the presence or absence of early sustained ventricular tachycardia or fibrillation, early sustained hypotension or cardiogenic shock, multivessel coronary artery disease, and an ejection fraction less than 40% in patients who had an infarction and were enrolled in three Thrombolysis and Angioplasty in Myocardial Infarction (TAMI) trials. Patients who had at least one of these risk factors (n = 149) had a 16% risk for a late complication; those who lacked these risk factors (n = 105) had a 6% risk. Mortality and reinfarction rates at 30 days were also significantly higher in the high-risk patients (6% compared with 0%). Others have confirmed these findings in larger patient populations [153]. No randomized trials have been done, but because of their poor prognosis, patients who have recurrent ischemia, electrical instability, mechanical complications (ventricular septal defect, mitral regurgitation, or free-wall rupture), or pump dysfunction should be strongly considered for more aggressive interventions, including cardiac catheterization [19, 117, 152, 154-168].
5.2 Identify Candidates for Early Discharge from the Hospital
At the other end of the spectrum, patients should be continually assessed for the possibility of early discharge; this assessment should be done on the basis of the absence of the above complications. Although length of stay must be determined for each patient individually on the basis of the declining frequency of complications as a function of time, a reasonable approach is to plan the final predischarge risk assessment at about 4 to 5 days after infarction for patients who have had an uncomplicated infarction [153, 169, 170]. This length of stay also creates an opportunity to increase the patient's level of activity (while monitoring for symptoms of ischemia or left ventricular dysfunction) and begin patient education and secondary prevention measures (see Section 6.7).
As noted above, several clinically based risk scores have been developed to identify patients with acute ischemic syndromes who have a low likelihood for acute complications or death [170-173]. Newby and colleagues [153] found that surviving patients in the Global Utilization of Streptokinase and TPA for Occluded Coronary Arteries (GUSTO)-I study who had “uncomplicated” infarctions (defined as the absence of reinfarction, ischemia, stroke, shock, heart failure, bypass surgery, balloon pumping, emergency catheterization, cardioversion, or defibrillation by day 4) comprised more than 57% of the population and had 30-day and 1-year mortality rates of 1% and 3.6%, respectively. Despite this low risk, the median length of stay in the intensive care unit was only slightly shorter in patients who had uncomplicated infarctions (3 compared with 4 days), and total hospital stays were actually the same length in the two groups. In economic terms, total inpatient costs among these patients could have been reduced by 25% by early identification and discharge [174]. Most recently, the Primary Angioplasty in Myocardial Infarction (PAMI)-2 investigators found that in low-risk patients (those <70 years of age who had an ejection fraction >45% and one- or two-vessel disease) who had acute catheterization and primary angioplasty, mortality and reinfarction rates on follow-up were 0.65% and 3.9%, respectively. These findings led these investigators to propose an even earlier discharge (at 3 days) for these low-risk patients [175].
6. Predischarge Risk Stratification
As noted above, many high-risk patients will declare themselves clinically as such during their convalescence. The challenge for the clinician during the predischarge phase is to distinguish the few patients who remain at higher risk from the many relatively low-risk patients. Multiple testing technologies have been developed to aid in this process, but, given the low overall event rate in these patients (1-year mortality rates of 2% to 5%), these tests must be highly sensitive and specific if they are to have clinical value.
Unfortunately, the studies that have evaluated noninvasive tests have contained many methodologic limitations, including a lack of power to examine prognostic end points. For example, of 54 studies of the prognostic use of postinfarction noninvasive stress testing that have been published since 1984, only 41% contained more than 10 mortality end points [7]. Other limitations of these studies include the use of retrospective analysis (determining the criteria for positive test results after patient outcomes are known); selection bias (patients were chosen selectively rather than randomly to receive a given testing method); ascertainment bias (incomplete follow-up); limited time horizons (for example, examining only short-term prognosis); and the use of “soft” end points (including cardiac revascularization), which can be driven by test results. Further, the treating clinicians were not blinded to test results, a fact that altered later patient management and, potentially, the test's prognostic accuracy. For instance, if a test accurately identifies patients who have left main coronary artery disease and these patients then have revascularization (thus improving long-term survival), the test's prognostic value with regard to mortality may diminish. Finally, as noted previously, few studies have determined whether their technologies add information to that obtained from the physicians' clinical assessment.
