The Clinical Challenge of Heart Failure: Comprehensive, Evidence-Based Management of the Hospitalized Patient With Acute Myocardial Decompensation—Diagnosis, Risk Stratification, and Outcome-Effective Treatment

Part I: Presentation, Differential Diagnosis, Laboratory Examination, and Prophylaxis Against Venous Thromboembolic Disease (VTED)

Authors: W. Frank Peacock, MD, FACEP, Emergency Department Clinical Operations Director, The Cleveland Clinic, OH; Benjamin J. Freda, DO, Department of Internal Medicine, Cleveland Clinic Foundation, OH.

Peer Reviewers: Charles L. Emerman, MD, Chairman, Department of Emergency Medicine, MetroHealth Medical Center, Cleveland Clinic Foundation, OH; Gregory A. Volturo, MD, FACEP, Vice Chairman and Professor, Department of Emergency Medicine, University of Massachusetts Medical School, Worcester.

Editorial Note: Pump failure, shock, dyspnea, anxiety, pulmonary edema, arrhythmias, venous thromboembolic disease, drug-related adverse events, and sometimes death—these clinical hallmarks of acute, decompensated heart failure are well known to the emergency physician, cardiologist, hospitalist, and intensivist. With nearly 5 million patients currently diagnosed with heart failure (HF) in the United States—and almost 500,000 new cases identified each year—this country is experiencing, and will continue to witness, a cardiovascular epidemic of staggering proportions.1 The facts surrounding this debilitating and, more often than not, life-threatening condition are enough to concern even the most confident clinician. Consider the following: The mortality rate of HF not only exceeds most cancers,1 but this condition is responsible for the debilitation and, eventually, death of more elderly individuals than any other single ailment.

Almost 300,000 patients die from HF or its complications each year, and this disease accounts for more hospitalizations (900,000 annually) and re-hospitalizations than any other illness in patients older than age 65.1 The financial burden that HF places on the Medicare system is as great as the two most common cancers—breast and lung—combined. In fact, at the 1999 Heart Failure Society of America Third Annual Scientific Assembly, it was estimated that inpatient costs for HF in that year would be $23.1 billion; it was estimated an additional $14.7 billion would be required for outpatient management of this growing patient population.

Although the incidence of HF is increasing at a dramatic pace, and its societal impact is as great as any other chronic illness, surprisingly few articles have been published in the emergency medicine literature outlining a systematic approach to managing patients with this life-threatening condition. This is unfortunate, because the diagnostic and therapeutic landscape for HF is undergoing a dramatic shift. Established therapies that include diuretics, inotropic agents, oxygen, and nitroglycerin are being supplemented by newer approaches that produce more rapid reductions in pulmonary wedge pressures and reduce such complications as deep venous thrombosis (DVT) and pulmonary embolism (PE) in patients at high risk. Considering recent advances in the comprehensive, multi-factorial management of patients with HF—among them, recent approval of nesiritide for treatment of acute, decompensated HF, as well as approval for enoxaparin to prevent DVT in hospitalized patients, including those with New York Heart Association (NYHA) Grade III/IV congestive heart failure (CHF)—there is an urgent need to review current, evidence-based strategies for optimizing outcomes in patients with HF.

Of the two newest strategies introduced for managing patients with HF, nesiritide (B-type natriuretic peptide [BNP]) is the first medication approved by the U.S. Food and Drug Administration (FDA) in more than a decade for treatment of the acutely decompensated patient. The recombinant DNA-manufactured form of endogenously synthesized BNP, nesiritide represents an amplification of the natural compensatory mechanism for neurohormonal and hemodynamic derangements that occur in HF. The results of such studies as the PRECEDENT trial have been strongly supportive of adopting nesiritide as a standard of care for an appropriately selected patient population. In this trial, which evaluated and compared varying doses of nesiritide or dobutamine, patients treated with dobutamine had higher ectopy rates, were more likely to suffer cardiac arrest, and had a higher rate of ventricular tachycardia as compared to nesiritide. Another study, evaluating six-month survival following short-term nesiritide therapy, showed a marked decrease in death rates compared to dobutamine. In short, there is evidence suggesting nesiritide is superior to the most commonly used inotropes for managing patients with HF.

Although augmentation and stabilization of left ventricular (LV) pump function, reduction of arrhythmia risk, and hemodynamic stabilization with pharmacological and/or surgical intervention remain the primary clinical and functional objectives of managing patients with HF, recent attention has been focused on preventing other complications, including venous thromboembolic disease (VTED), especially upon admission to the hospital setting. Not surprisingly, identifying a cost-effective strategy for preventing DVT and/or PE in HF patients has been an area of intense interest, inasmuch as recent consensus guidelines have introduced thromboprophylaxis recommendations with either low molecular weight heparin (LMWH) or unfractionated heparin (UFH) for general medical patients admitted to the hospital with risk factors for VTED; in all such guidelines, HF is cited as one of the principal indications.2

Based on such recommendations and other recent studies, there is compelling evidence that in immobilized patients—or those with significant restrictions in ambulation that place them at risk for DVT—who present to the hospital with CHF (NYHA Class III-IV), prophylaxis should be considered mandatory if there are no significant contraindications.3 Moreover, it should be emphasized that the American College of Chest Physicians (ACCP) guidelines and International Consensus Statement also cite CHF as a risk factor for VTED and emphasize the importance of this risk factor when assessing prophylaxis requirements for hospitalized medical patients.2,4

Clinical trials, albeit small, specifically have evaluated patients with HF. The PRINCE (Prevention in Cardiopulmonary Disease with Enoxaparin) study group conducted a randomized, multicenter trial in 665 hospitalized patients with severe cardiopulmonary diseases (332 with respiratory disorders, 333 with NYHA Grade III/IV heart failure).5 Patients were treated with enoxaparin (a type of LMWH) 40 mg subcutaneously (SC) daily or 5000 IU UFH SC TID for 8-12 days in an open-label study. Efficacy rates were evaluated in 454 patients using an intention-to-treat analysis. Although the rate of VTE was similar in both groups (8.4% in the enoxaparin group vs 10.4% in the UFH group), a subgroup analysis indicated that the incidence of VTE was greater in patients with HF than those with respiratory diseases.5 These and other studies stress the need for a comprehensive approach to the management of HF, i.e., one that includes pharmacological interventions aimed at improvement of LV function and overall functional status, as well as those aimed at preventing complications commonly encountered in hospitalized patients with HF.

