CHAPTER 69: THE EPIDEMIOLOGY OF HEART FAILURE

Heart failure (HF) is a clinical syndrome caused by abnormalities of cardiac structure or function that result in impaired cardiac filling, ejection of blood, or both and, consequently, an inability to meet the circulatory needs of the body. There is substantial variation in its etiology, associated comorbidities, prognosis, and response to therapy. HF carries a significant burden on society, accounting for approximately 1 million hospitalizations and 65,000 deaths at a cost of $31 billion annually in the United States alone. In this chapter, we review the epidemiology of HF, with a specific focus on chronic HF.

Commonly encountered symptoms and signs of HF, such as dyspnea and edema, have been recognized for more than 2,500 years, dating back to ancient Greece and Rome.¹ However, pathophysiologic understanding of HF was not well understood. Two millennia lapsed before William Harvey, in 1628, identified the heart as the pulsatile driver of blood through the vascular circulatory system.² This discovery set the stage for understanding hemodynamic abnormalities in HF. In 1918, Starling described the relationship between cardiac volumes and output in the normal and failing heart.³ The advent of cardiac catheterization in the 1940s brought hemodynamic assessments into clinical medicine, but even then therapies for HF were limited to palliation of symptoms with digitalis and the toxic mercurial diuretics.⁴

Around World War II, cardiovascular disease was recognized as the leading cause of death in the United States and, in 1944, Franklin D. Roosevelt was diagnosed with hypertensive heart disease and cardiac failure.⁵ In the context of increasing awareness of the public health burden of cardiovascular disease, the National Heart Act was signed in 1948, establishing what is now known as the National Heart, Lung, and Blood Institute and initiating a new epidemiologic study (the Framingham Heart Study).⁵ In 1953, the cardiologist Inge Edler and physicist Hellmuth Hertz (Lund University, Sweden) developed echocardiography, which enabled a noninvasive estimation of cardiac structure and function.⁶ In the late 1950s, thiazide diuretics were introduced to alleviate congestive symptoms in HF. The 1950s and 1960s also marked a time when knowledge about the cellular and molecular basis of HF was increasing, with seminal observations implicating inflammation, adrenergic and neurohormonal activation, abnormal energetics, myocyte hypertrophy, and altered excitation-contraction coupling.^7,8,9,10 A new era in HF therapy was also ushered in by the first heart transplant in 1967.¹¹

A major barrier to understanding the risk factors and natural history of HF was a lack of uniform diagnostic criteria. Therefore, investigators at the Framingham Heart Study developed a set of clinical criteria for the diagnosis of HF that was first reported in 1971 (Fig. 69–1).¹² Other algorithms for the diagnosis of HF, such as the Boston and Gothenburg criteria, were developed in the mid-to-late 1980s (see Fig. 69–1).^13,14 In each of these algorithms, specificity for the diagnosis of HF is enhanced by combining several diagnostic criteria.

Following the introduction of standardized criteria for HF, clinical research into understanding its risk factors, etiologies, prognosis, and therapies accelerated. In the Framingham Heart Study, hypertension and coronary artery disease were identified to be the major risk factors for HF, and the poor survival among patients with HF was demonstrated.¹² In the 1980s, large-scale randomized clinical trials began not only for treatment, but also prevention of symptomatic HF.^15,16,17 In 1990, a genetic cause for dilated cardiomyopathy was first described.¹⁸ The first clinical guidelines on HF management were published in 1995.¹⁹ By the mid-1990s, recognition that HF could occur in the setting of a preserved left ventricular ejection fraction (LVEF) was gaining traction.^20,21 The efficacy of mechanical (ventricular assist) and electrical (implantable cardioverter-defibrillators and cardiac resynchronization) devices were demonstrated in the early-to-mid 2000s. Nonetheless, despite the remarkable advances in the pathophysiologic understanding and treatment of HF over the past half century, it remains a leading cause of morbidity and mortality.

Criteria for the diagnosis of HF have not substantially changed in the four to five decades since the Framingham, Boston, and Gothenburg criteria were developed. In 2002, circulating natriuretic peptide levels were demonstrated to be robust markers for the differentiation of cardiac versus noncardiac causes of dyspnea, particularly in the acute setting (such as the emergency department), that may aid in the diagnosis of HF.^22,23 However, to date, neither the American Heart Association (AHA)/American College of Cardiology (ACC), nor European Society of Cardiology (ESC) guidelines have incorporated natriuretic peptide levels into the diagnostic criteria for HF. The current ESC guidelines do, however, require objective evidence of cardiac dysfunction (typically obtained noninvasively from echocardiography) in the setting of symptoms and signs to diagnose HF.²⁴

Few studies have compared the performance and agreement of the various criteria for HF.^25,26,27 Although no accepted gold standard for the diagnosis of HF currently exists, the Framingham Heart Study criteria are among the most commonly used criteria for the evaluation of HF in both population-based studies and clinical practice.^12,28,29,30 Population-based cohort studies and/or clinical trials have used the different diagnostic criteria described above to variable extents (some with expert physician panel review, others with automated algorithms or use of administrative claims data, eg, International Classification of Diseases Ninth Revision codes), thereby limiting comparability of the absolute rates for incident and prevalent HF across studies. Geographic variation in the use of diagnostic criteria for HF also has implications on inclusion/exclusion in clinical trials.^31,32 Other factors that influence the estimates of the incidence, prevalence, and prognosis in HF include use of inpatient hospitalized events, outpatient diagnoses, or both; variation in statistical methodology; and ability to apply consistent diagnostic criteria over time. Nonetheless, despite the lack of uniform criteria for defining HF, prior studies of the epidemiology of HF still yield reasonably consistent patterns of disease occurrence and associations with risk factors.