6.1 Cost Considerations
Before we discuss the relative clinical utility of the various diagnostic tests, it is important for the reader to have information about the economic aspects of predischarge testing. Table 2 lists the institutional and professional charges for these tests according to 1995 Medicare data and hospital fee schedules. Although charges may vary by institution and do not closely reflect the true costs of these tests, the figures are useful for general comparative purposes.
Along with the charge for an individual test, the total charges for a given testing strategy must be considered. Furthermore, both true-positive and false-positive results on initial screening tests will prompt further evaluation, increasing later use of resources. For example, if a physician wishes to assess both left ventricular function and the severity of coronary disease, he or she may do echocardiography and stress myocardial perfusion imaging (institutional charges, $806 and $1350, respectively), rest and exercise echocardiography or radionuclide ventriculography (charges, $846 and $1199, respectively), or cardiac catheterization (charge, $3168). In practice, however, physicians often order a series of tests, the results of which may provide overlapping clinical information. For example, in a national study of Medicare patients [176], 57% of persons having stress testing also had catheterization; 62% of those having catheterization had echocardiography at least once.
Unfortunately, few studies have compared the financial aspects of different management strategies. In a nonrandomized economic analysis of the Primary Angioplasty Registry [177], demographic, clinical, and angiographic factors explained only 19% of the variance in patient care costs. By comparison, a three-stage linear regression model showed that procedure use and complications accounted for 42% of the variance in costs. Charles and colleagues [178] compared inpatient hospital costs in a small substudy (two institutions; 310 patients) of the TIMI-IIb thrombolytic trial. In this report, patients randomly assigned to receive conservative care (predischarge stress radionuclide ventriculography) had fewer coronary interventions (angioplasty rates, 21% compared with 58%) but only minimal cost savings when compared with patients receiving an invasive strategy that included routine catheterization (mean hospital costs, $13 372 compared with $15 325). The invasive strategy was associated with fewer cardiac-related inpatient hospital days during the 1-year follow-up period [179].
6.2 Left Ventricular Function
Mortality and the risk for ischemic events have both been shown to be inversely related to ejection fraction [180-183]. Moreover, the prognostic importance of ejection fraction is equally strong in patients treated with and without thrombolysis (Figure 4) [180, 184]. Finally, randomized trials [185-188] have noted that appropriate medical interventions (such as therapy with angiotensin-converting enzyme inhibitors) may improve the long-term prognosis of patients who have left ventricular dysfunction after infarction. Thus, a clinical assessment of left ventricular function after infarction is indicated for both prognostic and therapeutic reasons.
Many patients who have left ventricular dysfunction are readily identified by physical criteria (such as rales, S3 gallop, or hypotension). Similarly, the electrocardiogram contains important clues to the extent of myocardial damage [189, 190]. Silver and coworkers [191] incorporated both clinical and electrocardiographic data into a clinical prediction rule for identifying patients who have preserved left ventricular function and thus obviating the need for specialized testing. In a retrospective analysis, they found that an interpretable electrocardiogram (lacking indications of left bundle-branch block, paced rhythms, and left ventricular hypertrophy with strain), an absence of previous Q-wave infarction, and an anterior infarction predicted an ejection fraction of 40% or more (positive predictive value, 0.98) for patients who had neither a history of congestive heart failure nor heart failure with the index infarction. This predictive rule was recently validated in 10 756 patients in GUSTO-I. Of the 46% of patients predicted to have preserved left ventricular function, 94% actually had an ejection fraction of 40% or more [174].
For quantitative assessment, patients may have radionuclide angiography, echocardiography, or left ventriculography during cardiac catheterization. These studies also allow for the calculation of ventricular volume, which had a higher predictive value than ejection fraction in two studies [192, 193]. They also provide information on regional wall-motion abnormalities, which may provide additional independent prognostic information. No study has clearly shown one ventricular imaging modality to be superior to another in direct comparisons. Thus, the choice among methods of assessing left ventricular function should be based on institutional availability and expertise, local cost considerations, and the individual clinical situation.