Overall, during the past few years, we have witnessed dramatic progress in both our understanding and treatment of HF. With these advances in clear focus, this landmark review will describe both established and recently introduced strategies for diagnostic evaluation of and therapy for patients with HF. Basing recommendations on evidence-based trials, it will outline promising technological advances that are likely to optimize outcomes in this challenging patient population.— The Editor

Clinical Presentation

From a clinical perspective, HF is characterized by dyspnea, weakness, fatigue, and compromised functional status and results when the myocardium cannot maintain the cardiac output that is required for normal metabolism and venous return. Presentation in the emergency department (ED) is characterized by a continuum that may begin with asymptomatic LV dysfunction; it then may progress to mild symptoms of dyspnea that occur only with significant exertion, or manifest as dyspnea at rest, and eventually terminate with severe LV failure, hypoperfusion, and cardiogenic shock. The acuity of this progression, the spectrum of symptoms, and the level of patient distress are determined by the rapidity at which LV dysfunction occurs. At the critical extreme, in the setting of an acute myocardial infarction (MI), which frequently is characterized by precipitous LV dysfunction, critical loss of myocardial performance results in severe dyspnea, hypotension, and altered mental status. This patient will present in extremis, and has a significant near-term mortality rate.

At the opposite end of the spectrum is the patient with chronic systolic dysfunction. This individual may be a regular or repeat visitor to the ED, and present with symptomatic exacerbation as a result of dietary and/or medication non-compliance, which may lead to edema, orthopnea, and moderate dyspnea on exertion. These patients usually are treated in the ED and hospitalized. However, they occasionally can be discharged home if there is good therapeutic response.

A third category of patients frequently presents to the ED, despite optimal medical care and meticulous medication compliance. These individuals complain of weakness, fatigue, malaise, and a general failure to thrive. Commonly euvolemic, or even dehydrated, their systemic symptoms result from terminal pump dysfunction. At this stage, medical treatment frequently is ineffective. While it is unlikely that this subgroup of patients will succumb in the ED, only invasive therapy (heart transplant, LV assist device, implantable cardiac device, biventricular pacing, and other cardiac devices) can alter what is, inevitably, an unfavorable prognosis.

Heart Failure: Definitions and Categories

The terminology of HF can be confusing. To facilitate accurate communication among health care providers and to orchestrate proper treatment plans, clinicians must be precise with the the broad range of qualifying terms. The descriptor "congestive" HF can be problematic. Loop diuretics may allow patients with HF to avoid congestive symptoms, despite severe LV dysfunction. Consequently, the term "congestive heart failure" describes a clinical state from which a patient with HF may or may not be suffering. Moreover, the terms high and low output, and forward and backward flow HF are best avoided.

Better descriptors are acute or chronic, and systolic or diastolic HF. Finally, some physicians differentiate left- and right-sided HF. Left-sided failure is described as having pulmonary symptoms in the absence of peripheral edema, jugular venous distention (JVD), or hepatojugular reflux (HJR). In contrast, right-sided HF is characterized by peripheral edema, JVD, and HJR, without pulmonary symptoms. Because cardiac architecture, from a mechanical perspective, is a closed system, abnormally elevated cardiac chamber pressures and volumes quickly are reflected into the contralateral system. This distinction should be reserved for patients with flow-limiting lesions (e.g., valvular dysfunction).

Systolic vs. Diastolic HF. In HF, the fundamental abnormality is impaired LV contractility, manifested by a downward shift in the Frank-Starling curve. In a healthy heart, increased preload (i.e., venous return), is associated with improved myocardial contractility (increased inotropicity) and, therefore, an increased stroke volume. This response is lost in HF, and represents the downward shift of the Frank-Starling mechanism. Consequently, any cardiovascular stress (e.g., walking) that increases venous return and cardiac pressures is not met by improved contractility, and ultimately leads to pulmonary congestion and edema.

The Law of LaPlace describes ventricular wall tension as a function of the product of pressure (afterload) and ventricular radius. With increased wall tension as the stimulus, myocardial remodeling occurs. Initially, healthy myocytes hypertrophy. An adverse re-sponse results when myocytes die (apoptosis) and become scar tissue. The underlying stimulus for apoptosis is unclear. The end result of remodeling is the determinant for the type of HF.

A normal ejection fraction (EF) is 60%. This means that the heart empties 60% of its blood volume with each cardiac cycle. Systolic dysfunction is defined as an EF no greater than 40%.6 In systolic HF, the ventricle has difficulty ejecting blood. The underlying pathologic cascade in systolic HF is impaired cardiac contractility, neurohormonal activation, increased intracardiac volume and pressure, and enhanced sensitivity to increased afterload.

Table 1. Diastolic Heart Failure Etiologies38

Restrictive Cardiomyopathy
• Cardiac amyloidosis
Constrictive Pericarditis
Ischemic Heart Disease
• Post-infarction scarring/remodeling
Hypertropic Heart Disease
• Hypertrophic cardiomyopathy
• Chronic hypertenison
• Aortic stenosis
Mitral or Tricuspid Stenosis

In diastolic failure, systolic function is preserved such that the EF may be normal or even supra-normal, but with an abnormal diastolic pressure/volume relationship. The primary pathophysiological abnormality is impaired ventricular relaxation. This results in a left ventricle that has difficulty receiving blood. The decrease in LV compliance from impaired myocardial relaxation necessitates higher atrial pressures to ensure adequate diastolic filling of the left ventricle. The frequency of diastolic dysfunction increases with age.7 Longstanding hypertension accompanied by the development of LV hypertrophy often are responsible for this syndrome. Coronary artery disease also contributes, as diastolic dysfunction is an early event in the ischemic cascade. Common etiologies for diastolic HF are summarized in Table 1. As many as 30-50% of HF patients have circulatory congestion on the basis of diastolic dysfunction8; treatment for volume overload is the same as it is for systolic dysfunction. However, patients with diastolic dysfunction are preload dependent and the use of excessive diuresis or venodilation may exacerbate the underlying deficit in ventricular filling, which can cause hypotension. Ultimately, after hemodynamics have been stabilized and congestion resolved, treatment of diastolic dysfunction requires consideration of therapy directed at the underlying etiology. Determining the type of HF can be difficult using history and physical findings; consequently, an echocardiogram is necessary.

