In recent years, much has been published on the clinical epidemiology of heart failure, including the secular trends in incidence, prevalence, etiology, and prognosis. A clear definition of any condition under study is essential but has proven difficult for heart failure, with no generally accepted gold standard. As a consequence, comparison between studies and countries, over time, is difficult. A full discussion of the diagnostic tests and strategies that can be used in making a diagnosis of heart failure is described in Chap. 28, but an overview of the approaches used in population-based studies is provided here.

Unquestionably, heart failure is an important health care issue. Developed countries spend 1% to 2% of their health care budget on patients with this condition, and heart failure is the single most common diagnostic-related group for admissions to US hospitals. Hospital readmission rates are high, particularly where chronic disease monitoring is poor, and admissions tend to be long, particularly in Europe. Much effort has been expended in improving the standards of care for patients with heart failure, with the publication of international guidelines and the development of chronic disease management programs key to the implementation of these guidelines.

Epidemiologic studies have provided a wealth of information about who is at risk of developing heart failure and how they fare after diagnosis. They also allow predictions to be made about the likely future burden of disease.

Most of the published data have come from studies in developed countries, initially North America, but more recently Europe. Much more limited information is available from the developing world. With epidemiologic transition, it is likely that the epidemiology of heart failure will become increasingly similar across the world, driven by the major etiologic forces of coronary artery disease, hypertension, and diabetes mellitus. Even now, cardiovascular disease is overtaking infectious disease as the leading cause of death worldwide.

Before describing the epidemiology of heart failure, it is important to appreciate the challenge of defining heart failure in a robust but practical manner suitable for population-based studies.

Heart failure is not a complete diagnosis in itself; this requires characterization of the syndrome in terms of its severity, the underlying cardiac abnormality, its etiology, and the manner in which the whole body has adjusted to the pump dysfunction. The advent of high-resolution noninvasive imaging, particularly echocardiography, has helped confirm underlying structural or functional cardiac abnormalities in patients with symptoms and signs suggestive of heart failure. Problems remain in determining what is outside normality, particularly in the aging heart.

The definition of heart failure has evolved over time. There is no currently widely accepted gold standard based on an objective test. Different definitions are used in epidemiologic studies, clinical trials, and clinical practice, making comparisons difficult.1,2 Recent population studies have attempted to apply more clinically based definitions but are labor intensive.

Heart failure has been defined as a “syndrome that develops as a consequence of cardiac disease and is recognized clinically by a constellation of symptoms and signs produced by complex circulatory and neurohormonal responses to cardiac dysfunction.“3 Although accurate, this definition does not lend itself to use in an epidemiologic study.

The European Society of Cardiology guideline on the diagnosis and treatment of heart failure suggests that symptoms should be present, either at rest or on exercise, with objective evidence of cardiac dysfunction provided preferably by echocardiography. Where an element of doubt persists, response to therapy directed toward heart failure can help confirm or refute the diagnosis.4 The most recent guidelines from the Heart Failure Society of America advocate a similar approach, with a careful history supplemented by physical examination and tests to assess cardiac structure and function.5 In both guidelines, the measurement of the plasma concentration of B-type natriuretic peptide is recommended to help make or rule out a diagnosis in patients when uncertainty remains.4,5 Such a clinical approach to diagnosis is time consuming but has recently been adopted in population-based studies, tending to replace the older approach of using scoring systems, such as the Framingham Heart Study score, which is discussed later in this chapter.

Signs and Symptoms of Heart Failure

Heart failure can manifest as a wide range of symptoms and signs. Few are specific for heart failure, and the sensitivity can be low and reduced further by cardioactive medication. Although orthopnea and paroxysmal nocturnal dyspnea are relatively specific for heart failure, they are not sensitive for a diagnosis of heart failure. Many persons with the syndrome will not demonstrate these features in the history. Similarly, a raised jugular venous pressure is highly specific but is insensitive and requires clinical expertise for reliable detection.

Clinical examination can help determine the underlying cardiac abnormality, such as significant valve disease, but cannot be used reliably to distinguish systolic from isolated diastolic abnormalities (Table 27–1).6

Patients admitted to the hospital with heart failure, understandably, have severe symptoms, and breathlessness predominates. Within the Acute Decompensated Heart Failure National Registry (ADHERE) of patients admitted to hospitals in the United States with acute heart failure, 34% presented with dyspnea at rest, 68% had lung rales, and 66% had signs of peripheral edema.7 Those presenting in the community are likely to have less severe symptoms and fewer clinical signs.