6.3 Risk for Electrical Instability
Several clinical factors denote patients at increased risk for sudden arrhythmic events: left ventricular dysfunction, anterior or Q-wave infarction, a Killip class of III or IV, multivessel coronary disease, and a larger infarction [194-197]. Several noninvasive methods have been proposed to further quantify risk for sudden death. One of the simplest, the presence of complex or frequent ventricular ectopy during electrocardiographic monitoring, has been known for several decades to be a predictor of in-hospital arrhythmic events. For example, an increase in the frequency of ventricular premature complexes, particularly to more than six per hour, appears to increase the risk for late death or recurrent ventricular tachycardia or fibrillation by approximately 60% in many studies [21, 198-203] but not all [150, 194, 204-211]. Similarly, patients who have ventricular fibrillation or sustained tachycardia more than 48 hours after infarction have an increased risk for subsequent sudden death [198, 207, 212, 213].
Signal-averaged electrocardiographic monitoring has been proposed as another noninvasive way to predict arrhythmic events. This procedure detects evidence of delayed conduction (the substrate for sustained reentrant ventricular arrhythmias) and has been reported to be a prognostic risk marker after infarction [210, 214-218]. The subgroups at highest risk include patients who have both an abnormal signal-averaged electrocardiogram and ventricular dysfunction or an abnormal 24-hour Holter recording [209, 210, 219]. Three major limitations inhibit the use of signal-averaged electrocardiography in risk stratification after infarction. First, the results of this procedure vary markedly during the initial days to weeks after infarction. Second, whereas a negative (or normal) test result has a high negative predictive value (>90% likelihood of 1-year event-free survival), the predictive value of a positive result is low (<20%). Finally, therapeutic trials (such as the Cardiac Arrhythmia Suppression Trial [CAST]) have not identified a way to correct abnormalities detected by this test. Thus, the test's ability to alter the management of patients after infarction is currently limited.
Tests for heart rate variability and baroreceptor sensitivity, which are measures of autonomic tone, also appear to identify high-risk populations [198, 220-227]. These tests have been studied in small, selected populations in trials that have not fully assessed the incremental value of the information gained through testing. Further, it remains unclear whether therapeutic interventions done on the basis of an abnormal test result will alter prognosis. The role of cardiac vagal activity in risk after infarction is being investigated in the 1200-patient Autonomic Tone and Reflexes After Myocardial Infarction (ATRAMI) trial. Pending the results of this and other studies, these modalities must be considered investigational.
The inducibility of sustained ventricular tachycardia during electrophysiology, including programmed electrostimulation, has been used for risk stratification among these patients. Unfortunately, nearly half of all reported trials have found programmed electrical stimulation to be ineffective in predicting later mortality or arrhythmic events [228-239]. Moreover, other noninvasive markers are equally effective in identifying patients at risk [237].
6.4 Risk for Myocardial Ischemia
Clearly, the severity of coronary disease and the amount of myocardial tissue at risk for recurrent ischemia have a major influence on long-term prognosis. To assess these risks, clinicians have two major testing strategies: noninvasive stress testing and cardiac catheterization.
Predischarge stress testing can assess the degree to which myocardial ischemia can be provoked. Myocardial ischemia can be assessed by patient symptoms (such as chest pain), electrocardiographic changes (such as ST-segment depression), or hemodynamic changes with exercise (for example, depressed blood pressure at maximal exertion or poor functional status). Imaging modalities-myocardial perfusion imaging, echocardiography, and radionuclide angiography-can provide additional measures of ischemia, including hyperemia after exercise and exercise-induced wall-motion abnormalities. These modalities can also localize and quantify the amount of myocardium at risk for ischemia.
6.5 Meta-Analysis of the Literature on Exercise Stress Testing
We did a meta-analysis of all studies of exercise electrocardiography, exercise perfusion imaging, radionuclide angiography, and stress echocardiography published from 1980 through 1994. Our objective was to assess the prognostic value of specific markers of myocardial ischemia or functional impairment (Table 3 and Table 4) [7]. We also did analyses after stratification by treatment (thrombolytic therapy compared with other therapies) to address the concern that test accuracy may be lower in patients who receive thrombolysis.