Clinical Pathophysiology

Our understanding of HF has evolved greatly during the past two decades.9 The genesis of HF no longer is explained as simple mechanical pump dysfunction. New studies have enabled investigators to discern the pathophysiological forces contributing to the syndrome. Typically, HF is set into motion by some myocardial injury or stressor. Myocardial injury can act directly on the heart tissue (e.g., ischemia/infarction, autoimmune insult, infectious disease, and other causes); it also can result from hemodynamic stress (chronic hypertension or valvular disease) or from arrythmogenic stress (tachycardia-induced cardiomyopathy). Extra-cardiac factors such as severe anemia, thyroid dysfunction, and sepsis can generate supra-normal cardiac work requirements and, as a result, HF. Although most cases of HF are believed to be due to hypertension and/or coronary artery disease, a large number of patients do not have a clearly identifiable precipitating cause for their syndrome. (See Table 2.)

Table 2. Major Etiologies of Heart Failure

Coronary artery disease
Complications of myocardial infarction
• Acute mitral regurgitation (papillary muscle rupture)
Sustained cardiac arrhythmia
Poorly controlled hypertension
Valvular rupture or disease
Postpartum cardiomyopathy
Acute pulmonary embolus
Pericardial disease/tamponade
• Effusion
• Constrictive pericarditis
Hyperkinetic states
• Anemia
• Thyrotoxicosis
• A-V fistula (e.g., dialysis)
Infiltrative disorders

Myocardial Injury. The sequence of events that leads to the progression of HF from a triggering insult is reasonably well identified. Myocardial injury activates several endogenous biochemical pathways in an initial attempt to maintain circulatory integrity and adequate arterial pressure. After an initial reduction in cardiac output, compensatory hormonal activation acts to preserve circulatory function. The mechanisms of circulatory preservation include sympathetic nervous system (SNS) activation, and increasing baseline levels of norepinephrine (NE), endothelin (ET, one of the most potent vasoconstrictors), and vasopressin. The renin-angiotensin-aldosterone system (RAAS) also is activated, resulting in elevated levels of angiotensin II and aldosterone. While initially protective to ensure tissue perfusion, these compensatory processes ultimately serve as the driving forces behind the pathology of chronic HF. Other substances implicated in the pathophysiology of HF include inflammatory cytokines such as tumor necrosis factor (TNF) and interleukin-6.10,11

Vasodilator peptides also are released by the failing heart. BNP and atrial natriuretic peptide (ANP) are secreted from the ventricles and atria in response to stretch stimulus. These natriuretic peptides promote sodium excretion, decrease vascular resistance, and act to decrease levels of aldosterone. In this regard, the heart functions as an endocrine organ. The natriurectic peptides serve as the endogenous counter-regulatory hormonal system to the RAAS, ET, and SNS. These peptides antagonize the compensatory hormones listed above and form the theoretical basis for newer HF diagnostic and therapeutic regimens.

The RAAS promotes salt and water retention. Angiotensin II also acts to promote peripheral vasoconstriction, resulting in increased afterload stress on the heart. Chronic adrenergic activation results in direct myocardial damage, increased vascular resistance, and increased risk of arrhythmias.9 Chronic elevation of arginine vasopression (AVP), NE, and ET are associated with higher death rates in HF. The combined effects of hormonal activation include sodium and water retention and increased vascular tone. While cardiac output may be maintained, it is at the cost of increased systemic vascular resistance and elevated intra-cardiac pressures. At this stage, the patient may be asymptomatic, but the mechanism already is in place to initiate the secondary pathologic process of cardiac remodeling.

Cardiogenic Shock

Cardiogenic shock (CS) is defined clinically as a state of decreased cardiac output, with evidence of tissue hypoperfusion, despite adequate intravascular volume.12 The etiology of CS usually is acute coronary ischemic disease; other causes are listed in Table 3. The most common cause of CS is acute myocardial infarction (AMI), and in this setting, CS has an overall mortality of 50-90%. In association with AMI, HF occurs if there is acute impairment of at least 25% of the left ventricle. If LV dysfunction exceeds 40%, CS ensues. However, CS may not occur immediately post-MI. In one large study, the median delay from AMI to clinical development of CS was seven hours.13

Table 3. Etiologies of Cardiogenic Shock

Extensive Myocardial Infarction (MI)
MI with Mechanical Complications
• Ventricular septal defect
• Acute mitral regurgitation
• Myocardial free wall rupture
Septic shock with myocardial depression
Pericardial tamponade
LV outflow obstruction
• Hypertrophic obstructive cardiomyopathy (HOCM)
• Aortic stenosis
End-stage cardiomyopathy
Cardiac contusion

Symptoms of CS commonly are the result of progressive myocardial dysfunction associated with impaired myocardial perfusion.12 Because a critical quantity of functioning myocardial tissue is lost in CS, cardiac output is decreased. To compensate for decreased stroke volume, tachycardia develops. The combination of hypotension and tachycardia drastically reduces coronary artery flow by decreasing perfusion pressure and diastolic filling time (the period during which the majority of coronary flow occurs). This may result in further ischemia and myocardial dysfunction. The signs of end organ dysfunction as a result of CS include altered mental status, severe respiratory distress, and decreased urine output. The Forrester classification relates clinical findings to hemodynamic states and mortality. (See Table 4.)

Table 4. Forrester Classification39
Class Description Cardiac Index PCWP Mortality

I No congestion/peripheral hypo perfusion 2.7 12 2
II Isolated congestion 2.3 23 10
III Isolated peripheral hypoperfusion 1.9 12 22
IV Both congestion and peripheral hypoperfusion 1.7 27 55
Key: PCWP = Pulmonary capillary wedge pressure

Diagnosis. Patients with CS can present with pulmonary edema as well as end organ hypoperfusion. Typically, they are congested and vasoconstricted. These patients may present with pulmonary rales and cool extremities, and frequently manifest mental obtundation. Although circulatory shock is diagnosed at the bedside, a cardiogenic source is confirmed by documentation of myocardial dysfunction and persistence of shock despite correction of hypoxemia, hypovolemia, and acidosis.12 Diagnostic criteria for CS are listed in Table 5. The BNP level (BNPL) will be elevated markedly and the chest x-ray (CXR) may show evidence of excess fluid. However, cardiomegaly may be absent in the acute presentation.

