American College of Cardiology

BRAUNWALD ET AL., MANAGEMENT OF PATIENTS WITH UNSTABLE ANGINA AND NON-ST-SEGMENT ELEVATION MYOCARDIAL INFARCTION UPDATE
http://www.acc.org/clinical/guidelines/unstable/incorporated/index.htm

ACC/AHA 2002 Guideline Update for the Management of Patients With Unstable Angina and Non-ST-Segment Elevation Myocardial Infarction

A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on the Management of Patients With Unstable Angina)

III. Hospital Care

Patients with UA/NSTEMI, recurrent symptoms and/or ECG ST-segment deviations, or positive cardiac markers who are stable hemodynamically should be admitted to an inpatient unit with continuous rhythm monitoring and careful observation for recurrent ischemia (a step-down unit) and managed according to the acute ischemia pathway (Figure 7). Patients with continuing discomfort and/or hemodynamic instability should be hospitalized for at least 24 h in a coronary care unit characterized by a nursing-to-patient ratio sufficient to provide 1) continuous rhythm monitoring, 2) frequent assessment of vital signs and mental status, 3) documented ability to perform defibrillation quickly after the onset of ventricular fibrillation, and 4) adequate staff to perform these functions. Patients should be maintained at that level of care until they have been observed for at least 24 h without any of the following major complications: sustained ventricular tachycardia or fibrillation, sinus tachycardia, atrial fibrillation or flutter, high-degree AV block, sustained hypotension, recurrent ischemia documented by symptoms or ST-segment change, new mechanical defect (ventricular septal defect or MR), or CHF.

Once a patient with documented high-risk ACS is admitted, standard medical therapy is indicated as discussed later. Unless a contraindication exists, these patients should be treated with aspirin (ASA), a beta-blocker, antithrombin therapy, and a GP IIb/IIIa inhibitor. Furthermore, critical decisions are required regarding the angiographic strategy. One option is a routine angiographic approach in which coronary angiography and revascularization are performed unless a contraindication exists. Within this approach, the most common strategy has called for a period of medical stabilization. Some physicians are now taking a more aggressive approach, with coronary angiography and revascularization performed within 24 h of admission; the rationale for the more aggressive approach is the protective effect of carefully administered antithrombin and antiplatelet therapy on procedural outcome. The alternative approach, commonly referred to as the "initially conservative strategy" (see Section III. D), is guided by ischemia, with angiography reserved for patients with recurrent ischemia or a "high-risk" stress test despite medical therapy. Regardless of the angiographic strategy, an assessment of LV function should be strongly considered in patients with documented ischemia because of the imperative to treat patients who have impaired LV function with angiotensin-converting enzyme (ACE) inhibitors (ACEIs) and beta-blockers and, when the coronary anatomy is appropriate (e.g., 3-vessel coronary disease), with CABG (see Section IV). When the coronary angiogram is obtained, a left ventriculogram can be obtained at the same time. When coronary angiography is not scheduled, echocardiography or nuclear ventriculography can be used to evaluate LV function.

A. Anti-Ischemic Therapy

Recommendations for Anti-Ischemic Therapy

Class I

1. Bed rest with continuous ECG monitoring for ischemia and arrhythmia detection in patients with ongoing rest pain. (Level of Evidence: C)
2. NTG, sublingual tablet or spray, followed by intravenous administration, for the immediate relief of ischemia and associated symptoms. (Level of Evidence: C)
3. Supplemental oxygen for patients with cyanosis or respiratory distress; finger pulse oximetry or arterial blood gas determination to confirm adequate arterial oxygen saturation (Sao₂ greater than 90%) and continued need for supplemental oxygen in the presence of hypoxemia. (Level of Evidence: C)
4. Morphine sulfate intravenously when symptoms are not immediately relieved with NTG or when acute pulmonary congestion and/or severe agitation is present. (Level of Evidence: C)
5. A beta-blocker, with the first dose administered intravenously if there is ongoing chest pain, followed by oral administration, in the absence of contraindications. (Level of Evidence: B)
6. In patients with continuing or frequently recurring ischemia when beta-blockers are contraindicated, a nondihydropyridine calcium antagonist (e.g., verapamil or diltiazem) as initial therapy in the absence of severe LV dysfunction or other contraindications. (Level of Evidence: B)
7. An ACEI when hypertension persists despite treatment with NTG and a beta-blocker in patients with LV systolic dysfunction or CHF and in ACS patients with diabetes. (Level of Evidence: B)

Class IIa

1. Oral long-acting calcium antagonists for recurrent ischemia in the absence of contraindications and when beta-blockers and nitrates are fully used. (Level of Evidence: C)
2. An ACEI for all post-ACS patients. (Level of Evidence: B)
3. Intra-aortic balloon pump (IABP) counterpulsation for severe ischemia that is continuing or recurs frequently despite intensive medical therapy or for hemodynamic instability in patients before or after coronary angiography. (Level of Evidence: C)

Class IIb

1. Extended-release form of nondihydropyridine calcium antagonists instead of a beta-blocker. (Level of Evidence: B)
2. Immediate-release dihydropyridine calcium antagonists in the presence of a beta-blocker. (Level of Evidence: B)

Class III

1. NTG or other nitrate within 24 h of sildenafil (Viagra) use. (Level of Evidence: C)
2. Immediate-release dihydropyridine calcium antagonists in the absence of a beta-blocker. (Level of Evidence: A)

The optimal management of UA/NSTEMI has the twin goals of the immediate relief of ischemia and the prevention of serious adverse outcomes (i.e., death or MI/[re]infarction). This is best accomplished with an approach that includes anti-ischemic therapy (Table 10), antiplatelet and antithrombotic therapy (see also Table 14), ongoing risk stratification, and the use of invasive procedures. Patients who are at intermediate or high risk for adverse outcome, including those with ongoing ischemia refractory to initial medical therapy and those with evidence of hemodynamic instability, should if possible be admitted to a critical care environment with ready access to invasive cardiovascular diagnosis and therapy procedures. Ready access is defined as ensured, timely access to a cardiac catheterization laboratory with personnel who have credentials in invasive coronary procedures, as well as to emergency or urgent cardiac surgery, vascular surgery, and cardiac anesthesia ⁽¹³²⁾.

The approach to the achievement of the twin goals described here includes the initiation of pharmacological management and planning of a definitive treatment strategy for the underlying disease process. Most patients are stable at presentation or stabilize after a brief period of intensive pharmacological management and, after appropriate counseling, will be able to participate in the choice of an approach for definitive therapy. A few patients will require prompt triage to emergency or urgent cardiac catheterization and/or the placement of an IABP. Some will prefer the continuation of a medical regimen without angiography. These patients require careful monitoring of the response to initial therapy with noninvasive testing and surveillance for persistent or recurrent ischemia, hemodynamic instability, or other features that may dictate a more invasive approach. Other patients prefer a more invasive strategy that involves early coronary angiography with a view toward revascularization.

1. General Care
The severity of symptoms dictates some of the general care that should be given during the initial treatment. Patients should be placed at bed rest while ischemia is ongoing but can be mobilized to a chair and bedside commode when symptom free. Subsequent activity should not be inappropriately restrictive; instead, it should be focused on the prevention of recurrent symptoms and liberalized as judged appropriate when response to treatment occurs. Patients with cyanosis, respiratory distress, or other high-risk features should receive supplemental oxygen. Adequate arterial oxygen saturation should be confirmed with direct measurement or pulse oximetry. No evidence is available to support the administration of oxygen to all patients with ACS in the absence of signs of respiratory distress or arterial hypoxemia. Oxygen use during initial evaluation should be limited to patients with questionable respiratory status or those with documented hypoxemia, because it consumes resources and evidence for its routine use is lacking. Inhaled oxygen should be administered if the arterial oxygen saturation (Sao₂) declines to less than 90%. Finger pulse oximetry is useful for the continuous monitoring of Sao₂ but is not mandatory in patients who do not appear to be at risk of hypoxia. Patients should undergo continuous ECG monitoring during their ED evaluation and early hospital phase, because sudden, unexpected ventricular fibrillation is the major preventable cause of death in this early period. Furthermore, monitoring for the recurrence of ST-segment shifts provides useful diagnostic and prognostic information, although the system of monitoring for ST-segment shifts must include specific methods intended to provide stable and accurate recordings.

2. Use of Anti-Ischemic Drugs
a. Nitrates
NTG reduces myocardial oxygen demand while enhancing myocardial oxygen delivery. NTG, an endothelium-independent vasodilator, has both peripheral and coronary vascular effects. By dilating the capacitance vessels (i.e., the venous bed), it increases venous pooling to decrease myocardial preload, thereby reducing ventricular wall tension, a determinant of myocardial oxygen consumption (MVO₂). More modest effects on the arterial circulation decrease systolic wall stress (afterload), contributing to further reductions in MVO₂. This decrease in myocardial oxygen demand is in part offset by reflex increases in heart rate and contractility, which counteract the reductions in MVO₂ unless a beta-blocker is concurrently administered. NTG dilates normal and atherosclerotic epicardial coronary arteries as well as smaller arteries that constrict with certain stressors (e.g., cold, mental or physical exercise). With severe atherosclerotic coronary obstruction and with less severely obstructed vessels with endothelial dysfunction, physiological responses to changes in myocardial blood flow are often impaired (i.e., loss of flow-mediated dilation), so maximal dilation does not occur unless a direct-acting vasodilator like NTG is administered. Thus, NTG promotes the dilation of large coronary arteries as well as collateral flow and redistribution of coronary blood flow to ischemic regions. Inhibition of platelet aggregation also occurs with NTG ⁽¹³³⁾, but the clinical significance of this action is not well defined.

Patients whose symptoms are not relieved with three 0.4-mg sublingual NTG tablets or spray taken 5 min apart (Table 11) and the initiation of an intravenous beta-blocker (when there are no contraindications), as well as all nonhypotensive high-risk patients (Table 6), may benefit from intravenous NTG, and such therapy is recommended in the absence of contraindications (i.e., the use of sildenafil [Viagra] within the previous 24 h or hypotension). Sildenafil inhibits the phosphodiesterase (PDE5) that degrades cyclic guanosine monophosphate (cGMP), and cGMP mediates vascular smooth muscle relaxation by nitric oxide. Thus, NTG-mediated vasodilatation is markedly exaggerated and prolonged in the presence of sildenafil. Nitrate use within 24 h after sildenafil or the administration of sildenafil in a patient who has received a nitrate within 24 h has been associated with profound hypotension, MI, and even death ⁽¹³⁴⁾.

Intravenous NTG may be initiated at a rate of 10 mcg per min through continuous infusion with nonabsorbing tubing and increased by 10 mcg per min every 3 to 5 min until some symptom or blood pressure response is noted. If no response is seen at 20 mcg per min, increments of 10 and, later, 20 mcg per min can be used. If symptoms and signs of ischemia are relieved, there is no need to continue to increase the dose to effect a blood pressure response. If symptoms and signs of ischemia are not relieved, the dose should be increased until a blood pressure response is observed. Once a partial blood pressure response is observed, the dosage increase should be reduced and the interval between increments should be lengthened. Side effects include headache and hypotension. Caution should be used when systolic blood pressure falls to less than 110 mm Hg in previously normotensive patients or to greater than 25% below the starting mean arterial blood pressure if hypertension was present. Although recommendations for a maximal dose are not available, a ceiling of 200 mcg per min is commonly used. Prolonged (2 to 4 weeks) infusion at 300 to 400 mcg per h does not increase methemoglobin levels ⁽¹³⁵⁾.

Topical or oral nitrates are acceptable alternatives for patients without ongoing refractory symptoms. Tolerance to the hemodynamic effects of nitrates is dose and duration dependent and typically becomes important after 24 h of continuous therapy with any formulation. Patients who require continued intravenous NTG beyond 24 h may require periodic increases in infusion rate to maintain efficacy. An effort must be made to use non-tolerance-producing nitrate regimens (lower dose and intermittent dosing). When patients have been free of pain and other manifestations of ischemia for 12 to 24 h, an attempt should be made to reduce the dose of intravenous NTG and to switch to oral or topical nitrates. It is not appropriate to continue intravenous NTG in patients who remain free of signs and symptoms of ischemia. When ischemia recurs during continuous intravenous NTG therapy, responsiveness to nitrates can often be restored by increasing the dose and then, after symptoms have been controlled for several hours, attempting to add a nitrate-free interval. This strategy should be pursued as long as symptoms are not adequately controlled. In stabilized patients, intravenous NTG should generally be converted within 24 h to a nonparenteral alternative (Table 11) administered in a non-tolerance-producing regimen to avoid the potential reactivation of symptoms.

Most studies of nitrate treatment in UA have been small and uncontrolled, and there are no randomized, placebo-controlled trials that address either symptom relief or reduction in cardiac events. One small, randomized trial compared intravenous NTG with buccal NTG and found no significant difference in the control of ischemia ⁽¹³⁶⁾. An overview of small studies of NTG in AMI from the prethrombolytic era suggested a 35% reduction in mortality rates ⁽¹³⁷⁾, although both the Fourth International Study of Infarct Survival (ISIS-4) ⁽¹³⁸⁾ and Gruppo Italiano per lo Studio della Sopravvivenza nell'infarto Miocardico (GISSI-3) ⁽¹³⁹⁾ trials formally tested this hypothesis in patients with suspected AMI and failed to confirm this magnitude of benefit. However, these large trials are confounded by frequent prehospital and hospital use of NTG in the "control" groups. The abrupt cessation of intravenous NTG has been associated with exacerbation of ischemic changes on the ECG ⁽¹⁴⁰⁾, and a graded reduction in the dose of intravenous NTG is advisable.

Thus, the rationale for NTG use in UA is extrapolated from pathophysiological principles and extensive, although uncontrolled, clinical observations ⁽¹³³⁾.

b. Morphine Sulfate
Morphine sulfate (1 to 5 mg intravenously [IV]) is recommended for patients whose symptoms are not relieved after 3 serial sublingual NTG tablets or whose symptoms recur despite adequate anti-ischemic therapy. Unless contraindicated by hypotension or intolerance, morphine may be administered along with intravenous NTG, with careful blood pressure monitoring, and may be repeated every 5 to 30 min as needed to relieve symptoms and maintain patient comfort.

Morphine sulfate has potent analgesic and anxiolytic effects, as well as hemodynamic effects that are potentially beneficial in UA/NSTEMI. No randomized trials have defined the unique contribution of morphine to the initial therapeutic regimen or its optimal administration schedule. Morphine causes venodilation and may produce modest reductions in heart rate (through increased vagal tone) and systolic blood pressure to further reduce myocardial oxygen demand. The major adverse reaction to morphine is an exaggeration of its therapeutic effect, causing hypotension, especially in the presence of volume depletion and/or vasodilator therapy. This reaction usually responds to supine or Trendelenburg positioning or intravenous saline boluses and atropine when accompanied by bradycardia; it rarely requires pressors or naloxone to restore blood pressure. Nausea and vomiting occur in approximately 20% of patients. Respiratory depression is the most serious complication of morphine; severe hypoventilation that requires intubation occurs very rarely in patients with UA/NSTEMI treated with this agent. Naloxone (0.4 to 2.0 mg IV) may be administered for morphine overdose with respiratory and/or circulatory depression. Meperidine hydrochloride can be substituted in patients who are allergic to morphine.

c. Beta-Adrenergic Blockers
Beta-blockers competitively block the effects of catecholamines on cell membrane beta-receptors. Beta1-adrenergic receptors are located primarily in the myocardium; inhibition of catecholamine action at these sites reduces myocardial contractility, sinus node rate, and AV node conduction velocity. Through this action, they blunt the heart rate and contractility responses to chest pain, exertion, and other stimuli. They also decrease systolic blood pressure. All of these effects reduce MVO₂. Beta2-adrenergic receptors are located primarily in vascular and bronchial smooth muscle; the inhibition of catecholamine action at these sites produces vasoconstriction and bronchoconstriction ⁽¹⁴¹⁾. In UA/NSTEMI, the primary benefits of beta-blockers are due to effects on beta1-adrenergic receptors that decrease cardiac work and myocardial oxygen demand. Slowing of the heart rate also has a very favorable effect, acting not only to reduce MVO₂ but also to increase the duration of diastole and diastolic pressure-time, a determinant of coronary flow and collateral flow.