Although meta-analytic methods allow the pooling of data, the literature on noninvasive techniques does not entirely lend itself to comparative estimates or to adjustment techniques related to the underlying risk in the population. The pretest risk for death varied threefold among the techniques; it ranged from 2.5% for dipyridamole echocardiography to 9.3% for exercise radionuclide angiography. Recent studies (primarily randomized trials) of patients treated with thrombolysis have also shown much lower mortality rates than have studies of patients not treated with thrombolysis. Because the positive predictive value of a test is strongly influenced by the pretest likelihood or baseline risk for the given event in the given population, the decline in positive predictive values seen in patients treated with thrombolysis (Table 4) reflects changes in the patient population, not changes in the operating characteristics of the tests.
6.5.1 Exercise Electrocardiography
On exercise electrocardiography, indicators that ischemia can be provoked have included significant ST-segment depression [≥ 0.1 mV horizontal or downsloping segments change from rest] and chest pain. Additional indirect measures of ischemia or underlying myocardial functional capacity include 1) impaired systolic blood pressure response to exercise and 2) exercise duration. The pooled prognostic values for each of these markers is shown in Table 3. The predictive ability of electrocardiographic stress testing may vary, however, depending on patient factors (such as age and sex), use of medication (for example, digoxin and β-blockers), and the testing protocol (type and timing and whether the testing is maximal or symptom-limited). Additional electrocardiographic variables, such as the number of Q waves, residual ST-T-wave changes, left ventricular hypertrophy, bundle-branch block, and ventricular rhythms, may further confound the determination of ischemic potential [240, 241].
In general, test results that are influenced by left ventricular function, such as impaired hemodynamic response to exercise or exercise duration, are modestly more sensitive predictors of mortality than are purely ischemic markers (such as ST-segment depression or chest pain). The positive predictive values (percentage of patients who have a positive test result and have an event), however, are low for all of these test results, given the low event rate in the tested populations (overall 1-year mortality rate, 3.3%). Conversely, the negative predictive value of exercise electrocardiography is approximately 0.90.
6.5.2 Exercise Myocardial Perfusion Imaging
Indicators of ischemia during myocardial perfusion imaging have focused on two measures: defect reversibility within and outside of the infarction zone. In our pooled data from 894 patients for prediction of cardiac death or reinfarction, the presence of a reversible perfusion defect was more sensitive than routine exercise electrocardiography, at a cost of reduced specificity (Table 3). A reversible defect outside of the infarction zone has been shown to occur more often in patients with multivessel disease [242-247] and was a predictor of death in one series [248, 249].
6.5.3 Exercise Ventricular Function Imaging
Impaired resting ejection fraction or its correlates can provide substantial information about survival. Several reports [159, 250, 251] have found that peak ejection fraction adds independent incremental information. In our analysis, a peak exercise ejection fraction of 40% or less had the highest positive predictive value (0.27) of all test measures evaluated (Table 3). This is partly explained, however, by the high underlying risk in the test population (1-year mortality rate, 9.3%). Preliminary series have also indicated that the development of new wall-motion abnormalities with stress (assessed by radionuclide angiography or echocardiography) has important diagnostic and prognostic implications, but larger, confirmatory studies are needed.
6.5.4 Pharmacologic Stress Testing
Exercise stress testing can supply information about functional status, but some patients cannot perform an exercise study. Patients who cannot do exercise studies because of cardiovascular limitations are obviously at high risk and should be considered for cardiac catheterization if they are candidates for revascularization. Others will be unable to exercise because of comorbid illness or other factors, such as paraplegia, peripheral vascular disease, or advanced age. For these patients, pharmacologic stress agents (for example, dipyridamole, adenosine, and dobutamine) and imaging protocols (such as thallium imaging, sestamibi imaging, and echocardiography) are available. Pharmacologic studies to date have been small and subject to many of the same biases noted for exercise reports; however, their sensitivity and specificity appear to be similar to those of exercise tests (Table 3).
6.6 Identify Candidates for Predischarge Cardiac Catheterization
Understanding the appropriate use of cardiac catheterization after infarction can be simplified by examining subgroups of patients. First, catheterization (or noninvasive stress testing, for that matter) is not indicated in patients for whom revascularization is not an option because of underlying comorbid conditions or patient preference for conservative care. Second, among patients for whom revascularization is an option, those who have complications after infarction are clearly at high risk and are generally recommended as candidates for catheterization. In contrast, the question of whether to use cardiac catheterization or noninvasive testing in patients who have uncomplicated infarction has been the focus of much controversy in the literature.