Table 5. Cardiogenic Shock Criteria
Systolic blood pressure < 90 mmHg (higher if chronically hypertensive)
Urine output < 0.5 cc/kg/hr
Evidence of end organ dysfunction:
– Renal failure
– Cerebral (confusion)
– Peripheral hypoperfusion (cool extremities)
If hemodynamic monitoring is available:
– PCWP > 18 mmHg and CI < 1.8 L/min/m2
Key: PCWP = Pulmonary capillary wedge pressure; CI = Cardiac index

Acute Pulmonary Edema

Acute pulmonary edema (APE) is characterized by fluid overload in the lungs. It may be cardiogenic in nature (i.e., secondary to increased LV filling pressures) or non-cardiogenic (i.e., fluid overload from renal failure). The development of APE is best understood as a downward spiral of events resulting in progressive myocardial dysfunction that highlights the failing heart’s sensitivity to increased afterload. The spiral starts when an individual with baseline LV dysfunction (diastolic or systolic) experiences an additional myocardial stressor. Filling pressures are increased, the myocardium is unable to compensate, and hence pulmonary congestion and dyspnea result. Catecholamine levels and vascular resistance are increased. This results in increases in blood pressure and afterload, further LV dysfunction, and progressive increases in filling pressures with more pulmonary congestion. A recent report has stressed the importance of diastolic dysfunction in the majority of patients presenting with hypertensive PE.14

Diagnosis. APE presents with clinical evidence of pulmonary congestion (tachypnea, rales, and x-ray changes) in the presence of elevated systemic blood pressure. Occasionally, patients with impending PE may have intense dyspnea in the absence of rales or edema. Patients with PE can progress rapidly to respiratory failure, cardiogenic shock, and cardiovascular collapse if the diagnosis is not recognized promptly and the condition is not treated aggressively. The arterial blood pressure often is elevated and serves as the focus of immediate therapy with vasodilators. Even patients with pronounced systolic dysfunction are capable of having significantly elevated blood pressure in this setting. Peripheral vasoconstriction often is severe, and the extremities are cool. (See Table 6 for the correlation between physical diagnosis and hemodynamic parameters.) As with cardiogenic shock, a search for underlying precipitants should be performed.

Table 6. Physical Diagnosis and Hemodynamic Status
Warm Extremities Dry (clear lungs) Wet (rales)
(good perfusion) Normal SVR Vasodilated (low SVR)
Lungs without rales Lungs congested (rales)
Normal CO Normal CO
Ex: Normal hemodynamics Ex: Septic shock
Cold Extremities
(poor perfusion) Vasoconstricted (high SVR) Vasoconstricted (high SVR)
Lungs without rales Lungs congested (rales)
Low CO Low CO
Ex: Dehydration and HF Ex: Decompensated HF
CO = Cardiac output 
SVR = Systemic vascular resistance 
HF = Heart failure

Decompensated Heart Failure

While a patient suffering from a new anterior infarction presents a clear reason for developing HF, most patients do not have a single, defining insult explaining the development of this syndrome. (See Figure 1.) However, most references report that the majority of HF is due to coronary artery disease (chronic ischemia or infarction) or hypertension.

The clinical spectrum of decompensated HF is broad. It includes asymptomatic patients, mildly symptomatic patients with dyspnea, and progresses to cardiogenic shock and acute pulmonary edema. Frequently, the diagnosis of HF has been established, and the presentation is the culmination of increasing congestive symptoms. HF frequently is decompensated due to an additional event superimposed upon the patient’s pre-existent pathology. Consequently, the physician should attempt to identify possible reasons for acute decompensation. Common precipitants are listed in Table 7. The results of a recent study of 323 patients are listed in Table 8. These patients were part of a clinical trial, and, therefore, the incidence of non-compliance may have been underestimated.

Table 7. Common Causes
of Heart Failure Decompensation

Acute myocardial ischemia or infarction
Uncontrolled hypertension
Superimposed infection
Atrial fibrillation or other arrythmia
Excessive alcohol
Worsening valvular lesion
Endocrine abnormalities (e.g., diabetes, hyperthyroidism, etc.)
Negative inotropic medications (e.g., verapamil, nifedipine, etc.)
Non-steroidal anti-inflammatory drugs
Treatment and dietary non-compliance


Table 8. Precipitants
of Decompensated Heart Failure



Excess salt 22%
Non-cardiac causes 20%
Anti-arrhythmic agents within 48 hours 15%
New arrhythmia 13%
Calcium blocker use 13%
Inappropriate therapy reduction 10%
Rapid atrial fibrillation 10%
Myocardial ischemia 8%
Medication non-compliance 7%
Myocardial infarction 2%
Uncontrolled hypertension 2%

Myocardial ischemia always should be considered as a possible precipitant of decompensated HF. In a retrospective review of 151 decompensated HF patients seen in the ED, 21 (14%) were found to have elevated cardiac enzymes. In these patients, chest pain occurred in only 30%, and the ECG was diagnostic for acute cardiac ischemia in fewer than 1%.15 Although controversial, elevated cardiac enzymes in HF patients are associated with an increased in-hospital mortality rate, poor six-month mortality rates, and a higher cardiac transplantation rate.

The NYHA class severity for individuals with HF represents an historical standard for categorizing the clinical severity of HF. This scale is based on the amount of effort required to produce symptoms of dyspnea and exercise intolerance. (See Table 9.) NYHA class is directly proportional to survival, with higher classes suffering higher morbidity and mortality rates; however, there is no direct relationship between NYHA and EF. For example, some patients are well-compensated with an EF of only 10%, while others with much higher EFs are profoundly symptomatic. In addition, NYHA class suffers from interobserver variability. The newly available BNP assay correlates with NYHA class, and may provide a more objective measure in the future.16,17

Table 9. New York Heart Association Class
Class Limitations Daily Living Symptoms BNPL (pg/mL)

I No limitation Asymptomatic during usual daily activities 100
II Slight limitation Mild symptoms during ordinary daily activities 200
III Moderate limitation Symptoms noted with minimal activities 450
IV Severe limitation Symptoms present even at rest >1000

Recently, the American College of Cardiology and the American Heart Association, in an effort to include patients who are in the asymptomatic initial stages of HF, published a new classification system. This system recognizes the fact that earlier intervention has the potential for greater morbidity and mortality benefits as compared to therapy in the late stages of disease. (See Table 10.)