Beta-blockers should be started early in the absence of contraindications. These agents should be administered intravenously followed by oral administration in high-risk patients as well as in patients with ongoing rest pain or orally for intermediate- and low-risk patients (Table 6).

The choice of beta-blocker for an individual patient is based primarily on pharmacokinetic and side effect criteria, as well as on physician familiarity (Table 12). There is no evidence that any member of this class of agents is more effective than another, except that beta-blockers without intrinsic sympathomimetic activity are preferable. The initial choice of agents includes metoprolol, propranolol, or atenolol. Esmolol can be used if an ultrashort-acting agent is required.

Patients with marked first-degree AV block (i.e., ECG PR interval [PR] of greater than 0.24 s), any form of second- or third-degree AV block in the absence of a functioning pacemaker, a history of asthma, or severe LV dysfunction with CHF should not receive beta-blockers on an acute basis ⁽²⁶⁾. Patients with significant sinus bradycardia (heart rate less than 50 bpm) or hypotension (systolic blood pressure less than 90 mm Hg) generally should not receive beta-blockers until these conditions have resolved. Patients with significant COPD who may have a component of reactive airway disease should be administered beta-blockers very cautiously; initially, low doses of a beta1-selective agent should be used. If there are concerns about possible intolerance to beta-blockers, initial selection should favor a short-acting beta1-specific drug such as metoprolol. Mild wheezing or a history of COPD mandates a short-acting cardioselective agent at a reduced dose (e.g., 2.5 mg metoprolol IV or 12.5 mg metoprolol orally or 25 mcg · kg^-1 · min^-1 esmolol IV as initial doses) rather than the complete avoidance of a beta-blocker.

In the absence of these concerns, several regimens may be used. For example, intravenous metoprolol may be given in 5-mg increments by slow intravenous administration (5 mg over 1 to 2 min), repeated every 5 min for a total initial dose of 15 mg. In patients who tolerate the total 15 mg IV dose, oral therapy should be initiated 15 min after the last intravenous dose at 25 to 50 mg every 6 h for 48 h. Thereafter, patients should receive a maintenance dose of 100 mg twice daily. Alternatively, intravenous propranolol is administered as an initial dose of 0.5 to 1.0 mg, followed in 1 to 2 h by 40 to 80 mg by mouth every 6 to 8 h. Intravenous esmolol is administered as a starting dose of 0.1 mg · kg^-1 · min^-1 with titration in increments of 0.05 mg · kg^-1 · min^-1 every 10 to 15 min as tolerated by the patient's blood pressure until the desired therapeutic response has been obtained, limiting symptoms develop, or a dosage of 0.3 mg · kg^-1 · min^-1 is reached. A loading dose of 0.5 mg per kg may be given by slow intravenous administration (2 to 5 min) for a more rapid onset of action. In patients suitable to receive a longer-acting agent, intravenous atenolol can be initiated with a 5-mg IV dose followed 5 min later by a second 5-mg IV dose and then 50 to 100 mg per day orally initiated 1 to 2 h after the intravenous dose. Monitoring during intravenous beta-blocker therapy should include frequent checks of heart rate and blood pressure and continuous ECG monitoring, as well as auscultation for rales and bronchospasm.

After the initial intravenous load, patients without limiting side effects may be converted to an oral regimen. The target resting heart rate is 50 to 60 bpm, unless a limiting side effect is reached. Selection of the oral agent should be based on the clinician's familiarity with the agent. Maintenance doses are given in Table 12.

Initial studies of beta-blocker benefits in ACS were small and uncontrolled. An overview of double-blind, randomized trials in patients with threatening or evolving MI suggests an approximately 13% reduction in the risk of progression to AMI ⁽¹⁴²⁾. These trials lack sufficient power to assess the effects of these drugs on mortality rates for UA. However, randomized trials with other CAD patients (AMI, recent MI, stable angina with daily life ischemia, and heart failure) have all shown reductions in mortality and/or morbidity rates. Thus, the rationale for beta-blocker use in all forms of CAD, including UA, is very compelling and in the absence of contraindications is sufficient to make beta-blockers a routine part of care, especially in patients who are to undergo cardiac or noncardiac surgery.

In conclusion, evidence for the beneficial effects of the use of beta-blockers in patients with UA is based on limited randomized trial data, along with pathophysiological considerations and extrapolation from experience with CAD patients who have other types of ischemic syndromes (stable angina, AMI, or heart failure). The recommendation for the use of intravenous beta-blockers in high-risk patients with evolving pain is based on the demonstrated benefit in AMI patients, as well as the hemodynamic objectives to reduce cardiac work and myocardial oxygen demand. The duration of benefit with long-term oral therapy is uncertain.

d. Calcium Antagonists
These agents reduce cell transmembrane inward calcium flux, which inhibits both myocardial and vascular smooth muscle contraction; some also slow AV conduction and depress sinus node impulse formation. Agents in this class vary in the degree to which they produce vasodilation, decreased myocardial contractility, AV block, and sinus node slowing. Nifedipine and amlodipine have the most peripheral arterial dilatory effect but little or no AV or sinus node effects, whereas verapamil and diltiazem have prominent AV and sinus node effects and some peripheral arterial dilatory effects as well. All 4 of these agents, as well as the newer agents, have coronary dilatory properties that appear to be similar. Although different members of this class of agents are structurally diverse and may have somewhat different mechanisms of action, no reliable data demonstrate the superiority of 1 agent (or groups of agents) over another in ACS, except for the risks posed by rapid-release, short-acting dihydropyridines (Table 13). Beneficial effects in ACS are believed to be due to variable combinations of decreased myocardial oxygen demand that relate to decreased afterload, contractility, and heart rate and improved myocardial flow that relates to coronary artery and arteriolar dilation ^(141,143). These agents also have theoretical beneficial effects on LV relaxation and arterial compliance. Major side effects include hypotension, worsening CHF, bradycardia, and AV block.

Calcium antagonists may be used to control ongoing or recurring ischemia-related symptoms in patients who are already receiving adequate doses of nitrates and beta-blockers, in patients who are unable to tolerate adequate doses of 1 or both of these agents, or in patients with variant angina (see Section VI. F). In addition, these drugs have been used for the management of hypertension in patients with recurrent UA ⁽¹⁴³⁾. Rapid-release, short-acting dihydropyridines (e.g., nifedipine) must be avoided in the absence of adequate concurrent beta-blockade in ACS because controlled trials suggest increased adverse outcomes ^(144-146). Verapamil and diltiazem should be avoided in patients with pulmonary edema or evidence of severe LV dysfunction ^(147,148). Amlodipine and felodipine, however, appear to be well tolerated by patients with chronic LV dysfunction ⁽¹⁴⁹⁾. The choice of an individual calcium antagonist is based primarily on the type of agent; the hemodynamic state of the patient; the risk of adverse effects on cardiac contractility, AV conduction, and sinus node function; and the physician's familiarity with the specific agent. Trials in patients with acute CAD suggest that verapamil and diltiazem are preferred if a calcium antagonist is needed ^(148,149).

There are several randomized trials that involve the use of calcium antagonists in ACS. Results generally confirm that these agents relieve or prevent symptoms and related ischemia to a degree similar to that of beta-blockers. The largest randomized trial is the Danish Study Group on Verapamil in Myocardial Infarction (DAVIT) ^(150,151), in which 3,447 patients with suspected ACS were administered intravenous verapamil (0.1 mg per kg) at admission and then 120 mg 3 times daily vs. placebo. After 1 week, verapamil was discontinued in the patients (n = 2,011) without confirmed MI (presumably many of these patients had UA). Although there was no definitive evidence to suggest benefit (or harm) in this cohort, trends favored a reduction in the outcome of death or nonfatal MI. In the Holland Interuniversity Nifedipine/metoprolol Trial (HINT), nifedipine and metoprolol were tested in a 2 × 2 factorial design in 515 patients ⁽¹⁴⁶⁾. Nifedipine alone increased the risk of MI or recurrent angina relative to placebo by 16%, metoprolol decreased it by 24%, and the combination of metoprolol and nifedipine reduced this outcome by 20%. None of these effects, however, were statistically significant because the study was stopped early because of concern for harm with the use of nifedipine alone. However, in patients already taking a beta-blocker, the addition of nifedipine appeared favorable because the event rate was reduced significantly (risk ratio [RR] 0.68) ⁽¹⁵²⁾. Several meta-analyses that combined all of the calcium antagonists used in UA trials suggested no overall effect ^(142,153). However, in light of the aforementioned differences between the rapid-release dihydropyridines and the heart rate-slowing agents diltiazem and verapamil, such analyses are not appropriate. When the data for verapamil are considered alone, a beneficial effect in patients with ACS is apparent ⁽¹⁵⁰⁾.

Similarly, in the Diltiazem Reinfarction Study (DRS), 576 patients were administered diltiazem or placebo 24 to 72 h after the onset of non-Q-wave MI ⁽¹⁴⁵⁾. Diltiazem was associated with a reduction in CK-MB level-confirmed reinfarction and refractory angina at 14 days without a significant increase in mortality rates. Retrospective analysis of the non-Q-wave MI subset of patients in the Multicenter Diltiazem Postinfarction Trial (MDPIT) suggested similar findings without evidence of harm ⁽¹⁵⁴⁾.

However, retrospective analyses of DAVIT and MDPIT suggested that the administration of verapamil and diltiazem to suspected AMI patients who have LV dysfunction (many of whom had UA/NSTEMI) may have an overall detrimental effect on mortality rates ^(145,147). Although this caution is useful for clinical practice, more recent data suggest that this issue should be readdressed. For example, in DAVIT-2, verapamil was associated with a significant reduction in diuretic use compared with placebo ⁽¹⁵⁵⁾, suggesting that it did not further impair LV function. Furthermore, recent prospective trials with verapamil administered to AMI patients with heart failure who were receiving an ACEI strongly suggest benefit ^(147,156). The Diltiazem as Adjunctive Therapy to Activase (DATA) trial also suggests that intravenous diltiazem in AMI patients may be safe; death, MI, or recurrent ischemia decreased by 14% at 35 days and death, MI, or refractory ischemia decreased by 23% at 6 months ⁽¹⁵⁷⁾. These pilot data were confirmed in 874 AMI patients without heart failure in whom long-acting diltiazem (300 mg per day) was administered 36 to 96 h after thrombolysis ⁽¹⁵⁸⁾ (W.E. Boden, oral presentation, American Heart Association Scientific Sessions, Dallas, Texas, November 1998).

In conclusion, definitive evidence for benefit with all calcium antagonists in UA is predominantly limited to symptom control. For dihydropyridines, available randomized trial data are not consistent with a beneficial effect on mortality or recurrent infarction rates but in fact provide strong evidence for an increase in these serious events when they are administered early as a rapid-release, short-acting preparation without a beta-blocker. Thus, these guidelines recommend reservation of the dihydropyridine calcium antagonists as second or third choices after the initiation of nitrates and beta-blockers. For the heart rate-slowing drugs (verapamil and diltiazem), there is no controlled trial evidence for harm when they are administered early to patients with acute ischemic syndromes, and strong trends suggest a beneficial effect. Therefore, when beta-blockers cannot be used, heart rate-slowing calcium antagonists offer an alternative. When required for refractory symptom control, these agents can be used early during the hospital phase, even in patients with mild LV dysfunction, although the combination of a beta-blocker and calcium antagonist may act in synergy to depress LV function. The risks and benefits in UA of amlodipine and other newer agents relative to the older agents in this class reviewed here remain undefined.

e. Other
ACEIs have been shown to reduce mortality rates in patients with AMI or who recently had an MI and have LV systolic dysfunction ^{(159-161,161a)}, in diabetic patients with LV dysfunction ⁽¹⁶²⁾, and in a broad spectrum of patients with high-risk chronic CAD, including patients with normal LV function ⁽¹⁶³⁾. Accordingly, ACEIs should be used in such patients as well as in those with hypertension that is not controlled with beta-blockers and nitrates.

Other less extensively studied techniques for the relief of ischemia, such as spinal cord stimulation ⁽¹⁶⁴⁾ and prolonged external counterpulsation ^(165,166), are under evaluation. Most experience has been gathered with spinal cord stimulation in "intractable angina" ⁽¹⁶⁷⁾, in which anginal relief has been described.

The KATP channel openers have hemodynamic and cardioprotective effects that could be useful in UA/NSTEMI. Nicorandil is such an agent that is approved in a number of countries but not yet in the United States. In a pilot double-blind, placebo-controlled study of 245 patients with UA, the addition of this drug to conventional treatment significantly reduced the number of episodes of transient myocardial ischemia (mostly silent) and of ventricular and supraventricular tachycardia ⁽¹⁶⁸⁾. Further evaluation of this class of agents is under way.

B. Antiplatelet and Anticoagulation Therapy

Recommendations for Antiplatelet and Anticoagulation Therapy

Class I

1. Antiplatelet therapy should be initiated promptly. ASA should be administered as soon as possible after presentation and continued indefinitely. (Level of Evidence: A)
2. Clopidogrel should be administered to hospitalized patients who are unable to take ASA because of hypersensitivity or major gastrointestinal intolerance. (Level of Evidence: A)
3. In hospitalized patients in whom an early noninterventional approach is planned, clopidogrel should be added to ASA as soon as possible on admission and administered for at least 1 month (Level of Evidence: A) and for up to 9 months (Level of Evidence: B)
4. In patients for whom a PCI is planned, clopidogrel should be started and continued for at least 1 month (Level of Evidence: A) and up to 9 months in patients who are not at high risk for bleeding (Level of Evidence: B)
5. In patients taking clopidogrel in whom elective CABG is planned, the drug should be withheld for 5 to 7 days. (Level of Evidence: B)
6. Anticoagulation with subcutaneous LMWH or intravenous unfractionated heparin (UFH) should be added to antiplatelet therapy with ASA and/or clopidogrel. (Level of Evidence: A)
7. A platelet GP IIb/IIIa antagonist should be administered, in addition to ASA and heparin, to patients in whom catheterization and PCI are planned. The GP IIb/IIIa antagonist may also be administered just prior to PCI. (Level of Evidence: A)

Class IIa

1. Eptifibatide or tirofiban should be administered, in addition to ASA and LMWH or UFH, to patients with continuing ischemia, an elevated troponin or with other high-risk features in whom an invasive management strategy is not planned. (Level of Evidence: A)
2. Enoxaparin is preferable to UFH as an anticoagulant in patients with UA/NSTEMI, unless CABG is planned within 24 h. (Level of Evidence: A)
3. A platelet GP IIb/IIIa antagonist should be administered to patients already receiving heparin, ASA, and clopidogrel in whom catheterization and PCI are planned. The GP IIb/IIIa antagonist may also be administered just prior to PCI. (Level of Evidence: B)

Class IIb

Eptifibatide or tirofiban, in addition to ASA and LMWH or UFH, to patients without continuing ischemia who have no other high-risk features and in whom PCI is not planned. (Level of Evidence: A)

Class III

1. Intravenous fibrinolytic therapy in patients without acute ST-segment elevation, a true posterior MI, or a presumed new left bundle-branch block (LBBB). (Level of Evidence: A)
2. Abciximab administration in patients in whom PCI is not planned. (Level of Evidence: A)

Antithrombotic therapy is essential to modify the disease process and its progression to death, MI, or recurrent MI. A combination of ASA, UFH, and a platelet GP IIb/IIIa receptor antagonist represents the most effective therapy. The intensity of treatment is tailored to individual risk, and triple antithrombotic treatment is used in patients with continuing ischemia or with other high-risk features and in patients oriented to an early invasive strategy (Table 14). Table 15 shows the recommended doses of the various agents. An LMWH can be advantageously substituted for UFH, although experience with the former in PCI, in patients referred for urgent CABG, and in combination with a GP IIb/IIIa antagonist and with a thrombolytic agent is limited. In the ESSENCE ⁽¹⁶⁹⁾ and TIMI 11B trials ^(170,171) in UA and NSTEMI, enoxaparin was stopped 6 to 12 h before a planned percutaneous procedure or surgery and UFH was substituted. Trials are now under way to test the safety and efficacy of enoxaparin in patients undergoing PCI and of enoxaparin and dalteparin combined with a GP IIb/IIIa antagonist and with a lytic agent. A pilot double-blind, randomized study of 55 patients compared the combination of tirofiban and enoxaparin with the combination of tirofiban and UFH. The results suggested more reproducible inhibition of platelet aggregation with enoxaparin and less prolongation in bleeding time. There was an excess of minor bleeding with enoxaparin but not of major bleeding ⁽¹⁷²⁾. Furthermore, there may be problems in rapid reversal of the anticoagulation when required, such as before CABG.