6.6.1 Argument for Noninvasive Stress Testing Strategy
Proponents of provocative testing argue that it can identify significant myocardial ischemia and resting left ventricular dysfunction, obviating the need for cardiac catheterization (with its attendant risks and costs) in most patients. The major argument for a strategy of watchful waiting coupled with noninvasive provocative testing after uncomplicated infarction is derived from the aggregate results of several randomized trials. The results of these trials, all of which compared a strategy of routine catheterization and revascularization at different times with a more conservative strategy of catheterization only in patients who had ischemia after infarction, were recently summarized (Table 5) [252]. No survival advantage or reduction in nonfatal reinfarction has been found with the strategy of routine catheterization. Indeed, the TIMI-IIb trial [38] found that the routine catheterization strategy resulted in nearly three times the number of revascularization procedures but similar 1-year mortality and reinfarction rates.
6.6.2 Argument for Routine Catheterization
Proponents of routine cardiac catheterization argue that the accuracy of noninvasive diagnostic testing is low in patients who have had an infarction. Second, although trials have not shown that a strategy of routine cardiac catheterization is beneficial in terms of death or reinfarction, these trials studied patients who received thrombolytic therapy, had an average ejection fraction greater than 50%, and represented only a minority of patients with infarction. In addition, the combined studies are underpowered, potentially missing as much as a 20% risk reduction. Third, the need for readmission and subsequent procedures is higher in patients first treated conservatively [38], especially patients who eventually have non-Q-wave infarction [253]. Finally, Mark and colleagues [254] recently noted that in a nonrandomized comparison of U.S. and Canadian patients treated with thrombolysis (rates of later catheterization, 72% compared with 25%), U.S. patients had fewer cardiac symptoms and better functional status at 1 year. Although this study did not ascribe these benefits to a single aspect of care, it raises the possibility that aggressive intervention may incrementally improve quality of life.
The arguments for and against routine catheterization can be subdivided according to the presence or absence of left ventricular dysfunction. First, patients who have an ejection fraction less than 40% are more likely to have multivessel coronary disease than are those who have more normal left ventricular function, and substantial evidence suggests that coronary artery bypass graft surgery can improve the long-term survival of these patients [255-257]. Although some have argued that only patients with signs or symptoms of ischemia (that is, chest pain) benefit from revascularization, analysis from a combined data set of the seven largest randomized trials comparing bypass surgery with medical therapy [255] does not support this concept. Specifically, in patients who had left ventricular dysfunction and limited anginal symptoms after infarction, long-term mortality rates after bypass surgery (n = 164) were lower than those after medical therapy (n = 165) (Figure 5). Given this evidence, patients who have left ventricular dysfunction and are candidates for revascularization may benefit from routine catheterization.
Patients who show no evidence of global left ventricular dysfunction (those who have an ejection fraction >50%) have a much lower likelihood of multivessel disease and a much lower potential for benefit from revascularization than do patients who have ventricular dysfunction. In fact, routine angiography in these patients will identify many asymptomatic patients who have single-vessel disease and would not be predicted to benefit in terms of survival from revascularization. Thus, a conservative strategy of noninvasive stress testing is likely to provide equivalent outcomes and decreased costs for this group. For example, analysis of clinical trial data indicates that when angiography is done after otherwise uncomplicated infarction, the average length of stay is increased by 1 day.
6.7 Assess the Need To Modify Cardiac Risk Factors
Although the decision to do cardiac catheterization may be controversial, this and other in-hospital decisions represent only a small portion of risk intervention during the life of a patient with coronary artery disease. Given recent evidence about the ability of cardiac risk-factor modification to reduce the occurrence of later cardiac events, screening for the purposes of secondary prevention should be an important component of risk assessment after infarction.