Acute or Chronic Heart Failure?

Heart failure is commonly referred to as either acute or chronic, the former either being acute de novo heart failure or acute decompensation of chronic heart failure. Acute heart failure constitutes 5% of adult emergency medical admissions in the United Kingdom.8 Two surveys of such acute admissions have been devised: the ongoing ADHERE program in the United States7 and a series of snapshot surveys across Europe, under the auspices of the European Society of Cardiology.9 The characteristics of the patients identified in the two geographic areas are remarkably similar (Table 27–2).

Systolic or Diastolic Heart Failure

The underlying abnormality of cardiac function in an individual with heart failure can also be used to describe more fully the nature of the heart failure. Most trials of drug therapy have enrolled patients with obvious underlying systolic dysfunction of the left ventricle, often termed systolic heart failure. A cut point is usually chosen for ejection fraction (EF), below which the heart is said to have systolic dysfunction. Typically an EF of 35% or 40% (large heart) has been accepted. In some patients, however, heart failure occurs in the presence of good left ventricular systolic function. This has been termed heart failure with preserved systolic function or diastolic heart failure but is a heterogenous group including valve dysfunction, arrhythmia, pericardial disease, right ventricular dysfunction, or a stiff and noncompliant left ventricle. The latter abnormality is usually termed isolated diastolic heart failure but is used in different senses by different authors. Strictly speaking, the most rigorous definition would demand convincing evidence of good systolic function of both ventricles at rest and on exercise, no arrhythmia or valve disease, and a proven abnormality of diastolic function on pressure-volume loop calculation at the time of cardiac catheterization. Rarely is this strict definition used, mainly for pragmatic reasons, and the label of diastolic heart failure is applied to a patient with a clinical diagnosis of heart failure and normal systolic function (or at least a near normal calculated EF or a small- or normal-sized heart) on echocardiography. It is thus a diagnosis of exclusion, and such a nonspecific approach can lead to many patients being given a diagnosis of heart failure incorrectly. Much controversy has arisen around the best noninvasive methods of identifying heart failure caused by isolated diastolic dysfunction, with European guidelines suggesting that to make this diagnosis, there should be signs or symptoms of heart failure, normal (or only mildly abnormal) left ventricular systolic function, and evidence of abnormal left ventricular relaxation, filling, diastolic distensibility, or diastolic stiffness.10 Such a definition has proven difficult to use in clinical practice.

Vasan and colleagues have shied away from this strict definition and have suggested dividing the diagnosis of diastolic heart failure by the likely probability of the diagnosis being correct (definite, probable, or possible). Individuals are categorized depending on the presence or absence of symptoms and signs of heart failure, evidence of normal systolic function (EF >50%) during a heart failure event, and evidence of diastolic dysfunction chiefly from echocardiography.11 Such an approach has been endorsed by the current American College of Cardiology/American Heart Association (ACC/AHA) Task Force on Practice Guidelines,12 with the pragmatic suggestion that the term heart failure and normal left ventricular ejection fraction would be a more accurate description in many cases.

Heart failure with preserved systolic function was not considered to be common in many population-based studies, but recent reports from Olmsted County (Rochester Epidemiology Project) and a small nested case-control study from the Framingham Heart Study in North America suggest this can be as common as systolic dysfunction, particularly if the diagnosis of heart failure is based on accepting the clinician’s opinion or using the Framingham criteria and relying on echocardiography at any point during a hospitalization period.13-17 Most of the studies suggest that the probability of the systolic function of the left ventricle being preserved in a patient with heart failure is higher in the elderly, women, and the obese.17,18 In Europe, half of patients admitted to hospital with heart failure have a preserved EF (≥40%) at the time of measurement.9 It has been suggested that misdiagnosis of heart failure is common in the group of patients labeled as having preserved systolic function heart failure.18,19 Data from case-by-case expert review in population-based studies in the United Kingdom conducted by the author suggest that preserved systolic function is found in a lower proportion of new cases of heart failure—approximately 10% to 15%.20,21

Chest Radiograph

The main role for a chest radiograph is to exclude other causes for dyspnea—such as pleural effusion, pneumothorax, lung carcinoma, or pneumonia. Pulmonary edema supports a diagnosis of heart failure, although the reliability of identifying upper lobe venous blood diversion is poor. Cardiothoracic ratio is only of moderate value in identifying heart failure as the cause of breathlessness.22 Echocardiography has replaced chest radiography as the method of determining cardiac chamber dimensions.