1. Antiplatelet Therapy (Aspirin, Ticlopidine, Clopidogrel)
a. Aspirin
Some of the strongest evidence available about the long-term prognostic effects of therapy in patients with coronary disease pertains to ASA ⁽¹⁷³⁾. By irreversibly inhibiting cyclooxygenase-1 within platelets, ASA prevents the formation of thromboxane A2, thereby diminishing platelet aggregation promoted by this pathway but not by others. This platelet inhibition is the plausible mechanism for clinical benefit of ASA because it is fully present with low doses of ASA and because platelets represent one of the principal participants in thrombus formation after plaque disruption. Alternative or additional mechanisms of action for ASA are possible, such as an anti-inflammatory effect ⁽¹⁷⁴⁾, but they are unlikely at the low doses of ASA that are effective in UA/NSTEMI.

Among all clinical investigations with ASA, trials in UA/NSTEMI have most consistently documented a striking benefit of the drug independent of the differences in study design, such as time of entry after the acute phase, duration of follow-up, and doses used ^(175-178) (Figure 8).

No trial has directly compared the efficacy of different doses of ASA in patients who present with UA/NSTEMI. However, trials in secondary prevention of stroke, MI, death, and graft occlusion have not shown an added benefit for ASA doses of greater than 80 and 160 mg per d but have shown a higher risk of bleeding. An overview of trials with different doses of ASA in long-term treatment of patients with CAD suggests similar efficacy for daily doses ranging from 75 to 324 mg ⁽¹⁷³⁾. A dose of 160 mg per d was used in the Second International Study of Infarct Survival (ISIS-2) trial, which definitively established the efficacy of ASA in suspected MI ⁽¹⁸⁵⁾. It therefore appears reasonable to initiate ASA treatment in patients with UA/NSTEMI at a dose of 160 mg, as used in the ISIS-2 trial, or 325 mg. In patients who present with suspected ACS who are not already receiving ASA, the first dose may be chewed to rapidly establish a high blood level. Subsequent doses may be swallowed. Thereafter, daily doses of 75 to 325 mg are prescribed.

The prompt action of ASA and its ability to reduce mortality rates in patients with suspected AMI enrolled in the ISIS-2 trial led to the recommendation that ASA be initiated immediately in the ED as soon as the diagnosis of ACS is made or suspected. In patients who are already receiving ASA, it should be continued. The protective effect of ASA has been sustained for at least 1 to 2 years in clinical trials in UA. Longer-term follow-up data in this population are lacking. Given the relatively short-term prognostic impact of UA/NSTEMI in patients with coronary disease, long-term efficacy can be extrapolated from other studies of ASA therapy in CAD. Studies in patients with prior MI, stroke, or transient ischemic attack have shown statistically significant benefit during the first 2 years and some additional but not statistically significant benefit during the third year ⁽¹⁷³⁾. In the absence of large comparison trials of different durations of antiplatelet treatment in patients with cardiovascular disease or in primary prevention, it seems prudent to continue ASA indefinitely unless side effects are present ^(5,26,173). Thus, patients should be informed of the evidence that supports the use of ASA in UA/NSTEMI and CAD in general and instructed to continue the drug indefinitely, unless a contraindication develops.

Contraindications to ASA include intolerance and allergy (primarily manifested as asthma), active bleeding, hemophilia, active retinal bleeding, severe untreated hypertension, an active peptic ulcer, or another serious source of gastrointestinal or genitourinary bleeding. Gastrointestinal side effects such as dyspepsia and nausea are infrequent with the low doses. Acute gout due to impaired urate excretion is rarely precipitated. Primary prevention trials have reported a small excess in intracranial bleeding, which is offset in secondary prevention trials by the prevention of ischemic stroke. It has been proposed that there is a negative interaction between ACEIs and ASA with a reduction in the vasodilatory effects of ACEIs, presumably because ASA inhibits ACEI-induced prostaglandin synthesis. This interaction does not appear to interfere with the clinical benefits of therapy with either agent ⁽¹⁸⁶⁾. Therefore, unless there are specific contraindications, ASA should be administered to all patients with UA/NSTEMI.

b. Adenosine Diphosphate Receptor Antagonists and Other Antiplatelet Agents
Two thienopyridines-ticlopidine and clopidogrel-are adenosine diphosphate (ADP) antagonists that are currently approved for antiplatelet therapy ⁽¹⁸⁷⁾. The platelet effects of ticlopidine and clopidogrel are irreversible but take several days to become completely manifest. Because the mechanisms of the antiplatelet effects of ASA and ADP antagonists differ, a potential exists for additive benefit with the combination.

Ticlopidine has been used successfully for the secondary prevention of stroke and MI and for the prevention of stent closure and graft occlusion. In an open-label trial ⁽¹⁸⁸⁾, 652 patients with UA were randomized to receive 250 mg of ticlopidine twice a day or standard therapy without ASA. At 6-month follow-up, ticlopidine reduced the rate of fatal and nonfatal MI by 46% (13.6% vs. 7.3%, p = 0.009). The benefit of ticlopidine in the study developed after only 2 weeks of treatment, which is consistent with the delay of the drug to achieve full effect.

The adverse effects of ticlopidine limit its usefulness: gastrointestinal problems (diarrhea, abdominal pain, nausea, vomiting), neutropenia in approximately 2.4% of patients, severe neutropenia in 0.8% of patients, and, rarely, thrombotic thrombocytopenia purpura (TTP) ⁽¹⁸⁹⁾. Neutropenia usually resolves within 1 to 3 weeks of discontinuation of therapy but very rarely may be fatal. TTP, which also is a very uncommon life-threatening complication, requires immediate plasma exchange. Monitoring of ticlopidine therapy requires a complete blood count that includes a differential count every 2 weeks for the first 3 months of therapy.

Most clinical experience with clopidogrel is derived from the Clopidogrel versus Aspirin in Patients at Risk of Ischaemic Events (CAPRIE) trial ⁽¹⁹⁰⁾. A total of 19,185 patients were randomized to receive 325 mg per day ASA or 75 mg per day clopidogrel. Entry criteria consisted of atherosclerotic vascular disease manifested as recent ischemic stroke, recent MI, or symptomatic peripheral arterial disease. Follow-up extended for 1 to 3 years. The RR of ischemic stroke, MI, or vascular death was reduced by 8.7% in favor of clopidogrel from 5.83% to 5.32% (p = 0.043). There was a slightly increased, but minimal, incidence of rash and diarrhea with clopidogrel treatment and slightly more bleeding with ASA. There was no excess neutropenia with clopidogrel, which contrasts with ticlopidine. The results provide evidence that clopidogrel is at least as effective as ASA and may be modestly more effective. In a recent report, 11 severe cases of TTP were described as occurring within 14 days after the initiation of clopidogrel; plasma exchange was required in 10 of the patients, and 1 patient died ⁽¹⁹¹⁾. These cases occurred among more than 3 million patients treated with clopidogrel.

Ticlopidine or clopidogrel is reasonable antiplatelet therapy for secondary prevention with an efficacy at least similar to that of ASA. These drugs are indicated in patients with UA/NSTEMI who are unable to tolerate ASA due to either hypersensitivity or major gastrointestinal contraindications, principally recent significant bleeding from a peptic ulcer or gastritis. Care must be taken during the acute phase with these drugs because of the delays required to achieve a full antiplatelet effect. Clopidogrel is preferred to ticlopidine because it more rapidly inhibits platelets and appears to have a more favorable safety profile. Experience is being acquired with this drug in acute situations with a loading dose (300 mg) to achieve more rapid platelet inhibition. Initial treatment with heparin (UFH or LMWH) and probably with a GP IIb/IIIa antagonist is especially important in patients with UA/NSTEMI who are treated with 1 of the thienopyridines because of their delayed onset of antiplatelet activity compared with ASA.

Two randomized trials were recently completed in which clopidogrel was compared with ticlopidine. In 1 study, 700 patients who successfully received a stent were randomized to receive 500 mg of ticlopidine or 75 mg of clopidogrel, in addition to 100 mg of ASA, for 4 weeks ⁽¹⁹²⁾. Cardiac death, urgent target vessel revascularization, angiographically documented thrombotic stent occlusion, or nonfatal MI within 30 days occurred in 3.1% of patients who received clopidogrel and 1.7% of patients who received ticlopidine (p = 0.24), and noncardiac death, stroke, severe peripheral vascular hemorrhagic events, or any adverse event that resulted in the discontinuation of the study medication occurred in 4.5% and 9.6% of patients, respectively (p = 0.01). The CLopidogrel ASpirin Stent International Cooperative Study (CLASSICS) (P. Urban, A.H. Gershlick, H.-J. Rupprecht, M.E. Bertrands, oral presentation, American Heart Association Scientific Sessions, Atlanta, Ga, November 1999) was conducted in 1,020 patients. A loading dose of 300 mg of clopidogrel followed by 75 mg per day was compared to a daily dose of 75 mg without a loading dose and with a loading dose of 150 mg of ticlopidine followed by 150 mg twice a day (patients in each of the 3 arms also received ASA). The first dose was administered 1 to 6 h after stent implantation; the treatment duration was 28 days. The trial showed better tolerance to clopidogrel with or without a loading dose than to ticlopidine. Stent thrombosis or major complications occurred at the same frequency in the 3 groups.

The Clopidogrel in Unstable angina to prevent Recurrent ischemic Events (CURE) trial randomized 12,562 patients with UA and NSTEMI presenting within 24 h to placebo or clopidogrel (loading dose of 300 mg followed by 75 mg daily) and followed them for 3 to 12 months. All patients received aspirin. Cardiovascular death, MI, or stroke occurred in 11.5% of patients assigned to placebo and 9.3% assigned to clopidogrel (RR = 0.80, p less than 0.001). In addition, clopidogrel was associated with significant reductions in the rate of inhospital severe ischemia and revascularization, as well as the need for thrombolytic therapy or intravenous GP IIb/IIIa receptor antagonists. These results were observed across a wide variety of subgroups. A reduction in recurrent ischemia was noted within the first few hours after randomization.

There was an excess of major bleeding (2.7% in the placebo group vs. 3.7% in the clopidogrel group, p = 0.003) as well as of minor but not of life-threatening bleeding. The risk of bleeding was increased in patients undergoing CABG surgery within the first 5 days of stopping clopidogrel. CURE was conducted at centers in which there was no routine policy of early invasive procedures; revascularization was performed during the initial admission in only 23% of the patients. Although the addition of a platelet GP IIb/IIIa inhibitor in patients receiving ASA, clopidogrel, and heparin in CURE was well tolerated, fewer than 10% of patients received this combination. Therefore, additional information on the safety of the addition of heparin (LMWH or UFH) and a GP IIb/IIIa inhibitor in patients already receiving ASA and clopidogrel should be obtained. Also, it is not yet clear whether clopidogrel improved the outcome in patients who received GP IIb/IIIa antagonists.

This trial provides strong evidence for the addition of clopidogrel to ASA on admission in the management of patients with UA and NSTEMI, in whom a noninterventional approach is intended-an especially useful approach in hospitals that do not have a routine policy of early invasive procedures. The optimal duration of therapy with clopidogrel has not been determined, but the favorable results in CURE were observed over a period averaging 9 months.

In the PCI-CURE study, 2,658 patients undergoing PCI had been randomly assigned double-blind treatment with clopidogrel (n = 1,313) or placebo (n = 1,345). Patients were pretreated with aspirin and the study drug for a median of 10 days. After PCI, most patients received open-label thienopyridine for about 4 weeks, after which the study drug was restarted for a mean of 8 months. Fifty-nine patients (4.5%) in the clopidogrel group had the primary end point-a composite of cardiovascular death, MI, or urgent target-vessel revascularization-within 30 days of PCI compared with 86 (6.4%) in the placebo group (relative risk 0.70 [95% Confidence Interval {CI} 0.50-0.97], p = 0.03). Overall (including events before and after PCI), there was a 31% reduction in cardiovascular death or MI (p = 0.002). Thus, in patients with UA and NSTEMI receiving ASA and undergoing PCI, a strategy of clopidogrel pretreatment followed by at least 1 month and probably longer-term therapy is beneficial in reducing major cardiovascular events compared with placebo ⁽⁵²²⁾. Therefore, clopidogrel should be used routinely in patients who undergo PCI.

There now appears to be an important role for clopidogrel in patients with UA/NSTEMI, both those who are managed conservatively as well as those who undergo PCI, especially stenting. However, it is not entirely clear how long therapy should be maintained. Since clopidogrel, when added to ASA, increases the risk of bleeding during major surgery in patients who are scheduled for elective CABG, clopidogrel should be withheld for at least 5 days ⁽⁵²¹⁾, and preferably for 7 days before surgery ⁽⁵²³⁾. In many hospitals in which patients with UA/NSTEMI undergo diagnostic catheterization within 24 to 36 h of admission, clopidogrel is not started until it is clear that CABG will not be scheduled within the next several days. A loading dose of clopidogrel can be given to a patient on the catheterization table if a PCI is to be carried out immediately. If PCI is not carried out, the clopidogrel can be begun after the catheterization.

Sulfinpyrazone, dipyridamole, prostacyclin, and prostacyclin analogs have not been associated with benefit in UA or NSTEMI and are not recommended. The thromboxane synthase blockers and thromboxane A2 receptor antagonists have been evaluated in ACS but have not shown any advantage over ASA. A number of other antiplatelet drugs are currently available, and still others are under active investigation. Oral GP IIb/IIIa receptor blockers were tested in 1 PCI trial and 3 UA/NSTEMI trials; the 4 trials failed to document a benefit and 2 showed an excess mortality rate ^{(193,193a,193b)}.

Clopidogrel is the preferred thienopyridine because of its more rapid onset of action, especially after a loading dose ^(524,525) and better safety profile than ticlopidine ⁽⁵²⁶⁾.