In many patients, a myocardial infarction is the first manifestation of coronary disease. This event, while signaling the presence of disease, also provides an ideal opportunity to motivate patients to make lifestyle changes. Secondary prevention strategies can substantially reduce morbidity and mortality [258]. For example, a meta-analysis of secondary cholesterol intervention trials found that aggressive cholesterol-lowering treatment resulted in a significant overall reduction (22%) in rates of fatal and nonfatal reinfarction [259]. The recent Scandinavian Simvastatin Survival Study [260] found that patients who received a cholesterol-lowering agent compared with patients who received routine medical care had a nearly 70% reduction in major cardiac events.
Although cholesterol levels can be depressed below baseline levels for as long as 2 months after infarction, lipid panels drawn within 24 to 48 hours of infarction accurately reflect lipid status [261]. Similarly, total cholesterol levels elevated beyond this window remain indicative of risk. Thus, we suggest that a lipid panel (including tests for total and high-density lipoprotein cholesterol levels) be drawn at admission from all patients who present within 24 hours of infarction. Treatment of elevated serum cholesterol levels should then begin as outlined by the National Cholesterol Education Program [261].
Patients who can be motivated to quit smoking can achieve benefit within 3 years by reducing their risk for nonfatal infarction to a level similar to that of patients who have never smoked [262]. A pooled analysis of several intervention trials [263] yielded a minimal cost per year of life saved for smoking cessation programs. Measurement of blood pressure and serum glucose levels are part of the routine admission tests for patients who may have had an infarction, and these tests may help to screen for hypertension and diabetes mellitus in this population. Other potential markers of risk, such as levels of certain coagulation factors, can also affect prognosis [264, 265]. Until a successful therapeutic intervention is identified, however, special laboratory tests for heritable clotting abnormalities should be reserved for research settings and unusual clinical situations, such as the evaluation of patients with no known risk factors who present with infarction at a very young age.
Hormonal status may also be an important, potentially modifiable risk factor in women who have myocardial infarction. A pooled analysis of 32 epidemiologic studies [266] showed that postmenopausal estrogen-replacement therapy was associated with a 30% to 40% lower likelihood of cardiovascular events. The cardioprotective benefit of hormone-replacement therapy has been reported to be even greater in women who have already had a cardiovascular event [267]. Although this epidemiologic evidence is strong, it is based predominantly on observational data from nonrandomized trials. Confirmation of these results is needed from randomized clinical trials. Current guidelines therefore suggest that the risks and potential benefits of hormone-replacement therapy be assessed and discussed with female patients [266]. This policy appears to be particularly appropriate for women who have previously had myocardial infarction.
A patient's emotional state and social support network may have important effects on long-term prognosis [268-270]. Lack of social support has been identified as a significant, independent risk factor for death after adjustment for clinical risk factors, coronary anatomy, and left ventricular function [270]. Stress-reduction techniques may also reduce the incidence of subsequent overall cardiac events [271, 272]. Randomized trials (including one of serotonin-reuptake inhibitors) are currently assessing whether pharmacologic interventions for stress and depression after infarction can also reduce this risk. Pending the results of these trials, it may be reasonable to classify clinically depressed or socially isolated patients as higher risk and consider intervention.
Structured cardiac rehabilitation programs (educational programs after discharge that last 8 to 12 weeks) include a comprehensive schedule of aerobic exercise and risk-factor modification. A pooled analysis of numerous studies [273] has shown that patients who exercise after myocardial infarction have a 20% to 25% reduction in mortality compared with controls. In addition to reducing risk, cardiac rehabilitation programs provide substantial improvements in quality of life without a substantial economic burden compared with other health care strategies [274, 275]. A patient's suitability for cardiac rehabilitation can often be evaluated as part of the ischemic risk assessment before discharge.
Finally, an important way to improve long-term prognosis after myocardial infarction is to ensure that patients receive the medical therapies that have been shown to improve survival. These include therapy with aspirin [276, 277], β-blockers [278, 279], and angiotensin-converting enzyme inhibitors (for patients who have left ventricular dysfunction) [185-188]. Although it is natural to assume that nearly all patients who have no contraindications would receive these therapies, recent community-based studies have found contrary results. Ellerbeck and colleagues [280] found that among 16 124 Medicare patients who were considered “ideal” candidates for therapy, only 83% were receiving aspirin and only 45% were receiving β-blockers [280].