Electrocardiogram

In clinical practice, the electrocardiogram (ECG) is used to detect arrhythmia and can provide evidence suggestive of previous myocardial infarction or ventricular hypertrophy. Many studies have suggested that a completely normal ECG is unlikely in a person with heart failure, but its positive predictive value is low in the elderly where ECG abnormality is common.23

Plasma B-Type Natriuretic Peptide

B-type natriuretic peptide (BNP) is secreted by the heart, and the plasma concentration is elevated in left ventricular hypertrophy or dysfunction (systolic or diastolic) and particularly in those with heart failure. Several studies have confirmed its value as a rule-out test for heart failure in patients presenting with new symptoms in either primary or secondary care settings.22-24 Raised plasma concentration can occur in other conditions, such as acute myocardial infarction, pulmonary embolism, and renal failure; and normal values are higher in the elderly and women.25 The current European and North American heart failure guidelines4,5,12 suggest that the measurement of the plasma concentration of BNPs can be useful in confirming or refuting a diagnosis of heart failure, particularly at the time of first presentation in the acute setting. A recent individual patient data meta-analysis of diagnostic studies confirmed the utility of plasma BNP concentration in the workup of patients with suspected heart failure, with greater diagnostic value than the ECG.26

Echocardiography

Transthoracic echocardiography is a simple, safe, and effective method for assessing cardiac structure and function. It is the main imaging method used in cardiology and lends itself to population-based studies. Interpretation of normality can be difficult, particularly for those with poor images because of obesity or chronic airway disease, but satisfactory images should be obtainable in 80% to 90% of free-living subjects.27

Magnetic Resonance Imaging

Magnetic resonance imaging provides high-resolution images of cardiac structure and ventricular function. Contrast agents such as gadolinium can provide information on inflammation, fibrosis, and myocardial perfusion. Valve function can also be assessed, although with less reliability than myocardial structure and function. Although becoming mainstream in large hospitals, the expense of the equipment and its lack of portability make magnetic resonance imaging of limited use in population-based studies.

Cardiac Catheterization

Cardiac catheterization allows measurement of intracardiac pressures, estimation of cardiac output, detection of valve abnormalities, quantification of left ventricular EF, and detection of epicardial coronary artery disease. Diastolic function can be assessed in detail. Its use in epidemiologic studies is limited by the invasive nature of the procedures and the consequent risk to study participants. Cardiac catheterization has been used in a population-based study of incident (new) heart failure in England, where it demonstrated that the presence and importance of coronary artery disease are underestimated by noninvasive assessment methods.21

Epidemiologic Methods

The modern clinical approach to the diagnosis of heart failure contrasts with the methods used in many early epidemiologic studies. The first large study to address heart failure was the Framingham Heart Study. A cohort of a little more than 5000 individuals was examined every 2 years from the study inception in 1948. Heart failure was considered present if on examination there were two “major“ or one “major“ and two “minor“ criteria fulfilled28 (Table 27–3). Between 1949 and 1988, 652 new cases of heart failure were identified—giving an incidence of 2.3 cases per 1000 per year in men and 1.4 cases per 1000 per year in women and indicating a steep increase with age in both sexes.29 The Framingham Score has been used in several other studies in North America, including the Rochester Epidemiology Project.16

Although these criteria are not used in clinical practice to diagnose heart failure, the advantage of the Framingham Heart Study is that the same criteria for heart failure have been applied for many decades, making possible conclusions on secular trends in incidence and prognosis.

Other scoring-based systems have been devised.30 These have not been used widely, and more recent studies in Europe have adopted a more clinical approach, with review of symptoms, signs, and results of chest radiograph and echocardiography by a consensus expert panel. This method appears reproducible20,21 and has been used in several studies in England and the Netherlands.31