Table 10. ACC/AHA Heart Failure Classification37
ACC/AHA Class Stage Patient Description NYHA

A High risk for developing left ventricular dysfunction (LVD) Hypertension 
Coronary artery disease
Diabetes mellitus
Family history of cardiomyopathy
B Asymptomatic LVD Previous MI 
LV systolic dysfunction
Asymptomatic valvular disease
C Symptomatic LVD Known structural heart disease
Shortness of breath and fatigue
Reduced exercise tolerance
D Refractory end-stage HF Marked symptoms at rest despite maximal medical therapy IV

Diagnosis. The majority of patients with decompensated HF will have a history of HF and present to the ED without florid PE or systemic hypoperfusion. They may complain of dyspnea (on exertion, nocturnally, or with recumbency), peripheral edema, cough, weight gain, fatigue, abdominal pain, or generalized weakness. These patients suffer from a progression of their previously compensated HF.

Low Cardiac Output

HF traditionally has been considered a congestive state with an inexorable course characterized by increasing symptoms, frequent hospitalizations, and ultimately, death. With the advent of powerful loop diuretics, vasodilator therapy, ACE inhibitors, and other agents, treatment not only forestalls development of symptoms, but prolongs the quality of life and decreases mortality. Consequently, a relatively new class of HF patient may present to the ED. This patient generally is compliant with medications, meticulous in fluid restriction, and has a long-established course of HF therapy. With progressive ventricular dysfunction, these individuals present with a low cardiac output syndrome, in the absence of clinically detectable systemic or pulmonary congestion. Peripheral perfusion may be compromised, and there is severe fatigue, shortness of breath, and limitation in physical activity on the basis of poor cardiac output. As with decompensated HF, a superimposed precipitant added to the patient’s baseline myocardial dysfunction must be considered. In addition, intravascular volume depletion (i.e., secondary to overzealous diuresis) may contribute to clinical deterioration.

Diagnosis. These patients report a history of HF, and already are taking appropriate medications. They may relate a history of overwhelming fatigue, dyspnea on exertion, nausea, malaise, and orthostatic symptoms. These symptoms occur despite meticulous compliance with medications, and in the absence of significant weight gain or congestive findings. They usually are not fluid overloaded, and are generally without pulmonary congestion or extremity edema. These patients will require work-up for a progression of their HF, as well as a search for reversible precipitants of decompensation. These patients represent the "cold and dry" patient in Table 6.

Differential Diagnosis

A number of clinical conditions can mimic the acute presentation of HF. Diagnostic accuracy is required in these patients, as omissions in treatment will prevent an optimal response in HF; moreover, therapy directed against another acute condition could have grave consequences. Because breathlessness is a common presenting symptom, other conditions causing dyspnea should be considered. Acute coronary syndrome always must be excluded as the primary cause of the patient’s complaints, and as the underlying etiology of HF exacerbation. Table 11 summarizes diagnoses that must be considered in the differential diagnosis of HF.

Table 11. Heart Failure Differential Diagnosis
Dyspneic States
• Chronic obstructive pulmonary disease exacerbation
• Asthma exacerbation
• Pulmonary embolus
• Acute myocardial infarction
• Physical deconditioning
• Obesity
• Pleural effusions
• Pneumonia/pulmonary infection
• Pneumothorax
Fluid Retentive States
• Renal failure/nephrotic syndrome
• Liver failure/cirrhosis
• Portal vein thrombosis
• Hypoproteinemia
• Dependent edema
Low Cardiac Output States
• Acute myocardial infarction
• Tension pneumothorax
• Pericardial tamponade
• Sepsis

As might be expected, the diagnosis of HF can be especially problematic in the patient with co-existing obstructive pulmonary disease. As is the case with acute exacerbation of HF, patients with chronic obstructive pulmonary disease (COPD) may present with acute dyspnea and have abnormal lung sounds. In fact, an exacerbation of HF can result in worsening of COPD and vice versa. The history may provide clues to the cause of the exacerbation. Non-compliance with diuretics, or the presence of fever with increased sputum production, should suggest the presence of HF or pneumonia, respectively. A careful physical examination, searching for findings that permit the clinician to distinguish between HF and COPD, should be performed. Severe hypertension and peripheral vasoconstriction suggest acute HF, even in the presence of audible wheezing in a patient with known COPD. A directed physical exam is necessary, although it is limited in sensitivity and specificity. This includes checking for JVD, HJR, the presence of extra heart sounds, and evaluating peripheral perfusion and extremity edema. The chest x-ray can be of assistance, but it does have limitations.

Ventilation perfusion scan, helical thoracic computed tomography (CT), or pulmonary angiography may be required to confirm the diagnosis. Finally, peripheral edema is a common finding in HF, but it is not specific for this condition. It also is found in patients with hypoproteinemic states, hepatic or renal failure, and may be the result of primary venous disease.

Diagnosis. The diagnosis of HF may be difficult when the patient has minimal symptoms and when there is no supporting data from clinical laboratory or echocardiography. With disease progression, the most common ED complaint is a respiratory problem (e.g., dyspnea, orthopnea), although edema, weakness, difficulty sleeping, fatigue, dizziness, cough, or abdominal pain are associated symptoms. When limited to using only history and physical, the diagnosis of HF frequently is in error. In a family practice clinic environment, the diagnosis of HF was correct in only 36% of males, and 18% of females.18 Misdiagnoses were attributed to obesity, deconditioning, or another dyspneic condition. The diagnosis of HF is more accurate in the ED, but this is probably because symptoms are more acute. Among 250 patients presenting to an urgent care environment with HF, physicians misdiagnosed 30 patients. Of these 30 patients, 15 were over-diagnosed and 15 were underdiagnosed as having HF.17

History. In general, a brief history should be elicited from the patient or family with special attention to risk factors, signs, or symptoms of myocardial ischemia, arrhythmia, potential infectious processes and, as appropriate, PE. Once stabilized, the patient and family should be questioned about salt intake, medicine compliance (with attention to financial concerns and side effects), alcohol and drug use, and recent changes in their therapeutic regimen. Clinicians also should evaluate the possibility that iatrogenic, drug-related factors, such as digitalis intoxication, use of nonsteroidal anti-inflammatory drugs, or negative inotropism, may have played a role in exacerbating HF.

Physical Examination. After the ABCs have been addressed, the physical exam should focus initially on murmurs indicative of ventricular outflow tract obstruction, such as those found with aortic stenosis and hypertrophic obstructive cardiomyopathy. These valvular lesions require specialized care and prompt detection inasmuch as administration of nitrates, arterial vasodilators, or diuretics may worsen their clinical status.