2. Anticoagulants
Anticoagulants available for parenteral use include UFH, various LMWHs, and hirudin, and for oral use, the antivitamin K drugs are available. Synthetic pentasaccharides and synthetic direct thrombin inhibitors (argatroban, bivaluridine) as well as oral direct and indirect thrombin inhibitors are under clinical investigation. Hirudin is approved as an anticoagulant in patients with heparin-induced thrombocytopenia and for the prophylaxis of deep vein thrombosis after hip replacement.

Heparin exerts its anticoagulant effect by accelerating the action of circulating antithrombin, a proteolytic enzyme that inactivates factor IIa (thrombin), factor IXa, and factor Xa. It prevents thrombus propagation but does not lyse existing thrombi ⁽¹⁹⁴⁾. UFH is a heterogeneous mixture of chains of molecular weights that range from 5,000 to 30,000 and have varying effects on anticoagulant activity. UFH binds to a number of plasma proteins, blood cells, and endothelial cells. The LMWHs are obtained through chemical or enzymatic depolymerization of the polysaccharide chains of heparin to provide chains with different molecular weight distributions. About 25% to 50% of the pentasaccharide-containing chains of LMWH preparations contain greater than 18 saccharide units, and these are able to inactivate both thrombin and factor Xa. However, LMWH chains that are less than 18 saccharide units retain their ability to inactivate factor Xa but not thrombin. Therefore, LMWHs are relatively more potent in the catalyzation of the inhibition of factor Xa by antithrombin than in the inactivation of thrombin. Distinct advantages of LMWH over UFH include decreased binding to plasma proteins and endothelial cells and dose-independent clearance with a longer half-life that results in more predictable and sustained anticoagulation with once- or twice-a-day subcutaneous administration. A major advantage of LMWHs is that they do not usually require laboratory monitoring of activity. The pharmacodynamic and pharmacokinetic profiles of the different commercial preparations of LMWHs vary, with their mean molecular weights ranging from 4,200 to 6,000. Accordingly, their ratios of anti-Xa factor to anti-IIa factor vary, ranging from 1.9 to 3.8 ⁽¹⁹⁵⁾.

By contrast, the direct thrombin inhibitors very specifically block thrombin effects without the need for a cofactor such as antithrombin. Hirudin binds directly to the anion binding site and the catalytic sites of thrombin to produce potent and predictable anticoagulation ⁽¹⁹⁶⁾. Several large trials (see later) that compare hirudin with UFH in UA/NSTEMI have demonstrated a modest short-term reduction in the composite end point of death or nonfatal MI with a modest increase in the risk of bleeding.

Bivalurudin (Hirulog) is a synthetic analog of hirudin that binds reversibly to thrombin. It has been compared with UFH in several small trials in UA/NSTEMI and in PCI with some evidence of a reduction in death or MI and less bleeding than with UFH ^(197-199).

a. Unfractionated Heparin
Seven randomized, placebo-controlled trials with UFH have been reported ^(200-205). A placebo-controlled study performed by Theroux et al. ⁽¹⁷⁷⁾ between 1986 and 1988 tested treatments that consisted of ASA and a 5,000-U IV bolus of UFH followed by 1,000 U per h in a 2 × 2 factorial design. UFH reduced the risk of MI by 89% and the risk of recurrent refractory angina by 63%. An extension of this study compared ASA and UFH in UA patients. MI (fatal or nonfatal) occurred in 3.7% of patients who received ASA and 0.8% of patients who received UFH (p = 0.035) ⁽²⁰²⁾.

The Research Group in Instability in Coronary Artery Disease (RISC) trial was a double-blind, placebo-controlled trial with a 2 × 2 factorial design that was conducted in men with UA or NSTEMI ⁽¹⁷⁸⁾. ASA significantly reduced the risk of death or MI. UFH alone had no benefit, although the group treated with the combination of ASA and UFH had the lowest number of events during the initial 5 days. Neri-Seneri et al. ⁽²⁰³⁾ suggested that symptomatic and silent episodes of ischemia in UA could be prevented by an infusion of UFH but not by bolus injections or by ASA. Taken together, these trials indicate that the early administration of UFH is associated with a reduction in the incidence of AMI and ischemia in patients with UA/NSTEMI.

The results of the studies that have compared the combination of ASA and heparin with ASA alone are shown in Figure 8. In the trials that used UFH, the reduction in the rate of death or MI during the first week was 54% (p = 0.016), and in the trials that used either UFH or LMWH, the reduction was 63%. Two published meta-analyses have included different studies. In 1 meta-analysis, which involved 3 randomized trials and an early end point (less than 5 days) ⁽¹⁷⁹⁾, the risk of death or MI with the combination of ASA and heparin was reduced by 56% (p = 0.03). In the second meta-analysis, which involved 6 trials and end points that ranged from 2 to 12 weeks, the RR was reduced by 33% (p = 0.06) ⁽²⁰⁶⁾. Most of the benefits of the various anticoagulants are short term, however, and not maintained on a long-term basis. Reactivation of the disease process after the discontinuation of anticoagulants may contribute to this loss of early gain that has been described with UFH ⁽²⁰⁷⁾, dalteparin ⁽¹⁸¹⁾, and hirudin ^(208,209). The combination of UFH and ASA appears to mitigate this reactivation in part ^(207,210), although there is hematologic evidence of increased thrombin activity after the cessation of intravenous UFH even in the presence of ASA ⁽²¹¹⁾. Uncontrolled observations suggested a reduction in the "heparin rebound" by switching from intravenous to subcutaneous UFH for several days before the drug is stopped.

UFH has important pharmacokinetic limitations that are related to its nonspecific binding to proteins and cells. These pharmacokinetic limitations of UFH translate into poor bioavailability, especially at low doses, and marked variability in anticoagulant response among patients ⁽²¹²⁾. As a consequence of these pharmacokinetic limitations, the anticoagulant effect of heparin requires monitoring with the activated partial thromboplastin time (aPTT), a test that is sensitive to the inhibitory effects of UFH on thrombin (factor IIa), factor Xa, and factor IXa. Many clinicians have traditionally prescribed a fixed initial dose of UFH (e.g., 5,000-U bolus, 1,000 U per h initial infusion); clinical trials have indicated that a weight-adjusted dosing regimen could provide more predictable anticoagulation than the fixed-dose regimen ^(213-215). The weight-adjusted regimen is recommended with an initial bolus of 60 to 70 U per kg (maximum 5,000 U) and an initial infusion of 12 to 15 U · kg^-1 · h^-1·(maximum 1,000 U per h). The therapeutic range of the various nomograms differs due to variation in the laboratory methods used to determine aPTT. The American College of Chest Physicians consensus conference ⁽²¹²⁾ has therefore recommended dosage adjustments of the nomograms to correspond to a therapeutic range equivalent to heparin levels of 0.3 to 0.7 U per mL by anti-factor Xa determinations, which correlates with aPTT values between 60 and 80 s. In addition to body weight, other clinical factors that affect the response to UFH include age, which is associated with higher aPTT values, and smoking history and diabetes mellitus, which are associated with lower aPTT values ^(212,216).

Thus, even though weight-based UFH dosing regimens are used, the aPTT should be monitored for adjustment of UFH dosing. Because of variation among hospitals in the control aPTT values, nomograms should be established at each institution that are designed to achieve aPTT values in the target range (e.g., for a control aPTT of 30 s, the target range [1.5 to 2.5 times control] would be 45 to 75 s). Measurements should be made 6 h after any dosage change and used to adjust UFH infusion until the aPTT exhibits a therapeutic level. When 2 consecutive aPTT values are therapeutic, the measurements may be made every 24 h and, if necessary, dose adjustment carried out. In addition, a significant change in the patient's clinical condition (e.g., recurrent ischemia, bleeding, hypotension) should prompt an immediate aPTT determination, followed by dose adjustment, if necessary.

Serial hemoglobin/hematocrit and platelet measurements are recommended at least daily during UFH therapy. In addition, any clinically significant bleeding, recurrent symptoms, or hemodynamic instability should prompt their immediate determination. Serial platelet counts are necessary to monitor for heparin-induced thrombocytopenia. Mild thrombocytopenia may occur in 10% to 20% of patients who are receiving heparin, whereas severe thrombocytopenia (platelet count less than 100,000) occurs in 1% to 2% of patients and typically appears after 4 to 14 days of therapy. A rare but dangerous complication (less than 0.2% incidence) is autoimmune UFH-induced thrombocytopenia with thrombosis ⁽²¹⁷⁾. A high clinical suspicion mandates the immediate cessation of all heparin therapy (including that used to flush intravenous lines).

Most of the trials that evaluate the use of UFH in UA/NSTEMI have continued therapy for 2 to 5 days. The optimal duration of therapy remains undefined.

b. Low-Molecular-Weight Heparin
In a pilot open-label study, 219 patients with UA were randomized to receive ASA (200 mg per d), ASA plus UFH, or ASA plus nadroparin, an LMWH. The combination of ASA and LMWH significantly reduced the total ischemic event rate, the rate of recurrent angina, and the number of patients requiring interventional procedures ⁽¹⁸⁰⁾.

The FRISC study ⁽¹⁸¹⁾ randomized 1,506 patients with UA or non-Q-wave MI to receive subcutaneous administration of the LMWH dalteparin (120 IU per kg twice daily) or placebo for 6 days and then once a day for the next 35 to 45 days. Dalteparin was associated with a 63% risk reduction in death or MI during the first 6 days (4.8% vs. 1.8%, p = 0.001), matching the favorable experience observed with UFH. Although an excess of events was observed after the dose reduction to once daily after 6 days, a significant decrease was observed at 40 days with dalteparin in the composite outcome of death, MI, or revascularization (23.7% vs. 18.0%, p = 0.005), and a trend was noted in a reduction in rates of death or MI (10.7% vs. 8.0%, p = 0.07).

Because the level of anticoagulant activity cannot be easily measured in patients receiving LMWH (e.g., aPTT or activated clotting time [ACT]), interventional cardiologists have expressed concern about the substitution of LMWH for UFH in patients scheduled for catheterization with possible PCI. However, Collett et al. ⁽⁵²⁷⁾ showed in a study involving 293 patients with UA/NSTEMI who received the usual dose of enoxaparin that PCI can be performed safely.

In the National Investigators Collaborating on Enoxaparin Trial (NICE-1), an observational study, intravenous enoxaparin (1.0 mg per kg) was used in 828 patients undergoing elective PCI ⁽⁵²⁸⁾ without an intravenous GP IIb/IIIa inhibitor. The rate of bleeding (1.1% for major bleeding and 6.2% for minor bleeding in 30 days) was comparable to historical controls given UFH.

An alternative approach is to use LMWH during the period of initial stabilization. The dose can be withheld on the morning of the procedure, and if an intervention is required and more than 8 hours has elapsed since the last dose of LMWH, UFH can be used for PCI according to usual practice patterns. Because the anticoagulant effect of UFH can be more readily reversed than that of LMWH, UFH is preferred in patients likely to undergo CABG within 24 h.

c. LMWH Versus UFH
Four large randomized trials have directly compared an LMWH with UFH (Figure 9). In the FRagmin In unstable Coronary artery disease (FRIC) study, 1482 patients with UA/NSTEMI received open-label dalteparin (120 IU per kg subcutaneously twice a day) or UFH for 6 days ⁽²¹⁸⁾. At day 6 and until day 45, patients were randomized a second time to double-blind administration of dalteparin (120 IU per kg once a day) or placebo. During the first part of the study, the risk of death, MI, or recurrent angina was nonsignificantly increased with dalteparin (9.3% vs. 7.65%, p = 0.33), and the risk of death or MI was unaffected (3.9% vs. 3.6%, p = 0.8); death also tended to occur more frequently with dalteparin (1.5% vs. 0.4% with UFH, p = 0.057). Between days 6 and 45, the rates of death, MI, and recurrence of angina were comparable between the active treatment and placebo groups.

The ESSENCE trial ⁽¹⁶⁹⁾ compared enoxaparin (1 mg per kg twice daily subcutaneous administration) with standard UFH (5,000 U bolus), followed by an infusion titrated to an aPTT of 55 to 86 s, administered for 48 h to 8 days (median duration in both groups of 2.6 days) ⁽¹⁶⁹⁾. With UFH, only 46% of patients reached the target aPTT within 12 to 24 h. The composite outcome of death, MI, or recurrent angina was reduced by 16.2% at 14 days with enoxaparin (19.8% UFH vs. 16.6% enoxaparin, p = 0.019) and by 19% at 30 days (23.3% vs. 19.8%, p = 0.017). The rates of death were unaffected, whereas there were trends to reductions in the rates of death and MI by 29% (p = 0.06) at 14 days and by 26% (p = 0.08) at 30 days.

The TIMI 11B trial randomized 3,910 patients with UA/NSTEMI to enoxaparin (30 mg IV initial bolus immediately followed by subcutaneous injections of 1 mg per kg every 12 h) or UFH (70 U per kg bolus followed by an infusion of 15 U · kg^-1 · h^-1 titrated to a target aPTT 1.5 to 2.5 times control) ⁽¹⁷⁰⁾. The acute phase therapy was followed by an outpatient phase, during which enoxaparin or placebo for patients who were initially randomized to UFH was administered in a double-blind manner twice a day. Enoxaparin was administered for a median of 4.6 days, and UFH was administered for a median of 3.0 days. The composite end point of death, MI, or need for an urgent revascularization (defined as an episode of recurrent angina prompting the performance of coronary revascularization during the index hospitalization or after discharge leading to rehospitalization and coronary revascularization) was reduced at 8 days from 14.5% to 12.4% (p = 0.048) and at 43 days from 19.6% to 17.3% (p = 0.048). The rates of death or MI were reduced from 6.9% to 5.7% (p = 0.114) at 14 days and from 8.9% to 7.9% (p = 0.276) at 43 days. No incremental benefit was observed with outpatient treatment, whereas the risk of major bleeding was significantly greater during the outpatient treatment. The risk of minor bleeding was also increased both in and out of hospital with enoxaparin.

The FRAXiparine in Ischaemic Syndrome (FRAXIS) trial had 3 parallel arms and compared the LMWH nadroparin administered for 6 or 14 days with control treatment with UFH ⁽²¹⁹⁾. Three thousand four hundred sixty-eight patients with UA or NSTEMI were enrolled. The composite outcome of death, MI, or refractory angina occurred at 14 days in 18.1% of patients in the UFH group, 17.8% of patients treated with nadroparin for 6 days, and 20.0% of patients treated with nadroparin for 14 days; the values at 3 months were 22.2%, 22.3%, and 26.2% of patients, respectively (p less than 0.03 for the comparison of 14-day nadroparin therapy with UFH therapy). Trends to more frequent death and to more frequent death or MI were observed at all time points in nadroparin-treated patients.

Thus, 2 trials with enoxaparin have shown a moderate benefit over UFH, and 2 trials (1 with dalteparin and 1 with nadroparin) showed neutral or unfavorable trends. Whether the heterogeneous results are explained by different populations, study designs, various heparin dose regimens, properties of the various LMWHs (more specifically different molecular weights and anti-factor Xa/anti-factor IIa ratios), or other unrecognized influences is a matter of speculation. A meta-analysis of the 2 trials with enoxaparin that involved a total of 7,081 patients showed a statistically significant reduction of approximately 20% in the rate of death, MI, or urgent revascularization at 2, 8, 14, and 43 days and in the rate of death or MI at 8, 14, and 43 days. At 8, 14, and 43 days, there was a trend toward a reduction in death as well ⁽¹⁷¹⁾.