7. Conclusions
Assessment of risk after myocardial infarction has traditionally been driven by a narrow view of intervention; interventions have been initiated primarily as a result of manifest signs or symptoms. A multitude of recent studies provides an empirical basis for a new and more comprehensive strategy of early and continual risk assessment, which can offer both physician and patient the benefits of targeted, evidence-based intervention, risk reduction, and enhanced survival based on known and discoverable risk factors as well as clinical manifestations.
Risk assessment should begin with the initial clinical assessment. Many patients further declare themselves clinically to be at high risk during their hospitalization (with recurrent ischemia, congestive failure, malignant ventricular arrhythmias, or other complications of infarction). These patients probably benefit from a more aggressive approach, including catheterization (if they are candidates for revascularization). In contrast, randomized studies have not shown that routine cardiac catheterization is more effective than a conservative strategy for patients who have uncomplicated infarctions. This population can be further stratified, however, by left ventricular function. Those who have an ejection fraction less than 40% have an increased risk for death, often have multivessel coronary disease, and may benefit from cardiac catheterization (if they are candidates for revascularization). Patients who have nearly normal left ventricular function have a very low risk for death and may therefore be candidates for early discharge from the hospital; predischarge stress testing should be used for further risk stratification within this group of patients. Finally, risk stratification, with identification of modifiable cardiac risk factors and initiation of secondary prevention measures where indicated, should continue after discharge.
Dr. Shaw: Duke Clinical Research Institute, Bay C-1, 2024 West Main Street, Durham, NC 27705.
- Copyright ©2004 by the American College of Physicians
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- 126.↵
- 127.↵
- 128.↵
- 129.↵
- 130.↵
- 131.↵
- 132.↵
- 133.↵
- 134.↵
- 135.↵
- 136.↵
- 137.↵
- 138.↵
- 139.↵
- 140.↵
- 141.↵
- 142.↵
- 143.↵
- 144.↵
- 145.↵
- 146.↵
- 147.↵
- 148.↵
- 149.↵
- 150.↵
- 151.↵
- 152.↵
- 153.↵
- 154.↵
- 155.↵
- 156.↵
- 157.↵
- 158.↵
- 159.↵
- 160.↵
- 161.↵
- 162.↵
- 163.↵
- 164.↵
- 165.↵
- 166.↵
- 167.↵
- 168.↵
- 169.↵
- 170.↵
- 171.↵
- 172.↵
- 173.↵
- 174.↵
- 175.↵
- 176.↵
- 177.↵
- 178.↵
- 179.↵
- 180.↵
- 181.↵
- 182.↵
- 183.↵
- 184.↵
- 185.↵
- 186.↵
- 187.↵
- 188.↵
- 189.↵
- 190.↵
- 191.↵
- 192.↵
- 193.↵
- 194.↵
- 195.↵
- 196.↵
- 197.↵
- 198.↵
- 199.↵
- 200.↵
- 201.↵
- 202.↵
- 203.↵
- 204.↵
- 205.↵
- 206.↵
- 207.↵
- 208.↵
- 209.↵
- 210.↵
- 211.↵
- 212.↵
- 213.↵
- 214.↵
- 215.↵
- 216.↵
- 217.↵
- 218.↵
- 219.↵
- 220.↵
- 221.↵
- 222.↵
- 223.↵
- 224.↵
- 225.↵
- 226.↵
- 227.↵
- 228.↵
- 229.↵
- 230.↵
- 231.↵
- 232.↵
- 233.↵
- 234.↵
- 235.↵
- 236.↵
- 237.↵
- 238.↵
- 239.↵
- 240.↵
- 241.↵
- 242.↵
- 243.↵
- 244.↵
- 245.↵
- 246.↵
- 247.↵
- 248.↵
- 249.↵
- 250.↵
- 251.↵
- 252.↵
- 253.↵
- 254.↵
- 255.↵
- 256.↵
- 257.↵
- 258.↵
- 259.↵
- 260.↵
- 261.↵
- 262.↵
- 263.↵
- 264.↵
- 265.↵
- 266.↵
- 267.↵
- 268.↵
- 269.↵
- 270.↵
- 271.↵
- 272.↵
- 273.↵
- 274.↵
- 275.↵
- 276.↵
- 277.↵
- 278.↵
- 279.↵
- 280.↵