Some aspects of the physical exam can narrow the differential diagnosis of HF. The most useful findings for confirming the diagnosis of HF is the presence of elevated JVD or a cardiac third heart sound (S3). All other physical exam parameters have a poorer correlation with HF. (See Table 12.)17

Table 12. Accuracy of Various Heart Failure Parameters17
Variable Sensitivity Specificity Accuracy

History of HF 62 94 90
Dyspnea 56 53 54
Orthopnea 47 88 72
Rales 56 80 70
S3 20 99 66
JVD 39 94 72
Edema 67 68 68

ECG and Laboratory Evaluation of Heart Failure

The ECG is a pivotal diagnostic tool in patients with suspected HF. ECG changes diagnostic of acute MI or acute ischemia mandate admission to the cardiac intensive care unit. An ECG also can assess cardiac rhythm, suggest the need for evaluation of a possible electrolyte abnormality (e.g., hyperkalemia), and may indicate the presence of potential drug toxicity (e.g., junctional bradycardia with digoxin toxicity). It also can provide prognostic information. In dilated cardiomyopathy, the presence of abnormal Q waves, QRS duration greater than 0.12 ms, or left bundle-branch block, predict an increased five-year mortality rate.8 Note that in patients with systolic dysfunction, the resting ECG is relatively insensitive and nonspecific in identifying severe underlying coronary disease (i.e., presence or absence of old Q-waves).

Cardiac Monitoring. A period of cardiac monitoring in the ED is necessary for patients with HF. This may help exclude cardiac arrhythmia as the cause of the current exacerbation. Continued monitoring, either on the inpatient unit or in the observation unit may be of utility when indicated by the clinical presentation (i.e., recent palpitations or near-syncope). Aggressive treatment strategies, such as repetitive diuretic dosing, may result in electrolyte abnormalities whose effect can be pro-arrhythmic.

Laboratory. As emphasized, asymptomatic cardiac ischemia can precipitate acute HF, or it may produce decompensation in a previously stable HF patient. Because of CK-MB kinetics, a positive result may diagnose but not exclude acute ischemia. Similarly, troponin T may provide evidence of cardiac injury as long as 10 days prior to presentation, but cannot exclude acute myocardial damage.15 As compared to AMI in non-HF patients, elevated cardiac ischemia markers in HF may not necessarily represent epicardial coronary artery occlusion. However, cardiac marker elevation is associated with an increased risk of adverse outcome.19-25 Arterial blood gases routinely are not necessary, and can be limited to the cohort at concurrent risk of CO2 retention and in those who are severely ill. Other suggested laboratory tests include serum electrolytes.

Testing for specific drug levels (e.g., digoxin) should be guided by the presentation. Alcohol analysis and drug screening may be considered in selected patients. Magnesium levels should be considered when either arrhythmia or severe hypokalemia is present.

B-type Natriuretic Peptide—An Emerging Tool

Natriuretic peptides (NP) are a series of proteins released as a result of volume stimulus. ANP originates from the cardiac atria and increases in response to atrial distension. BNP is stored in ventricular myocardial granules and is released in response to ventricular overload. Lastly, CNP is an endothelially derived NP. These peptides serve as the hormonal counterpoint to the renin-angiotensin-aldosterone system, which produces vasoconstriction, sodium retention, and SNS activation in patients with HF. NPs result in vasodilation, decreased aldosterone levels, inhibition of the renin-angiotensin-aldosterone system, and decreases in SNS activity, therefore modulating the processes known to be at the core of HF pathophysiology.

BNP is cleared by three mechanisms: 1) binding to cell surface clearance receptors with internalization and proteolysis; 2) proteolytic cleavage by neutral endopeptidases; and 3) renal filtration. The elimination half-life of exogenously administered BNP is approximately 18 minutes.

Measuring the BNP level (BNPL) has the potential to improve diagnostic accuracy with suspected HF in the ED in as much as BNPL correlates well with cardiac function and clinical presentation. In one study, in the absence of clinical HF, BNPL was a mean of 38 pg/mL, compared to patients with HF, in whom the mean BNPL was 1076 pg/mL.17 Accordingly, BNPL is an accurate tool for the diagnosis of HF, and it correlates with NYHA class.16 (See Table 9.)

A BNPL is most useful in patients who present a diagnostic challenge, i.e., those with comorbid conditions that may confound historical and physical findings. In one study, patients whose dyspnea was due to an exacerbation of COPD had a BNPL lower than 100, compared to those with HF-induced dyspnea, in whom the BNPL was in the range of 1000. Another analysis comparing edema in HF and non-HF patients reported similar results.17

Studies also suggest that BNP can discern future outcomes. A BNPL greater than 480 predicts higher six-month death and rehospitalization rate due to HF, compared to those with levels less than 230 pg/mL.26 In addition, the one-year mortality rate is increased markedly if the BNPL exceeds 73, in comparison to those with a lower BNPL.27 Finally, some data suggest ED disposition can be guided by BNPL. Patients with a BNPL greater than 700 were more likely to require inpatient admission, while those treated successfully as outpatients had BNPL less than 254 pg/mL.17 While these findings need to be duplicated by other trials, BNPLs have demonstrated a clear trend for having predictive value in making patient dispositions in patients with HF.26

The diagnosis of HF is suggested by a suggestive clinical presentation and a BNPL greater than 100 pg/mL. At this cutoff point, the BNPL demonstrates both a sensitivity and specificity of 94%. The positive predictive value is 92%, and the negative predictive value is 96%, for an overall accuracy of 94%. Although this can be a useful test for helping ED physicians confirm the presence or absence of congestive HF,17 the BNPL also is elevated in some non-HF conditions. It is increased relatively in the elderly, in women, and in those conditions that would be predicted to increase intracardiac pressures, such as dialysis, cirrhosis, primary pulmonary hypertension, possibly hormone replacement therapy, and PE. Finally, this test has a coefficient of variation of about 10%. In the ranges that BNP occurs clinically in the ED, this is insignificant.

Clinical Utility of BNP Assay. Guidelines for incorporating a BNPL into patient management have not yet been accepted uniformly. However, studies suggest a BNPL is useful for excluding the diagnosis of HF when it would be considered in the differential.17 In these situations, a normal BNPL should prompt consideration of a non-HF diagnosis. Because non-HF conditions can elevate BNPL, the clinical context of a high BNPL must be considered. A positive BNPL should prompt testing to verify the diagnosis and determine the cause of HF (e.g., ECG, CXR, echocardiogram, cardiac stress testing).