Although it is tempting to compare the relative treatment effects of the different LMWH compounds in Figure 9, the limitations of such indirect comparisons must be recognized. The only reliable method of comparing 2 treatments is through a direct comparison in a well-designed clinical trial or series of trials. The comparison of different therapies (e.g., different LMWHs) with a common therapy (e.g., UFH) in different trials does not allow a conclusion to be made about the relative effectiveness of the different LMWHs because of the variability in both control group and experimental group event rates due to protocol differences, differences in concomitant therapies due to geographical and time variability, and the play of chance. Similar considerations apply to comparisons among platelet GP IIb/IIIa inhibitors.

However, in the Enoxaparin Versus Tinzaparin (EVET) trial, 2 LMWHs, enoxaparin and tinzaparin, administered for 7 days were compared in 438 patients with UA/NSTEMI. A preliminary report stated that both the recurrence of UA and the need for revascularization were significantly lower in the enoxaparin group ⁽⁵²⁹⁾. The advantages of LMWH preparations are the ease of subcutaneous administration and the absence of a need for monitoring. Furthermore, the LMWHs stimulate platelets less than UFH ⁽²²⁰⁾ and are less frequently associated with heparin-induced thrombocytopenia ⁽²²¹⁾. They are associated with more frequent minor, but not major, bleeding. In the ESSENCE trial, minor bleeding occurred in 11.9% of enoxaparin patients and 7.2% of UFH patients (p less than 0.001), and major bleeding occurred in 6.5% and 7.0%, respectively. In TIMI 11B, the rates of minor bleeding in hospital were 9.1% and 2.5%, respectively (p less than 0.001), and the rates of major bleeding were 1.5% and 1.0% (p = 0.143). In the FRISC study, major bleeding occurred in 0.8% of patients with dalteparin and in 0.5% of patients with placebo, and minor bleeding occurred in 8.2% (61 of 746 patients) and 0.3% (2 of 760 patients) of patients, respectively. The anticoagulation provided with LMWH is less effectively reversed with protamine than it is with UFH. In addition, LMWH administered during PCI does not permit monitoring of the ACT to titrate the level of anticoagulation. In the ESSENCE and TIMI 11B trials, special rules were set to discontinue enoxaparin before PCI and CABG. UFH was administered during PCI to achieve ACT values of greater than 350 s. An economic analysis of the ESSENCE trial suggested cost savings with enoxaparin ⁽²²²⁾. For patients who are receiving subcutaneous LMWH and in whom CABG is planned, it is recommended that LMWH be discontinued and UFH be used during the operation. Additional experience with regard to the safety and efficacy of the concomitant administration of LMWHs with GP IIb/IIIa antagonists and thrombolytic agents is currently being acquired.

The FRISC, FRIC, TIMI 11B, and Fast Revascularization During Instability in Coronary Artery Disease (FRISC II) trials evaluated the potential benefit of the prolonged administration of an LMWH after hospital discharge. The first 3 of these trials did not show a benefit of treatment beyond the acute phase. In the FRISC trial, doses of dalteparin were administered between 6 days and 35 to 45 days; in FRIC, patients were rerandomized after the initial 6-day treatment period to receive dalteparin for an additional 40 days; and the outpatient treatment period lasted 5 to 6 weeks in TIMI 11B and 1 week in the FRAXIS trial. The FRISC II trial used a different study design. Dalteparin was administered to all patients for a minimum of 5 days ⁽²²³⁾. Patients were subsequently randomized to receive placebo or the continued administration of dalteparin twice a day for up to 90 days. Analysis of the results from the time of randomization showed a significant reduction with dalteparin in the composite end point of death or MI at 30 days (3.1% vs. 5.9%, p = 0.002) but not at 3 months (6.7% vs. 8.0%, p = 0.17). The composite of death, MI, or revascularization during the total treatment period was reduced at 3 months (29.1% vs. 33.4%, p = 0.031). The benefits of prolonged dalteparin administration were limited to patients who were managed medically and to patients with elevated TnT levels at baseline. These results may make a case for the prolonged use of an LMWH in selected patients who are managed medically or in whom angiography is delayed.

d. Hirudin and Other Direct Thrombin Inhibitors
Hirudin, the prototype of the direct thrombin inhibitors, has been extensively studied. The GUSTO-IIb trial randomly assigned 12,142 patients to 72 h of therapy with either intravenous hirudin or UFH ⁽²²⁴⁾. Patients were stratified according to the presence of ST-segment elevation on the baseline ECG (4,131 patients) or its absence (8,011 patients). The primary end point of death, nonfatal MI, or reinfarction at 30 days occurred in 9.8% of the UFH group vs. 8.9% of the hirudin group (odds ratio [OR] 0.89, p = 0.058). For patients without ST-segment elevation, the rates were 9.1% and 8.3%, respectively (OR 0.90, p = 0.22). At 24 h, the risk of death or MI was significantly lower in the patients who received hirudin than in those who received UFH (2.1% vs. 1.3%, p = 0.001). However, the Thrombolysis and Thrombin Inhibition in Myocardial Infarction (TIMI) 9B trial of hirudin as adjunctive therapy to thrombolytic therapy in patients with STEMI showed no benefit of the drug over UFH either during study drug infusion or later ⁽²²⁵⁾. The GUSTO-IIb and TIMI 9B trials used hirudin doses of 0.1 mg per kg bolus and 0.1 mg · kg^-1 · h^-1 infusion for 3 to 5 days after the documentation of excess bleeding with the higher doses used in the GUSTO-IIA and TIMI 9A trials (0.6 mg per kg bolus and 0.2 mg · kg^-1 · h^-1 infusion) ^(224,226).

The Organization to Assess Strategies for Ischemic Syndromes (OASIS) program evaluated hirudin in patients with UA or non-Q-wave MI. OASIS 1 ⁽²²⁷⁾ was a pilot trial of 909 patients that compared the low hirudin dose of 0.1 mg per h infusion and the medium hirudin dose of 0.15 mg per h infusion with UFH. The latter dose provided the best results, with a reduction in the rate of death, MI, or refractory angina at 7 days (6.5% with UFH vs. 3.3% with hirudin, p = 0.047). This medium dose was used in the large OASIS 2 ⁽²²⁸⁾ trial that consisted of 10,141 patients with UA/NSTEMI who were randomized to receive UFH (5000 IU bolus plus 15 U · kg^-1 · h^-1) or recombinant hirudin (0.4 mg per kg bolus and 0.15 mg · kg^-1 · h^-1) infusion for 72 h. The primary end point of cardiovascular death or new MI at 7 days occurred in 4.2% in the UFH group vs. 3.6% patients in the hirudin group (RR 0.84, p = 0.064). A secondary end point of cardiovascular death, new MI, or refractory angina at 7 days was significantly reduced with hirudin (6.7% vs. 5.6%, RR 0.83, p = 0.011). There was an excess of major bleeds that required transfusion with hirudin (1.2% vs. 0.7% with heparin, p = 0.014) but no excess in life-threatening bleeds or strokes. A meta-analysis of the GUSTO-IIB, TIMI 9B, OASIS 1, and OASIS 2 trials showed risks of death or MI at 35 days relative to heparin after randomization of 0.90 (p = 0.015) with hirudin compared with UFH; RR values were 0.88 (p = 0.13) for patients receiving thrombolytic agents and 0.90 (p = 0.054) for patients not receiving thrombolytic agents ⁽²²⁸⁾. At 72 h, the RR values of death or MI were 0.78 (p = 0.003), 0.89 (p = 0.34), and 0.72 (p = 0.002), respectively. Additional trials of direct antithrombins in UA/NSTEMI appear warranted.

Hirudin (lepirudin) is presently indicated only for anticoagulation in patients with heparin-induced thrombocytopenia ⁽²²¹⁾ and for the prophylaxis of deep vein thrombosis in patients undergoing hip replacement surgery. It should be administered as a 0.4 mg per kg IV bolus over 15 to 20 s followed by a continuous intravenous infusion of 0.15 mg · kg^-1 · h^-1, with adjustment of the infusion to a target range of 1.5 to 2.5 times the control aPTT values.

e. Long-Term Anticoagulation
The long-term administration of warfarin has been evaluated in a few pilot studies. Williams et al. ⁽²⁰¹⁾ randomized 102 patients with UA to UFH for 48 h followed by open-label warfarin for 6 months and reported a 65% risk reduction in the rate of MI or recurrent UA. In the Antithrombotic Therapy in Acute Coronary Syndromes (ATACS) trial, Cohen et al. ⁽¹⁷⁹⁾ randomized 214 patients with UA/NSTEMI to ASA alone or the combination of ASA plus UFH followed by warfarin. At 14 days, there was a reduction in the composite end point of death, MI, and recurrent ischemia with the combination therapy (27.0% vs. 10.5%, p = 0.004). In a small randomized pilot study of 57 patients allocated to warfarin or placebo in addition to ASA, less evidence was noted of angiographic progression in the culprit lesion after 10 weeks of treatment with warfarin (33% for placebo vs. 4% for warfarin) and more regression was observed ⁽²²⁹⁾. The OASIS pilot study ⁽²³⁰⁾ compared a fixed dosage of 3 mg per d coumadin or a moderate dose titrated to an international normalized ratio (INR) of 2 to 2.5 in 197 patients for 7 months after the acute phase. Low-intensity warfarin had no benefit, whereas the moderate-intensity regimen reduced the risk of death, MI, or refractory angina by 58% and the need for rehospitalization for UA by 58%. However, these results were not reproduced in the larger OASIS 2 trial ⁽²²⁸⁾ of 3712 patients randomized to the moderate-intensity regimen of warfarin or standard therapy, with all patients receiving ASA. The rate of cardiovascular death, MI, or stroke after 5 months was 7.65% with the anticoagulant and 8.4% without (p = 0.37) ⁽²³¹⁾. Thus, the role, if any, of long-term warfarin in patients with UA/NSTEMI remains to be defined.

The Coumadin Aspirin Reinfarction Study (CARS) conducted in post-MI patients was discontinued prematurely due to a lack of evidence of benefit of reduced-dose ASA (80 mg per d) with either 1 or 3 mg of warfarin daily compared with 160 mg per d ASA alone ⁽²³²⁾. The Combination Hemotherapy And Mortality Prevention (CHAMP) study found no benefit of the use of warfarin (to an INR of 1.5 to 2.5) plus 81 mg per d ASA vs. 162 mg per d ASA with respect to total mortality, cardiovascular mortality, stroke, and nonfatal MI (mean follow-up of 2.7 years) after an index AMI (Oral presentation, The Combination Hemotherapy and Mortality Prevention (CHAMP) Study: Presented at the American Heart Association Annual Scientific Sessions, November 1999, Atlanta, GA). Low- or moderate-intensity anticoagulation with fixed-dose warfarin is not recommended for routine use after hospitalization for UA/NSTEMI. Warfarin should be prescribed, however, for UA/NSTEMI patients with established indications for warfarin, such as atrial fibrillation and mechanical prosthetic heart valves.

3. Platelet GP IIb/IIIa Receptor Antagonists
The GP IIb/IIIa receptor is abundant on the platelet surface. When platelets are activated, this receptor undergoes a change in configuration conformation that increases its affinity for binding to fibrinogen and other ligands. The binding of molecules of fibrinogen to receptors on different platelets results in platelet aggregation. This mechanism is independent of the stimulus for platelet aggregation and represents the final and obligatory pathway for platelet aggregation ⁽²³⁴⁾. The platelet GP IIb/IIIa receptor antagonists act by occupying the receptors, preventing fibrinogen binding, and thereby preventing platelet aggregation. Experimental and clinical studies have suggested that occupancy of greater than or equal to 80% of the receptor population and inhibition of platelet aggregation to ADP (5 to 20 micromoles per L) by greater than or equal to 80% results in potent antithrombotic effects ⁽²³⁵⁾. The various GP IIb/IIIa antagonists, however, possess significantly different pharmacokinetic and pharmacodynamic properties ⁽²³⁶⁾.

Abciximab is a Fab fragment of a humanized murine antibody that has a short plasma half-life but strong affinity for the receptor, resulting in some receptor occupancy that persists for weeks. Platelet aggregation gradually returns to normal 24 to 48 h after discontinuation of the drug. Furthermore, abciximab is not specific for GP IIb/IIIa and inhibits the vitronectin receptor (alpha_vbeta₃) on endothelial cells and the MAC-1 receptor on leukocytes ^(237,238). The clinical relevance of occupancy of these receptors is not presently known.

Eptifibatide is a cyclic heptapeptide that contains the KGD (Lys-Gly-Asp) sequence; tirofiban and lamifiban (a drug that is not yet approved) are nonpeptide mimetics of the RGD (Arg-Gly-Asp) sequence of fibrinogen ^{(236,239-241)}. Receptor occupancy with these 3 synthetic antagonists is in general in equilibrium with plasma levels. They have a half-life of 2 to 3 h and are highly specific for the GP IIb/IIIa receptor, with no effect on the vitronectin receptor (alpha_vbeta₃ integrin). Thus, the median percent inhibition of platelet aggregation to 5 micromoles per L ADP achieved after a loading dose of 0.4 mcg · kg^-1 · min^-1 of tirofiban for 30 min is 86%, and the inhibition is sustained with an infusion of 0.1 mcg · kg^-1 · min^-1. A higher dose of 10 mcg per kg over 3 min followed by an infusion of 0.15 mcg · kg^-1 · min^-1 achieves 90% inhibition within 5 min. Platelet aggregation returns to normal in 4 to 8 h after discontinuation of the drug, a finding that is consistent with the relatively short half-life of the drug ⁽²⁴²⁾. GP IIb/IIIa antagonists may bind different sites on the receptor and result in somewhat different binding properties that may modify their platelet effects and potentially, paradoxically, activate the receptor ⁽²⁴³⁾. Oral antagonists to the receptor are currently under investigation, although these programs have been slowed by the aforementioned negative results of 4 large trials of 3 of these compounds ^{(193,193a,193b)}.

The efficacy of GP IIb/IIIa antagonists in prevention of the complications associated with percutaneous interventions has been documented in numerous trials, many of them composed totally or largely of patients with UA ^{(182,244-246)} (see Figures 13 and 14 in Section IV). Two trials with tirofiban and 1 trial with eptifibatide have also documented their efficacy in UA/NSTEMI patients, only some of whom underwent interventions ^(10,21). A trial has been completed with lamifiban ⁽¹⁸³⁾, and one is ongoing with abciximab. Because the various agents have not been compared directly with each other, their relative efficacy is not known.

Abciximab has been studied primarily in PCI trials, in which its administration consistently showed a significant reduction in the rate of MI and the need for urgent revascularization (Table 16). The CAPTURE trial enrolled patients with refractory UA ⁽¹⁸²⁾. After angiographic identification of a culprit lesion suitable for angioplasty, patients were randomized to either abciximab or placebo administered for 20 to 24 h before angioplasty and for 1 h thereafter. The rate of death, MI, or urgent revascularization within 30 days (primary outcome) was reduced from 15.9% with placebo to 11.3% with abciximab (RR 0.71, p = 0.012). At 6 months, death or MI had occurred in 10.6% of the placebo-treated patients vs. 9.0% of the abciximab-treated patients (p = 0.19). The longer action of abciximab makes it less optimal in patients likely to need CABG than tirofiban or eptifibatide, whose action is shorter. Abciximab is approved for the treatment of UA/NSTEMI as an adjunct to PCI or when PCI is planned within 24 h.

The GUSTO IV-ACS trial ⁽⁵³⁰⁾ enrolled 7,800 patients with UA/NSTEMI who were admitted to the hospital with more than 5 min of chest pain and either ST-segment depression and/or elevated TnT or TnI concentration. All received ASA and either UFH or LMWH. They were randomized to placebo, an abciximab bolus and 24-h infusion, or an abciximab bolus and 48-h infusion. In contrast to other trials with GP IIb/IIIa antagonists, GUSTO IV-ACS enrolled patients in whom early (less than 48 h) revascularization was not intended. At 30 days, death or MI occurred in 8.0% of patients taking placebo, 8.2% of patients taking 24-h abciximab, and 9.1% of patients taking 48-h abciximab, differences that were not statistically significant. At 48 h, death occurred in 0.3%, 0.7%, and 0.9% of patients in these groups, respectively (placebo vs. abciximab 48 h, p = 0.008). The lack of benefit of abciximab was observed in most subgroups, including patients with elevated concentrations of troponin who were at higher risk. Although the explanation for these results is not clear, they indicate that abciximab at the dosing regimen used in GUSTO IV-ACS is not indicated in the management of patients with UA or NSTEMI in whom an early invasive management strategy is not planned.

Tirofiban was studied in the Platelet Receptor Inhibition in Ischemic Syndrome Management (PRISM) ⁽¹⁸⁴⁾ and Platelet Receptor Inhibition in Ischemic Syndrome Management in Patients Limited by Unstable Signs and Symptoms (PRISM-PLUS) ⁽²¹⁾ trials. The former trial directly compared tirofiban with heparin in 3,232 patients with accelerating angina or angina at rest and ST-segment or T-wave changes and with enzyme elevation, a previous MI, or a positive stress test or angiographically documented coronary disease. The primary composite outcome (death, MI, or refractory ischemia at the end of a 48-h infusion period) was reduced from 5.6% with UFH to 3.8% with tirofiban (RR 0.67, p = 0.01). At 30 days, the frequency of the composite outcome was similar in the 2 groups (17.1% for UFH vs. 15.9% for tirofiban, p = 0.34), but a trend toward reduction in the rate of death or MI was present with tirofiban (7.1% vs. 5.8%, p = 0.11), and a significant reduction in mortality rates was observed (3.6% vs. 2.3%, p = 0.02). The benefit of tirofiban was mainly present in patients with an elevated TnI or TnT concentration at baseline ⁽⁹⁰⁾.

The PRISM-PLUS trial enrolled 1,915 patients with clinical features of UA within the previous 12 h and the presence of ischemic ST-T changes or CK and CK-MB elevation. Patients were randomized to tirofiban alone, UFH alone, or the combination for a period varying from 48 to 108 h. The tirofiban-alone arm was dropped during the trial because of an excess mortality rate. The combination of tirofiban and UFH compared with UFH alone reduced the primary composite end point of death, MI, or refractory ischemia at 7 days from 17.9% to 12.9% (RR 0.68, p = 0.004). This composite outcome was also significantly reduced by 22% at 30 days (p = 0.03) and by 19% at 6 months (p = 0.02). The end point of death or nonfatal MI was reduced by 43% at 7 days (p = 0.006), 30% at 30 days (p = 0.03), and 22% at 6 months (p = 0.06).

Computer-assisted analysis of coronary angiograms obtained after 48 h of treatment in 1,491 patients in the PRISM-PLUS trial showed a significant reduction in the thrombus load at the site of the culprit lesion and improved coronary flow as assessed according to the TIMI criteria in patients who received the combination of tirofiban and UFH ⁽²⁴⁷⁾. Tirofiban, in combination with heparin, has been approved for the treatment of patients with ACS, including patients who are managed medically as well as those undergoing PCI.

Eptifibatide was studied in the PURSUIT trial, which enrolled 10,948 patients who had chest pain at rest within the previous 24 h and ST-T changes or CK-MB elevation ⁽¹⁰⁾. The study drug was added to standard management until hospital discharge or for 72 h, although patients with normal coronary arteries or other mitigating circumstances had shorter infusions. The infusion could be continued for an additional 24 h if an intervention was performed near the end of the 72-h infusion period. The primary outcome rate of death or nonfatal MI at 30 days was reduced from 15.7% to 14.2% with eptifibatide (RR 0.91, p = 0.042). Within the first 96 h, a substantial treatment effect was seen (9.1% vs. 7.6%, p = 0.01). The benefits were maintained at 6-month follow-up. Eptifibatide has been approved for the treatment of patients with ACS (UA/NSTEMI) who are treated medically or with PCI. It is usually administered with ASA and heparin.

The cumulative event rates observed during the phase of medical management and at the time of PCI in the CAPTURE, PRISM-PLUS, and PURSUIT trials are shown in Figure 10 ⁽²⁴⁸⁾. By protocol design, almost all patients underwent PCI in CAPTURE. In PRISM-PLUS, angiography was recommended. A percutaneous revascularization was performed in 30.5% of patients in PRISM-PLUS and in 13.0% of patients in PURSUIT. Each trial has shown a statistically significant reduction in the rate of death or MI during the phase of medical management; the reduction in event rates was magnified at the time of the intervention.

Although it is tempting to evaluate the drug effect by comparing patients who had intervention with those who did not, such an analysis is inappropriate. Patients who do not undergo intervention include many low-risk patients, patients who died before having the opportunity for intervention, patients with contraindications, and patients with uncomplicated courses in countries and practices that use the ischemia-guided approach; there is no way to adjust for these imbalances. Accordingly, the analysis in Figure 10 includes the event rates for all patients during the time when they were treated medically. It then begins the analysis anew in patients who underwent PCI at the time of angiography while on drug or placebo. In the PRISM-PLUS trial, 1,069 patients did not undergo early PCI. Although tirofiban treatment was associated with a lower incidence of death, MI or death, or of MI or refractory ischemia at 30 days, these reductions were not statistically significant ⁽⁵³¹⁾. In a high-risk subgroup of these patients not undergoing PCI (TIMI risk score greater than or equal to 4) ⁽⁵¹⁷⁾ tirofiban appeared to be beneficial whether they underwent PCI (OR 0.60, 95% CI 0.35-1.01) or not (OR 0.69, 95% CI 0.49-0.99). However, no benefit was observed in patients at lower risk ^(519,532). In the PURSUIT trial eptifibatide reduced the incidence of death or MI from 15.7% to 14.2% (RR 0.91, 95% CI = 0.79-1.00, p = 0.032) ⁽⁵³⁴⁾.

Boersma et al. carried out a meta-analysis of GP IIb/IIIa antagonists of all six large randomized placebo-controlled trials (including GUSTO IV ^[530] involving 31,402 patients with UA/NSTEMI not routinely scheduled to undergo coronary revascularization ^[535]). A small reduction in the odds of death or MI in the active treatment arm (11.8% vs 10.8%, OR = 0.91, 95% CI 0.84-0.98, p = 0.015) was observed. Unexpectedly, no benefit was observed in women (test for interaction between treatment assignment and gender, p less than 0.0001). In the meta-analysis, reductions in the end points of death or nonfatal MI considered individually did not achieve statistical significance.

Although not scheduled for coronary revascularization procedures, 11,965 of the 31,402 patients (38%) actually underwent PCI or CABG within 30 days, and in this subgroup the OR for death or MI in the patients assigned to GP IIb/IIIa antagonists was 0.89 (95% CI 0.80-0.98). In the other 19,416 patients who did not undergo coronary revascularization, the OR for death or MI in the GP IIb/IIIa group was 0.95 (0.86-1.05, NS). Major bleeding complications were increased in the GP IIb/IIIa antagonist-treated group compared to those who received placebo (2.4% vs. 1.4%, p less than 0.0001). The authors concluded that in patients with UA/NSTEMI not routinely scheduled for early revascularization and at high risk of thrombotic complications, "treatment with a GP IIb/IIIa inhibitor might therefore be considered" ⁽⁵³⁵⁾. Thus, GP IIb/IIIa inhibitors are of substantial benefit in patients with UA/NSTEMI who undergo PCI; they are of modest benefit in patients who are not routinely scheduled to undergo PCI (but who may do so), and they are of questionable benefit in patients who do not undergo PCI.

Although there is a temptation to use the comparison of each of these GP IIb/IIIa inhibitors with placebo to draw conclusions about relative efficacy, such an exercise could be misleading. Each trial had different entry criteria, different approaches to angiographic evaluation, and different criteria for end point measurement and took place in different locations in different time periods. The effects of these differences cannot be accounted for in an indirect comparison. Head-to-head (direct) comparisons will be required to draw reliable conclusions about the relative efficacy of these different molecules.

Treatment with a GP IIb/IIIa antagonist increases the risk of bleeding, which is typically mucocutaneous or involves the access site of vascular intervention. Unfortunately, each trial also used a different definition of bleeding and reported differently with regard to bleeding related to CABG. In the PRISM trial with no interventions on treatment, major bleeding (excluding CABG) occurred in 0.4% of patients who received tirofiban and 0.4% of patients who received UFH. In the PRISM-PLUS trial, major bleeding according to the TIMI criteria was reported in 1.4% of patients who received tirofiban and 0.8% of patients who received placebo (p = 0.23), whereas PURSUIT reported major bleeding in 10.6% of patients who received eptifibatide and 9.1% of patients who received placebo (p = 0.02). In the PURSUIT trial, with the exclusion of patients who underwent CABG, the rates were 3.0% with eptifibatide and 1.3% with placebo (p less than 0.001). No trials have shown an excess of intracranial bleeding with a GP IIb/IIIa inhibitor. As with the efficacy data, the temptation to make indirect comparisons should be tempered by the variability in protocol, circumstances, and definitions of the trial.

ASA has been used with the intravenous GP IIb/IIIa receptor blockers in all trials. A strong case can also be made for the concomitant use of heparin with GP IIb/IIIa receptor blockers. The tirofiban arm without UFH in the PRISM-PLUS trial was discontinued early because of an excess of deaths. In addition, the PURSUIT trial reported a higher event rate in the 11% of patients who were not treated with concomitant heparin ⁽¹⁰⁾. Current recommendations call for the concomitant use of heparin with GP IIb/IIIa inhibitors. It should be noted that an interaction exists between heparin and GP IIb/IIIa inhibitors with a higher ACT for the combination and a need for lower doses of heparin than usually recommended to achieve the best outcomes in the setting of PCI. Information is currently being gained concerning the safety and efficacy of the combination of LMWH and GP IIb/IIIa inhibitors.

Blood hemoglobin and platelet counts should be monitored and patient surveillance for bleeding should be carried out daily during the administration of GP IIb/IIIa receptor blockers. Thrombocytopenia is an unusual complication of this class of agents. Severe thrombocytopenia defined by nadir platelet counts of less than 50,000 mL-1 is observed in 0.5% of patients, and profound thrombocytopenia defined by nadir platelet counts of less than 20,000 mL-1 is observed in 0.2% of patients. Although reversible, thrombocytopenia is associated with an increased risk of bleeding ^(249,250).

Although the data are not definitive, it does appear that GP IIb/IIIa inhibitors can be used with LMWH. In the Antithrombotic Combination Using Tirofiban and Enoxaparin (ACUTE II) study ⁽⁵³⁶⁾, UFH and enoxaparin were compared in patients with UA/NSTEMI receiving tirofiban. The incidence of major and minor bleeding was similar, and there was a trend to fewer adverse events in the patients receiving enoxaparin. A number of other open-label studies have examined the safety of combining enoxaparin with abciximab, eptifibatide, or tirofiban in patients with UA/NSTEMI being treated with PCI or conservatively ⁽⁵³⁶⁾; of combining enoxaparin with abciximab in patients undergoing elective PCI ⁽⁵³⁷⁾; of combining dalteparin with abciximab during PCI ⁽⁵³⁸⁾ (J.J. Ferguson, oral presentation, ACC Annual Scientific Sessions, Orlando, Florida, March 2001); and of administering dalteparin to patients with UA/NSTEMI receiving abciximab who were treated conservatively (L.C. Wallentin, oral presentation, Congress of the European Society of Cardiology, Amsterdam, The Netherlands, August 2000). Although the majority of these studies relied on historical controls, none suggested that the combination of LMWH and a GP IIb/IIIa inhibitor was associated with excess bleeding, whether or not the patient also underwent PCI.

a. Thrombolysis
The failure of intravenous thrombolytic therapy to improve clinical outcomes in the absence of AMI with ST-segment elevation or bundle-branch block was clearly demonstrated in the TIMI IIIB, ISIS-2, and Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto-1 (GISSI) 1 trials ^(19,251,252). A meta-analysis of thrombolytic therapy in UA patients showed no benefit of thrombolysis vs. standard therapy for the reduction of AMI ⁽¹⁹⁾. Thrombolytic agents had no significant beneficial effect and actually increased the risk of MI ⁽¹⁹⁾. Consequently, such therapy is not recommended for the management of patients with an ACS without ST-segment elevation, a posterior wall MI, or a presumably new LBBB (see ACC/AHA Guidelines for the Management of Patients With Acute Myocardial Infarction ^[5]).

C. Risk Stratification

Recommendations

Class I

1. Noninvasive stress testing in low-risk patients (Table 6) who have been free of ischemia at rest or with low-level activity and of CHF for a minimum of 12 to 24 h. (Level of Evidence: C)
2. Noninvasive stress testing in patients at intermediate risk (Table 6) who have been free of ischemia at rest or with low-level activity and of CHF for a minimum of 2 or 3 days. (Level of Evidence: C)
3. Choice of stress test is based on the resting ECG, ability to perform exercise, local expertise, and technologies available. Treadmill exercise is suitable in patients able to exercise in whom the ECG is free of baseline ST-segment abnormalities, bundle-branch block, LV hypertrophy, intraventricular conduction defect, paced rhythm, preexcitation, and digoxin effect. (Level of Evidence: C)
4. An imaging modality is added in patients with resting ST-segment depression (greater than or equal to 0.10 mV), LV hypertrophy, bundle-branch block, intraventricular conduction defect, preexcitation, or digoxin who are able to exercise. In patients undergoing a low-level exercise test, imaging modality may add sensitivity. (Level of Evidence: B)
5. Pharmacological stress testing with imaging when physical limitations (e.g., arthritis, amputation, severe peripheral vascular disease, severe COPD, general debility) preclude adequate exercise stress. (Level of Evidence: B)
6. Prompt angiography without noninvasive risk stratification for failure of stabilization with intensive medical treatment. (Level of Evidence: B)

Class IIa

A noninvasive test (echocardiogram or radionuclide angiogram) to evaluate LV function in patients with definite ACS who are not scheduled for coronary arteriography and left ventriculography. (Level of Evidence: C)

The management of ACS patients requires continuous risk stratification. Important prognostic information is derived from careful initial assessment, the patient's course during the first few days of management, and the patient's response to anti-ischemic and antithrombotic therapy. The Braunwald classification ^(8,111a) has been validated prospectively and represents an appropriate clinical instrument to help predict outcome ⁽²⁵³⁾. Angina at rest, within 48 h in the absence of an extracardiac condition (primary UA) (Braunwald Class III), and UA in the early postinfarction period (Braunwald Class C), along with age, male sex, hypertension, and maximal intravenous antianginal/anti-ischemic therapy, were independent predictors for death or nonfatal MI. The baseline ECG on presentation was also found to be extremely useful for risk stratification in the TIMI III registry ⁽⁶⁰⁾. For example, patients with ST-segment depression of greater than or equal to 0.1 mV had an 11% rate of death or nonfatal MI at 1 year. Those with LBBB had rates of 22.9%. The majority of patients had no ECG change or only isolated T-wave changes, with 6.8% to 8.2% rates of death or MI, respectively, at 1 year. In another study, the rates of death or MI associated with these initial ECG findings in ACS patients were even higher ⁽²⁵⁴⁾ (Figure 11). In many cases, noninvasive stress testing provides a very useful supplement to such clinically based risk assessment. In addition, as pointed out previously, troponins are very helpful in risk assessment.