Serial BNPLs can be used to monitor adequacy of HF therapy.28 In general, levels correlate well with treatment response. Following treatment for HF, a declining BNPL portends a more favorable outcome, whereas a rising BNPL suggests a greater risk of adverse outcome. A rising BNPL indicates a more aggressive treatment strategy may be warranted. Finally, a BNPL 48 hours post-MI is a strong prognostic indicator for the development of HF or death within one year, and appropriate treatment and monitoring strategies should be considered in this group.29,30

Changes in BNPL correlate dynamically with pulmonary capillary wedge pressure (PCWP). It has been suggested that levels may be utilized to follow acute responses to treatment. Further research is needed, but BNP appears to be a laboratory surrogate for PCWP.31 The BNP test takes 15 minutes to perform and it is available as a Clinical Laboratories Improvement Act (CLIA)-defined moderately complex bedside point of care assay.

Radiographic Modalities in Heart Failure

Chest Radiograph. A PA and lateral CXR should be obtained in all patients with suspected HF. Importantly, a negative radiograph does not exclude the diagnosis of abnormal left ventricular function. However, in the acutely dyspneic patient, it may exclude other significant, confounding diagnoses, among them, pneumothorax, infiltrative processes, pneumonia, and others. CXR findings suggesting HF include cardiomegaly, Kerley’s lines, increased pulmonary vascularity, and pleural effusions. Because radiographic abnormalities may follow the clinical presentation by hours, therapy should not be withheld waiting for the CXR.

It should be stressed that the CXR has poor sensitivity for cardiomegaly,32 and the performance technique significantly can impact the interpretation. Supine radiographs, commonly used following endotracheal intubation, have poor sensitivity (67%) and specificity (70%) for detecting pulmonary fluid.33 Furthermore, in chronic HF, the CXR has unreliable sensitivity, specificity, and predictive value in identifying patients with PCWP greater than 20 mmHg. In addition, both physical and x-ray signs of congestion have poor predictive values for an elevated PCWP in patients with chronic HF. In one study, x-ray evidence of pulmonary congestion was absent in 53% of HF patients with mild to moderately elevated PCWPs (16-29 mmHg) and in 39% of those with markedly elevated PCWP (> 30 mmHg).34

Cardiomegaly is a useful finding for diagnosing HF, and a cardiothoracic ratio of 60% is associated with a higher five-year mortality rate.8 However, the CXR is insensitive for cardiomegaly. In 49 cases of echocardiographically proven cardiomegaly, 22% had cardiothoracic ratios less than 50%.32 Undetectable cardiomegaly by CXR is explained by intrathoracic cardiac rotation. While cardiomegaly is detected accurately by echocardiography, this modality is not available in all EDs.

Pleural effusions are common in HF, but they easily are missed by CXR. This is of concern in intubated patients, since the supine technique further degrades the diagnostic accuracy for the detection of pleural effusions. In one study, 34 patients with pleural effusions proven by decubitus CXRs, the sensitivity, specificity, and accuracy of the supine CXR was reported as 67%, 70%, and 67%, respectively.33

Although a portable CXR is obtained readily in the ED, there are drawbacks to relying upon it for diagnostic purposes. (See Table 13 for the common findings of HF.) In 22 "mild" HF patients, only dilated upper lobe vessels were found in more than 60%. The frequency of findings increased with increasing HF severity. In severe HF, CXR abnormalities occurred in at least 67%, except a prominent vena cava in 44%, and Kerley’s lines in 11%.35 The "gold standard" for documenting intrathoracic fluid is unclear, although CT-scanning has been proposed. While CT may offer improved imaging, it is difficult to perform in unstable patients, and cannot be recommended routinely for ED use in HF.

Table 13. Most Common Portable CXR Findings (Descending Order of Frequency)35

1. Dilated upper lobe vessels
2. Cardiomegaly
3. Interstitial edema
4. Enlarged pulmonary artery
5. Pleural effusion
6. Alveolar edema
7. Prominent superior vena cava
8. Kerley’s lines

Echocardiography. Echocardiography is the "gold standard" test for confirming the presence of HF. It provides excellent information about contractility, chamber size, valve status, and it can evaluate EF. As the single most important measurement in HF,36 quantitative determination of EF can establish the diagnosis and determine the type of HF, but there is no predictable association between symptoms and EF. Once the diagnosis of systolic failure is established, repetitive measurements usually are unnecessary. However, if an echocardiogram more than one year prior showed a normal EF, a repeat study may be considered. While echocardiography can be done by portable technique, it is relatively unavailable in real time for most emergency physicians. Knowledge of a patient’s EF can impact therapy. Since diastolic HF is relatively pre-load dependent, a less aggressive treatment regimen can avoid hypotension resulting from rapid lowering of preload in this population.

Bioimpedance Monitoring. Bioimpedance monitoring is FDA-approved for non-invasive hemodynamic measurement of cardiac function. The placement of four electrodes, similar to ECG electrodes, allows the calculation of cardiac output, systemic vascular resistance (SVR), and total thoracic fluid content. Output and resistance parameters may be normalized by index, similar to data generated from a Swan-Ganz catheter. While maximizing cardiac output has little effect on mortality, improving SVR with vasodilators results in an acute improvement in dyspnea, and chronically decreases mortality. Determining SVR may permit more aggressive use of vasodilator medication in the acute treatment of decompensated HF. Secondly, bioimpedance measurement of thoracic fluid content correlates with the amount of thoracic fluid noted on CXR. This measurement may be useful for rapid evaluation of suspected HF, or when the chest radiograph is equivocal.


1. O’Connell JB, et al. J Heart Lung Transplant. 1994;13:S107-S112.

2. Second Thromboembolic Risk Factors (THRIFT II) Consensus Group. Risk of and prophylaxis for venous thromboembolism in hospital patients. Phlebology 1998:13:87-97.

3. Clagett GP, Andersen FA, Heit JA, et al. Prevention of venous thromboembolism. Chest 1998;114:531S-560S.

4. Nicolaides AN, Bergquist D, Hull R, et al. Consensus statement. Prevention of venous thromboembolism. Int Angiol 1997:16:3-38.