Some patients, however, are at such high risk for an adverse outcome that noninvasive risk stratification would not be likely to identify a subgroup with sufficiently low risk to avoid coronary angiography to determine whether revascularization is possible. These patients include those who manifest, despite intensive medical therapy, recurrent rest angina, hemodynamic compromise, or severe LV dysfunction. Such patients should be considered directly for early coronary angiography without noninvasive stress testing. However, referral for coronary angiography is not reasonable if they are unwilling to consider revascularization or have severe complicating illnesses that preclude revascularization. Other patients may have a very low likelihood of CAD after the initial clinical evaluation with the risk of an adverse outcome so low that no abnormal test finding would be likely to prompt therapy that would further reduce the already very low risk for adverse outcomes (e.g., a 35-year-old woman without CAD risk factors). Such patients would ordinarily not be considered for coronary angiography and revascularization unless the diagnosis is unclear. Patients who do not fall into these categories are reasonable candidates for risk stratification with noninvasive testing.

1. Care Objectives
The goals of noninvasive testing are to 1) determine the presence or absence of ischemia in patients with a low likelihood of CAD and 2) estimate prognosis. This information is key for the development of further diagnostic steps and therapeutic measures.

A detailed discussion of noninvasive stress testing in CAD is presented in the ACC/AHA Guidelines for Exercise Testing, ACC/AHA Guidelines for the Clinical Use of Cardiac Radionuclide Imaging, and ACC/AHA Guidelines for the Clinical Application of Echocardiography ^(255-257) (Tables 17 to 19). Briefly, the provocation of ischemia at a low workload, such as less than or equal to 6.5 metabolic equivalents (METs), a high-risk treadmill score (greater than or equal to 11) ⁽²⁵⁸⁾, implies severe limitation in the ability to increase coronary blood flow. This is usually the result of severe coronary artery obstruction and is associated with a high risk for adverse outcome and/or severe angina after discharge. Unless there are contraindications to revascularization, such patients generally merit referral for early coronary angiography to direct a revascularization procedure, if possible. On the other hand, the attainment of a higher workload (e.g., greater than 6.5 METs) without evidence of ischemia (low-risk treadmill score greater than or equal to 5) ⁽²⁵⁸⁾ is associated with functionally less severe coronary artery obstruction. Such patients have a better prognosis and can often be safely managed conservatively. Ischemia that develops at greater than 6.5 METs may be associated with severe coronary artery obstruction, but unless other high-risk markers are present (greater than 0.2-mV ST-segment depression or elevation, fall in blood pressure, ST-segment shifts in multiple leads reflecting multiple coronary regions, or prolonged [greater than 6 min of ST-segment shifts] recovery), these patients may also be safely managed conservatively (Table 17).

Stress radionuclide ventriculography or stress echocardiography (Table 18) provides an important alternative. Myocardial perfusion imaging with pharmacological stress (Table 19) is particularly useful in patients unable to exercise. The prognostic value of pharmacological stress testing appears similar to that of exercise testing with imaging, although there are few direct comparisons.

2. Noninvasive Test Selection
There are no conclusive data that either LV function or myocardial perfusion at rest and during exercise or pharmacological stress is superior in the assessment of prognosis. Both the extent of CAD and the degree of LV dysfunction are important in the selection of the appropriate therapy. Studies that directly compare prognostic information from multiple noninvasive tests for ischemia in patients after the stabilization of UA are hampered by small sample size. An exception may be the initial improved LV function, as seen with dobutamine stress echocardiography, which then deteriorates with increasing dobutamine doses ⁽²⁵⁶⁾. This test is particularly useful in patients with good acoustical windows because both resting LV function and the functional consequences of a coronary stenosis can be assessed.

The RISC study evaluated predischarge symptom-limited bicycle exercise testing in 740 men with UA/NSTEMI ⁽²⁵⁹⁾. Multivariate analysis showed that the extent of ST-segment depression expressed as the number of leads that showed ischemia at a low maximal workload was independently negatively correlated with infarct-free survival rates at 1 year. This and other smaller studies permit a comparison of the effectiveness of exercise ECG with exercise or dipyridamole thallium-201 study for risk stratification. All of these noninvasive tests show similar accuracy in dichotomization of the total population into low- and high-risk subgroups.

Selection of the noninvasive stress test should be based primarily on patient characteristics, local availability, and expertise in interpretation ⁽²⁶⁰⁾. Because of simplicity, lower cost, and widespread familiarity with performance and interpretation, the standard low-level exercise ECG stress test remains the most reasonable test in patients who are able to exercise and have a resting ECG that is interpretable for ST-segment shifts. Patients with an ECG pattern that would interfere with interpretation of the ST segment should have an exercise test with imaging. Patients who are unable to exercise should have a pharmacological stress test with imaging. A low-level exercise test (e.g., to completion of Bruce Stage II) may be carried out in low-risk patients (Table 6) who have been asymptomatic for 12 to 24 h. A symptom-limited test can be conducted in patients without evidence of ischemia for 7 to 10 days.

The optimal testing strategy in women remains less well defined than that in men (see Section VI. A), but there is evidence that imaging studies are superior to exercise ECG in women ^(260,261). Exercise testing has been reported to be less accurate for diagnosis in women. At least a portion of the lower reported accuracy derives from a lower pretest likelihood of CAD in women compared with men.

Results of a symptom-limited exercise test performed 3 to 7 days after UA/NSTEMI were compared with results of a test conducted 1 month later in 189 patients ⁽²⁶²⁾. The diagnostic and prognostic values of the tests were similar, but the earlier test identified patients who developed adverse events during the first month, and this represented about one half of all events that occurred during the first year. These data illustrate the importance of early noninvasive testing for risk stratification.

The Veterans Affairs Non-Q-Wave Infarction Strategies in Hospital (VANQWISH) trial used symptom-limited thallium exercise treadmill testing at 3 to 5 days to direct the need for angiography in the 442 non-Q-wave MI patients randomized to an early conservative strategy ⁽²⁶³⁾. This strategy included an effort to detect ischemia with noninvasive testing that would be associated with a high risk for adverse outcome. Cumulative death rates in the 238 conservative strategy patients directed to angiography on the basis of recurrent ischemia or high-risk stress test results were 3%, 10%, and 13% at 1, 6, and 12 months, respectively, whereas the rates were 1%, 3%, and 6% in the patients who were not directed to angiography (no recurrent ischemia or high-risk test). These findings support the concept that noninvasive stress testing can be used successfully to identify a high-risk subset of patients who could be directed to coronary angiography. It is unlikely that any angiographically directed early revascularization strategy could alter the very low early event rates observed in patients without a high-risk stress test.

Noninvasive tests are most useful for management decisions when risk can be stated in terms of events over time. A large population of patients must be studied to derive and test equations needed to accurately predict individual patient risk. No noninvasive study has been reported in a sufficient number of patients after the stabilization of UA to develop and test the accuracy of a multivariable equation to report test results in terms of absolute risk. Therefore, data from studies of stable angina patients must be used for risk reported as events over time. Although the pathological process that evokes ischemia may be different in the 2 forms of angina, it is likely that the use of prognostic nomograms derived from patients with stable angina are also predictive of risk in patients with recent UA after stabilization. With this untested assumption, the much larger literature derived from populations that include patients with both stable angina and UA provides equations for risk stratification that convert physiological changes observed during noninvasive testing into statements of risk expressed as events over time.

3. Selection for Coronary Angiography
In contrast to the noninvasive tests, coronary angiography provides detailed structural information to allow an assessment of prognosis and to provide direction for appropriate management. When combined with LV angiography, it also allows an assessment of global and regional LV function. Indications for coronary angiography are interwoven with indications for possible therapeutic plans such as PCI or CABG. The recently revised ACC/AHA Guidelines for Coronary Angiography present greater details on this subject ⁽²⁶⁴⁾.

Coronary angiography is usually indicated in patients with UA/NSTEMI who either have recurrent symptoms or ischemia despite adequate medical therapy or are at high risk categorized by clinical findings (CHF, malignant ventricular arrhythmias) or noninvasive test findings (significant LV dysfunction: ejection fraction [EF] less than 0.35, large anterior or multiple perfusion defects) (Tables 17 to 19), as discussed in Section III. B. Patients with UA who have had previous PCI or CABG should also in general be considered for early coronary angiography, unless prior coronary angiography data indicate that no further revascularization is likely to be possible. The placement of an IABP may be useful in patients with recurrent ischemia despite maximal medical management as well as in those with hemodynamic instability until coronary angiography and revascularization can be completed. Patients with suspected Prinzmetal's variant angina are also candidates for coronary angiography (see Section VI. F).

In all cases, the general indications for coronary angiography and revascularization are tempered by individual patient characteristics and preferences. Patient and physician judgments regarding risks and benefits are particularly important for patients who may not be candidates for coronary revascularization, such as very frail elderly persons and those with serious comorbid conditions (i.e., severe hepatic, pulmonary, or renal failure; active or inoperable cancer).

4. Patient Counseling
Results of testing should be discussed with the patient, his or her family, and/or his or her advocate in language that is understood. Test results should be used to help determine the advisability of coronary angiography, the need for adjustments in the medical regimen, and the need for secondary prevention measures (see Section V).

D. Early Conservative Versus Invasive Strategies
1. General Principles
Two different treatment strategies, termed "early conservative" and "early invasive," have evolved for patients with UA/NSTEMI. In the early conservative strategy, coronary angiography is reserved for patients with evidence of recurrent ischemia (angina at rest or with minimal activity or dynamic ST-segment changes) or a strongly positive stress test, despite vigorous medical therapy. In the early invasive strategy, patients without clinically obvious contraindications to coronary revascularization are routinely recommended for coronary angiography and angiographically directed revascularization if possible.

Recommendations:

Class I

1. An early invasive strategy in patients with UA/NSTEMI and any of the following high-risk indicators. (Level of Evidence: A)
   1. Recurrent angina/ischemia at rest or with low- level activities despite intensive anti-ischemic therapy
   2. Elevated TnT or TnI
   3. New or presumably new ST-segment depression
   4. Recurrent angina/ischemia with CHF symptoms, an S₃ gallop, pulmonary edema, worsening rales, or new or worsening MR
   5. High-risk findings on noninvasive stress testing
   6. Depressed LV systolic function (e.g., EF less than 0.40 on noninvasive study)
   7. Hemodynamic instability
   8. Sustained ventricular tachycardia
   9. PCI within 6 months
   10. Prior CABG
2. In the absence of these findings, either an early conservative or an early invasive strategy in hospitalized patients without contraindications for revascularization. (Level of Evidence: B)

Class IIa

1. An early invasive strategy in patients with repeated presentations for ACS despite therapy and without evidence for ongoing ischemia or high risk. (Level of Evidence: C)

Class III

1. Coronary angiography in patients with extensive comorbidities (e.g., liver or pulmonary failure, cancer), in whom the risks of revascularization are not likely to outweigh the benefits. (Level of Evidence: C)
2. Coronary angiography in patients with acute chest pain and a low likelihood of ACS. (Level of Evidence: C)
3. Coronary angiography in patients who will not consent to revascularization regardless of the findings. (Level of Evidence: C)

a. Rationale for the Early Conservative Strategy
Three multicenter trials have shown similar outcomes with early conservative and early invasive therapeutic strategies ^(19,265,266). The conservative strategy spares the use of invasive procedures with their risks and costs in all patients. Recent trials ^(266,267) have emphasized the early risk associated with revascularization procedures. When the early conservative strategy is chosen, a plan for noninvasive evaluation is required to detect severe ischemia that occurs spontaneously or at a low threshold of stress and to promptly refer these patients for coronary angiography and revascularization when possible. In addition, as in STEMI ⁽²⁶⁸⁾, an early echocardiogram should be considered to identify patients with significant LV dysfunction (e.g., EF less than 0.40). Such a finding prompts consideration for angiography to identify left main or multivessel CAD, because patients with multivessel disease and LV dysfunction are at high risk and may accrue a survival benefit from bypass surgery ^(269,270). In addition, a stress test (e.g., exercise or pharmacological stress) for the assessment of ischemia is recommended before discharge or shortly thereafter to identify patients who may also benefit from revascularization. The use of either LMWH or platelet GP IIb/IIIa receptor blockers has reduced the incidence of adverse outcomes in patients managed conservatively (see Section III. B) ^{(10,169-171,182,184,247,248)}, suggesting that the early conservative strategy may be advantageous because costly invasive procedures may be avoided in even more patients.

b. Rationale for the Early Invasive Strategy
In patients with UA/NSTEMI without recurrent ischemia in the first 24 h, the use of early angiography provides an invasive approach to risk stratification. It can identify the 10% to 15% of patients with no significant coronary stenoses and the approximately 20% with 3-vessel disease with LV dysfunction or left main CAD. This latter group may derive a survival benefit from bypass surgery (see Section IV). In addition, early percutaneous revascularization of the culprit lesion has the potential to reduce the risk for subsequent hospitalization and the need for multiple antianginal drugs compared with the early conservative strategy (TIMI IIIB) ⁽¹⁹⁾. Just as the use of improved antithrombotic therapy with LMWH and/or a platelet GP IIb/IIIa receptor blocker has improved the outcome in patients managed according to the early conservative strategy, the availability of these agents also makes the early invasive approach more attractive, because the early hazard of percutaneous intervention is lessened. The availability of GP IIb/IIIa receptor blockers has led to 2 alternatives for the invasive approach: immediate angiography or deferred angiography.

c. Immediate Angiography
Some believe that proceeding immediately to angiography is an efficient approach for the ACS patient. Patients found not to have CAD may be discharged rapidly or shifted to a different management strategy. Patients with obvious culprit lesions amenable to percutaneous intervention could have a procedure performed immediately, thus hastening discharge. Patients with left main CAD and patients with multivessel disease and LV dysfunction could be sent expeditiously to undergo bypass surgery, thereby avoiding a risky waiting period. However, only 1 observational study ⁽²⁷¹⁾ has addressed this approach directly, and the results are not definitive.

d. Deferred Angiography
In most reports that involve use of the early invasive strategy, angiography has been deferred for 12 to 48 h while antithrombotic and anti-ischemic therapies are intensified. Several observational studies ⁽⁵⁵²⁾ have found a lower rate of complications in patients undergoing percutaneous intervention more than 48 h after admission, during which heparin and ASA were administered, compared with early intervention. However, it should be noted that the value of medical stabilization before angiography has never been assessed formally.

2. Care Objectives
The objective is to provide a strategy that has the most potential to yield the best clinical outcome. The purpose of coronary angiography is to provide detailed information about the size and distribution of coronary vessels, the location and extent of atherosclerotic obstruction, and the suitability for revascularization. The LV angiogram, which is usually carried out along with coronary arteriogram, provides an assessment of the extent of focal and global LV dysfunction and of the presence and severity of coexisting disorders (e.g., valvular or congenital lesions). A detailed discussion of revascularization is presented in Section IV of these guidelines, as well as in the ACC/AHA Guidelines for Percutaneous Transluminal Coronary Angioplasty ⁽⁵⁵²⁾ and ACC/AHA Guidelines for Coronary Artery Bypass Graft Surgery ⁽²⁷⁴⁾. Although general guidelines can be offered, the selection of appropriate procedures and the decision to refer patients for revascularization require both clinical judgment and counseling with the patient and his or her family regarding expected risks and benefits. In this counseling, it is important to consider that large registry and controlled clinical trial data generally show no or limited evidence of reduced death or MI rates when early coronary angiography followed by revascularization is used in a routine and unselected manner in patients with UA/NSTEMI.