5. Kleber FX, Witt C, Flosbach K, et al. Study to compare the efficacy and safety for the LMWH enoxaparin and standard heparin in the prevention of thromboembolic events in medical patients with cardiopulmonary diseases (PRINCE study). Ann Hematol 1998;76:A93. Abstract P261.

6. Gomberg-Maitland M, Baran DA, Fuster V. Treatment of congestive heart failure: Guidelines for the primary care physician and the heart failure specialist. Arch Intern Med 2001;161:342-352.

7. Elesber AA, Redfield MM. Approach to patients with heart failure and normal ejection fraction. Mayo Clin Proc 2001;76:1047-1052.

8. Ghali JK, Kadakia S, Cooper RS, et al. Bedside diagnosis of preserved versus impaired left ventricular systolic function in heart failure. Am J Cardiol 1991;67:1002-1006.

9. Francis G. Pathophysiology of chronic heart failure. Am J Med 2001;110: 37S-46S.

10. Levine B, Kalman J, Mayer L, et al. Elevated circulating levels of tumor necrosis factor in severe chronic heart failure. N Engl J Med 1990;323:236-241.

11. Young JB. Heart failure is a fever. Clev Clin J Med 2000;67:801-803.

12. Hollenberg SM, Kavinsky CJ, Parrillo JE. Cardiogenic shock. Ann Intern Med 1999;131:47-59.

13. Hochman JS, Boland J, Sleeper LA, et al. Current spectrum of cardiogenic shock and effect of early revascularization on mortality. Circulation 1995;91:873-881.

14. Gandhi SK, Powers JC, Nomeir AM, et al. The pathogenesis of acute pulmonary edema associated with hypertension. N Engl J Med 2001;334:17-22.

15. Peacock WF, Emerman CL, McErlean ES, et al. Troponin T: Insensitive for outcome prediction in low risk suspected acute coronary syndrome ED observation unit patients. Acad EM 1998;5:519.

16. Biosite Diagnostics. Triage BNP Test package insert. San Diego, CA.

17. Dao Q, Krishnaswamy P, Kazanegra R, et al. Utility of B-type natriuretic peptide in the diagnosis of congestive heart failure in an urgent-care setting. J Am Coll Cardiol 2001; 37:379-385.

18. Remes J, Miettinen H, Reunanen A, et al. Validity of clinical diagnosis of heart failure in primary health care. Eur Heart J 1991;12:315-321.

19. Del Carlo CH, O’ Connor CM. Cardiac troponins in congestive heart failure. Am Heart J 1999;138:646-653.

20. Missov E, Mair J. A novel biochemical approach to congestive heart failure: Cardiac troponin T. Am Heart J 1999;138:95-99.

21. Missov E, Calzolari C, Pau B. Circulating cardiac troponin I in severe congestive heart failure. Circulation 1997;96:2953-2958.

22. Setsuta K, Seino Y, Takahashi N, et al. Clinical significance of elevated levels of cardiac troponin T in patients with chronic heart failure. Am J Cardiol 1999;84:608-611.

23. La Vecchia L, Mezzena G, Ometto R, et al. Detectable serum troponin I in patients with heart failure of nonmyocardial ischemic origin. Am J Cardiol 1997;80:88-89.

24. Healey JS, Smith ST, Davies RF, et al. Prognostic value of serum troponin T and troponin I in patients with HF. J Card Fail 2000;6:234.

25. Perna ER, Macia SM, Parras JI, et al. High levels of cardiac troponin T in decompensated HF are associated with poor long term prognosis. J Card Fail 2000;6:232.

26. Maisel A. Personal Communication, Sept. 11, 2001.

27. Tsutamoto T, Wada A, Maeda K, et al. Attenuation of compensation of endogenous cardiac natriuretic peptide system in chronic heart failure: prognostic role of plasma brain natriuretic peptide concentration in patients with chronic symptomatic left ventricular dysfunction. Circulation 1997;96: 509-516.

28. Nagaya N, Nishikimi T, Uematsu M, et al. Plasma brain natriuretic peptide as a prognostic indicator in patients with primary pulmonary hypertension. Circulation 2000;102:865-870.

29. Talwar S, Squire IB, Downie PF, et al. Profile of plasma N-terminal proBNP following acute myocardial infarction. EHJ 2000;21:1514-1521.

30. Bettencourt P, Ferreira AL, Pardal-Oliveira N, et al. Clinical significance of brain natriuretic peptide in patients with postmyocardial infarction. Clin Card 2000;23:921-927.

31. Kazanegra R, Cheng V, Garcia A, et al. A rapid test for B-type natriuretic peptide correlates with falling wedge pressures in patients treated for decompensated heart failure: A pilot study. J Card Fail 2001;7:21-29.

32. Kono T, Suwa M, Hanada H, et al. Clinical significance of normal cardiac silhouette in dilated cardiomyopathy—Evaluation based upon echocardiography and magnetic resonance imaging. Jpn Circ J 1992;56:359-65.

33. Ruskin JA, Gurney JW, Thorsen MK, et al. Detection of pleural effusions on supine chest radiographs. AJR 1987;148:681-683.

34. Chakko S, Woska D, Marinez H, et al. Clinical, radiographic, and hemodynamic correlations in chronic congestive heart failure: Conflicting results may lead to inappropriate care. Am J Med 1991;90:353-359.

35. Chait A, Cohen HE, Meltzer LE, et al. The bedside chest radiograph in the evaluation of incipient heart failure. Radiology 1972;105:563-566.

36. Packer M, Cohn JN. Consensus recommendations for the management of chronic heart failure. Am J Cardiol 1999;83:1A-38A.

37. Hunt SA, Baker DW, Chin MH, et al. American College of Cardiology/ American Heart Association guidelines for the evaluation and management of chronic heart failure in the adult: Executive summary. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to revise the 1995 Guidelines for the Evaluation and Management of Heart Failure). J Am Coll Cardiol 2001;38:2101-2113.

38. Grossman W. Diastolic dysfunction in congestive heart failure. N Engl J Med 1991;325:1557-1564.

39. Forrester JS, Diamond G, Swan HJC. Correlative classification of clinical and hemodynamic function after acute myocardial infarction. Am J Cardiol 1977;39:137.

40. Tsuyuki RT, McKelvie RS, Arnold JM, et al. Acute precipitants of congestive heart failure exacerbations. Arch Intern Med 2001;161:2337-2342.