Because the basis for acute ischemia is plaque rupture/erosion and/or severe obstructive CAD, it has been postulated that early revascularization would improve prognosis. This notion led to considerable investigation in patients with stable coronary syndromes as well as AMI in the 1970s. In selected circumstances, revascularization with CABG seems to be associated with lower morbidity and mortality rates compared with a more conservative strategy. These circumstances center around the documentation of severe ischemia (resting ECG and on noninvasive testing) or potential for severe ischemia (left main stenosis or severe multivessel CAD with impaired LV function). These data, however, are limited because both medical and surgical treatments have been markedly improved in the past 2 decades and the population of patients who present with CAD today has changed (e.g., a higher proportion of women, elderly persons, minorities, and diabetics).

Although 2 recent studies were not conducted in patients with UA/NSTEMI, they have addressed the value of stress testing in guiding therapy. The DANish trial in Acute Myocardial Infarction (DANAMI) studied 503 patients with inducible ischemia (i.e., a positive exercise stress test) after thrombolytic therapy for first MI and compared an ischemia-guided invasive strategy with a conservative strategy ⁽²⁷⁵⁾. The invasive strategy in the post-AMI patients with inducible ischemia resulted in a reduction in the incidence of reinfarction, hospitalizations for UA, and stable angina. Similarly, in the Asymptomatic Cardiac Ischemia Pilot (ACIP) ^(276,277), 558 clinically stable patients with ischemia on stress testing and during daily life (ST-segment depression on exercise treadmill testing or perfusion abnormality on radionuclide pharmacological stress test if unable to exercise, in addition to ST-segment depression on ambulatory ECG monitoring), of whom most had angina in the previous 6 weeks, were randomized to 1 of 3 initial treatment strategies: symptom-guided medical care, ischemia-guided medical care, and revascularization. More than one third of these patients had "complex" stenoses on angiography. Those randomized to early revascularization experienced less ambulatory ischemia at 12 weeks than did those randomized to initial medical care in whom revascularization was delayed and symptom driven.

In ACS patients with UA/NSTEMI, the purpose of noninvasive testing is to identify ischemia as well as to identify candidates at high risk for adverse outcome and to direct them to coronary angiography and revascularization when possible. However, both randomized trials ^{(19,265,266,278)} and observational data ^(279-281) do not uniformly support an inherent superiority for the routine use of early coronary angiography and revascularization. In fact, the VANQWISH trial suggests that an early conservative strategy, in which candidates for coronary angiography and revascularization are selected from the results of ischemia-guided noninvasive testing, may be associated with fewer deaths. Accordingly, the decision regarding which strategy to pursue for a given patient should be based on the patient's estimated outcome risk assisted by clinical and noninvasive test results, available facilities, previous outcome of revascularization by the team available and in the institution in which the patient is hospitalized, and patient preference.

Early coronary angiography may enhance prognostic stratification. This information may be used to guide medical therapy as well as to plan revascularization therapy, but it is important to emphasize that adverse outcome in ACS is very time dependent and that after 1 to 2 months, the risk for adverse outcome is essentially the same as that for low-risk chronic stable angina (Figure 3). Furthermore, numerous studies in patients with stable angina, including Research Group in Instability in Coronary Artery Disease (RITA)-2 ⁽²⁶⁷⁾, have documented the significant early risk of death or MI with an interventional strategy compared with medical management alone. Thus, the timing of coronary angiography and revascularization is critically important if patients at high risk are to benefit. Unfortunately, the total number of operative complications is increased when revascularization procedures are performed routinely, because some patients who are not in need of revascularization will be exposed to its hazards.

The population of patients with UA/NSTEMI includes a subgroup (i.e., those with left main coronary stenosis or multivessel stenoses with reduced LV function) at high risk for adverse outcome and therefore highly likely to benefit from revascularization. Clinical evaluation and noninvasive testing will aid in the identification of most of these high-risk patients who have markers of high risk such as advanced age (greater than 70 years), prior MI, revascularization, ST-segment deviation, CHF or depressed resting LV function (i.e., EF less than 0.40) on noninvasive study, or noninvasive stress test findings that suggest severe ischemia (see Section III. C). The remaining larger subgroup of patients, however, do not have the findings that portend a high risk for adverse outcomes. Accordingly, they are not likely to receive such benefit from routine revascularization, and invasive study is optional for them. It can be safely deferred pending further clinical developments. Decisions regarding coronary angiography in patients who are not high risk according to findings on clinical examination and noninvasive testing can be individualized on the basis of patient preferences.

The data on which these recommendations are based are from 4 randomized trials, TIMI IIIB ⁽¹⁹⁾, VANQWISH ⁽²⁶⁶⁾, Medicine versus Angiography in Thrombolytic Exclusion (MATE) ⁽²⁶⁵⁾, and FRISC II ⁽²⁷⁸⁾; a large prospective multinational registry, the OASIS registry ⁽²⁷⁹⁾; and 2 retrospective analyses ^(280,281).

In TIMI IIIB, 1,473 patients with UA (67%) or NSTEMI (33%) with chest pain of less than 24-h duration were randomized to either an invasive or early conservative strategy. At 42 days, 16.2% of the early invasive patients had died, had experienced a nonfatal MI, or had a strongly positive exercise test vs. 18.1% of early conservative patients (p = 0.33). Similarly, there was no difference in the outcome of death or MI in a comparison of treatment strategies (4,282). An analysis of factors associated with the failure of medical therapy in TIMI IIIB predicted patients who could be directed to a more invasive strategy in a cost-efficient manner. Among the 733 patients randomized to the conservative strategy, the factors that independently predicted failure of medical therapy included ST-segment depression on the qualifying ECG, prior ASA use, and older age. For most patients with 3 or more such risk factors, medical therapy had failed, defined as death, MI, rest angina, or markedly abnormal stress test results at 6 weeks. A combination of factors should be considered in the selection of patients for expedited angiography and revascularization ⁽²⁸²⁾.

NSTEMI represents a high-risk acute ischemic syndrome. The VANQWISH Investigators randomly assigned 920 patients with NSTEMI defined on the basis of CK-MB to either early invasive (462 patients) or conservative (458 patients) management within 72 h of the onset of an NSTEMI ⁽²⁶⁶⁾. The number of patients with either death or recurrent nonfatal MI and the number who died were higher in the invasive strategy group at hospital discharge (36 vs. 15 patients, p = 0.004 for death or nonfatal MI; 21 vs. 6, p = 0.007 for death), and these differences persisted at 1 month and at 1 year. Mortality rates during the almost 2-year follow-up also showed a strong trend toward reduction in patients assigned to the conservative strategy compared with those assigned to the invasive strategy (hazard ratio, 0.72; 95% CI, 0.51 to 1.01). The investigators concluded that most patients with NSTEMI do not benefit from routine, early invasive management and that a conservative, ischemia-guided initial approach is both safe and effective even in the predominantly high-risk male population of the VANQWISH.

The MATE trial ⁽²⁶⁵⁾ enrolled 201 patients with a variety of ACSs who were ineligible for thrombolytic therapy and were assigned to an early invasive or early conservative strategy. Although the incidence of total ischemic in-hospital events was lower in the early invasive strategy group, there were no differences between the groups in the incidence of death and reinfarction. Follow-up at a median of 21 months showed no significant differences in the cumulative incidences of death, MI, rehospitalization, or revascularization.

Most recently, in the FRISC II study, 3,048 ACS patients were treated with dalteparin for 5 to 7 days ⁽²⁷⁸⁾. Of these patients, 2,457 without acute problems who were not at high risk of a revascularization procedure (e.g., their age was not greater than 75 years, and they did not have prior CABG) were randomized (2 × 2 factorial design) to continue to receive either dalteparin or placebo (double blind) and either an invasive or a noninvasive treatment strategy. The latter patients were revascularized only for refractory or recurrent symptoms despite maximum medical therapy or severe ischemia (ST-segment depression greater than or equal to 0.3 mV) on symptom-limited exercise testing or AMI. At 6 months, there were no differences between continued dalteparin compared with placebo. However, death or MI occurred in 9.4% of patients assigned to the invasive strategy and in 12.1% of those assigned the noninvasive strategy (p less than 0.031). At 1 year the mortality rate in the invasive strategy group was 2.2% compared with 3.9% in the noninvasive strategy group (p = 0.016) ^(278a). It may be concluded from FRISC II that patients with UA/NSTEMI who are not at very high risk for revascularization and who first receive an average of 6 days of treatment with LMWH, ASA, nitrates, and beta-blockers have a better outcome at 6 months with a (delayed) routine invasive approach than with a routine conservative approach.

In the TACTICS-TIMI 18 trial ⁽⁵¹⁸⁾, 2,220 patients with UA or NSTEMI were treated with ASA, heparin, and the GP IIb/IIIa inhibitor tirofiban. They were randomized to an early invasive strategy with routine coronary angiography within 48 h followed by revascularization if the coronary anatomy was deemed suitable, or to a more conservative strategy. In the latter, catheterization was performed only if the patient had recurrent ischemia or a positive stress test. Death, MI, or rehospitalization for ACS at 6 months occurred in 15.9% of patients assigned to the invasive strategy vs. 19.4% assigned to the more conservative strategy (p = 0.025). Death or MI ⁽⁵³⁹⁾ was also reduced at 6 months (7.3% vs 9.5%, p less than 0.05). The beneficial effects on the outcome were observed in medium- and high-risk patients, as defined by an elevation of TnT greater than 0.01 ng per ml, the presence of ST-segment deviation, or a TIMI risk score of greater than 3 ⁽⁵¹⁷⁾. In the absence of these high-risk features, outcomes in patients assigned to the 2 strategies were similar. Rates of major bleeding were similar, and lengths of hospital stay were reduced in patients assigned to the invasive strategy. The benefits of the invasive strategy were achieved at no significant increase in the costs of care over the 6-month follow-up period.

Thus, both the FRISC II ⁽²⁷⁸⁾ and TACTICS-TIMI 18 ⁽⁵¹⁸⁾ trials, the 2 most recent trials comparing invasive vs. conservative strategies in patients with UA/NSTEMI, showed a benefit in patients assigned to the invasive strategy. In contrast to earlier trials, a large majority of patients undergoing PCI in these 2 trials received coronary stenting as opposed to balloon angioplasty alone. In FRISC II, the invasive strategy involved treatment for an average of 6 days in the hospital with LMWH, ASA, nitrates, and beta blockers prior to coronary angiography, an approach that would be difficult to adopt in US hospitals. In TACTICS-TIMI 18, treatment included the GP IIb/IIIa antagonist tirofiban, which was administered for an average of 22 h prior to coronary angiography. The routine use of the GP IIb/IIIa inhibitor in this trial may have eliminated the excess risk of early (within 7 days) acute MI in the invasive arm, an excess risk that was observed in FRISC II and other trials in which there was no routine "upstream" use of a GP IIb/IIIa blocker. Therefore, an invasive strategy is associated with a better outcome in UA/NSTEMI patients at high risk as defined in Table 6 and in TACTICS-TIMI 18 and who receive a GP IIb/IIIa inhibitor ⁽⁵¹⁸⁾. Although the benefit of intravenous GP IIb/IIIa inhibitors is established for UA/NSTEMI patients undergoing PCI, the optimum time to commence these drugs before the procedure has not been established. In the PURSUIT trial ⁽¹⁰⁾, in patients with UA/NSTEMI who were admitted to community hospitals, the administration of eptifibatide was associated with a reduced need for transfer to tertiary referral centers and improved outcomes ⁽⁵⁴¹⁾.

Some selected areas require additional comment. In a patient with UA, a history of prior PCI within the past 6 months suggests the presence of restenosis, which often can be effectively treated with repeat PCI. Coronary angiography without preceding functional testing is generally indicated. Patients with prior CABG represent another subgroup for whom a strategy of early coronary angiography is usually indicated. The complex interplay between the progression of native coronary disease and the development of graft atherosclerosis with ulceration and embolization is difficult to untangle noninvasively; all argue for early coronary angiography. In addition, patients with known or suspected reduced LV systolic function, including patients with prior anterior Q-wave MIs, those with prior measurements that show depressed LV function, and those who present with CHF, have sufficient risk that the possibility of benefit from revascularization procedures merits early coronary angiography without preceding functional testing.

In patients with UA/NSTEMI, coronary angiography typically shows the following profile: 1) no severe epicardial stenosis in 10% to 20%, 2) 1-vessel stenosis in 30% to 35%, 3) multivessel stenosis in 40% to 50%, and 4) significant (greater than 50%) left main stenosis in 4% to 10%. In the early invasive strategy in TIMI IIIB, no critical obstruction (less than 60% diameter stenosis) was found in 19% of patients, 1-vessel stenosis in 38%, 2-vessel stenosis in 29%, 3-vessel stenosis in 15%, and left main stenosis (greater than 50%) in 4%. Complex plaques are usually believed to be responsible for the culprit lesions. These usually are eccentric and sometimes have irregular borders, and correlate with intracoronary thrombi and an increased risk of recurrent ischemia at rest, MI, and cardiac death ⁽²⁸³⁾. Similar findings were noted in more than 80% of the patients in the VANQWISH trial, and more than 1 complex lesion was found in most patients ⁽²⁸⁴⁾. Interestingly, in TIMI IIIB, many of the patients without severe stenosis had reduced contrast clearance, which suggests microvascular dysfunction ⁽²⁸⁵⁾, which may contribute to impaired myocardial perfusion.

Patients with severe 3-vessel stenosis and reduced LV function and those with left main stenosis should be considered for early CABG (see Section IV). In low-risk patients, quality of life and patient preferences should be given considerable weight in the selection of a treatment strategy. Low-risk patients whose symptoms do not respond well to maximal medical therapy and who experience poor quality of life and functional status and are prepared to accept the risks of revascularization should be considered for revascularization.

The discovery that a patient does not have significant obstructive CAD can help avert improper "labeling" and prompt a search for the true cause of symptoms. Unfortunately, many such patients continue to have recurrent symptoms, become disabled, are readmitted to the hospital, and continue to consume healthcare resources even with repeated coronary angiography ^(286,287).

It is not presently possible to define the extent of comorbidity that would, in every case, make referral for coronary angiography and revascularization inappropriate. The high-risk patient with significant comorbidities requires thoughtful discussion among the physician, patient, and family and/or patient advocate. A decision for or against revascularization must be made on a case-by-case basis.

Examples of extensive comorbidity that usually preclude revascularization include 1) advanced or metastatic malignancy with a projected life expectancy of less than or equal to 1 year, 2) intracranial pathology that contraindicates the use of systemic anticoagulation or causes severe cognitive disturbance (e.g., Alzheimer's disease) or advanced physical limitations, 3) end-stage cirrhosis with symptomatic portal hypertension (e.g., encephalopathy, visceral bleeding), and 4) CAD that is known from previous angiography not to be amenable to revascularization. This list is not meant to be all inclusive. More difficult decisions involve patients with comorbidities not as serious as those listed here; examples include patients who have moderate or severe renal failure but are stable on dialysis.

Consultation with an interventional cardiologist and cardiac surgeon before coronary angiography is advised to define technical options and likely risks and benefits. The operators who perform coronary angiography and revascularization and the facility in which these procedures are carried out are important considerations because the availability of interventional cardiologists and cardiac surgeons who are experienced in high-risk and complex patients is essential. As a general principle, the potential benefits of coronary angiography and revascularization must be carefully weighed against the risks and the conflicting results of the clinical trials and registries.