HF is typically a progressive syndrome resulting from risk factors (stage A) that lead to asymptomatic abnormalities of cardiac structure and function (stage B), then eventually transitioning to symptomatic HF (stage C) and decompensations/refractory symptoms (stage D) (Fig. 69–2).³³ This staging system formulated by the AHA and the ACC highlights the potential importance of risk factor modification through lifestyle and pharmacologic means to prevent progression to symptomatic HF. In 2007, the prevalence of each HF stage was estimated in a population-based sample of 2029 residents of Olmsted County, Minnesota, over the age of 45 years, all of whom underwent standardized testing with symptom questionnaires, electrocardiogram (ECG), and echocardiography. In this study, 32% of individuals were classified as stage 0 (no risk factors and normal ventricular structure and function), 22% stage A, 34% stage B, 12% stage C, and 0.2% stage D.³⁴ It should be recognized that estimates of the prevalence of each stage will vary according to the definitions for, and the methods used to detect, abnormalities of cardiac structure and function (stage B) and the assessment of symptoms (stages C and D). In the Olmsted County study, stage B was defined as asymptomatic individuals with one or more of the following: a prior history of myocardial infarction, regional wall motion abnormalities, LVEF < 50%, left ventricular hypertrophy by ECG or echocardiography, left ventricular enlargement, diastolic dysfunction, or moderate or severe valvular heart disease.

For epidemiologic studies and clinical practice, HF can be classified according to other characteristics as well, including the degree of functional limitation, the LVEF, and whether the left, the right, or both ventricles are failing. The New York Heart Association (NYHA) classification is used to characterize functional capacity in patients with HF, ranging from no limitation (class I) to slight limitation during ordinary activities (class II), moderate limitation as a result of symptoms with less than ordinary activity (class III), or severe limitation with symptoms at rest (class IV).³⁵ For patients with advanced HF (ie, NYHA class III and IV with persistent limiting symptoms despite optimal medical therapy), the Interagency Registry of Mechanically Assisted Circulatory Support (INTERMACS) profile further refines classification of disease severity classification to inform risk assessment and the choice and the timing of mechanical circulatory support (Table 69–1).³⁶ For example, a HF patient with NYHA class III limitations may be further categorized as INTERMACS 5, 6, or 7, whereas a NYHA class IV patient could be further classified into INTERMACS categories 1, 2, 3, or 4.

The most common categorization based on cardiac structure and function, and one of substantial clinical importance, is whether the individual patient with HF has a preserved or a reduced LVEF. Heart failure with reduced ejection fraction (HFrEF) and HF with preserved ejection fraction (HFpEF) were previously referred to as systolic and diastolic HF, respectively. The terms systolic and diastolic HF were attempts to describe what was thought to be the primary underlying causal abnormality, namely, reduced pump function versus impaired ventricular filling, respectively. However, the terms have fallen out of favor because it is now well established that abnormalities of systolic function are present in HF patients even within the range of “normal” LVEF.³⁷ Similarly, patients with HFrEF have both systolic and diastolic dysfunction.

Recognition of HFpEF did not appear in the literature until the mid-1980s. In a study of 188 patients with symptomatic HF, approximately one in three had a “normal” LVEF (≥ 45%) quantified by radionuclide ventriculography.²⁰ Abnormal left ventricular filling, or diastolic dysfunction, was present in all 188 patients regardless of LVEF. At that time, LVEF was equated with systolic function; therefore, among the patients with normal LVEF, HF was attributed to diastolic dysfunction and labeled diastolic HF.

Transthoracic echocardiography is the preferred imaging modality for quantifying LVEF in individuals with HF because of its wide availability, lack of radiation, and ease of use.³³ According to the American Society of Echocardiography guidelines, the normal range of LVEF is 52% to 72% in men (average 62%) and 54% to 74% in women (average 64%), which is a slight change from prior guidelines in which the lower limit of normal LVEF was 55% for both men and women.^38,39 The threshold for defining preserved LVEF is controversial. The AHA, ACC, and ESC recommend differentiating a preserved from a reduced LVEF at a cut point of 50%, which is below the lower limit of the normal ranges in men and women.^24,33 Thus, preserved LVEF is not synonymous with normal LVEF. Clinical trials and community-based studies have used a variety of thresholds for defining preserved or reduced LVEF.^33,40,41 Clinical trials of HF tend to use lower LVEF values (≥ 40% or ≥ 45%) for defining preserved LVEF to include a broader range of eligible patients and to cover the LVEF spectrum for which medical therapies have not been demonstrated to improve morbidity and mortality, as described later in the chapter.⁴² Slightly higher LVEF thresholds (≥ 45%, ≥ 50%, or ≥ 55%) have typically been used in community-based registries.^40,41 There is also a consideration of individuals with HF with LVEF of 40% to 50% as a distinctive entity, although this view is not universally accepted.⁴³

The clinical importance of the distinction between reduced versus preserved LVEF relates to prognosis and therapy. This is addressed later in the chapter, but briefly, no specific therapy to date has been demonstrated in randomized clinical trials to be efficacious in reducing HF hospitalizations or risk of death among patients with HFpEF. This is in stark contrast to HFrEF, in which a number of pharmacologic and device-based therapies significantly improve morbidity and mortality. It is also noteworthy that a subset of patients with HFrEF may improve their LVEF over time, referred to as HF with recovered ejection fraction, an observation that suggests a possible dynamic nature of the condition instead of its classification into two distinctive subsets based on LVEF assessed at a single time point.^44,45

Another classification system for cardiac dysfunction is based on whether the left, right, or both sides of the heart are failing. Heart failure is most commonly a result of left ventricular dysfunction, but right ventricular dysfunction is not uncommon among patients with HF.^46,47 Pulmonary venous hypertension secondary to left HF is the most common cause for right HF. Cor pulmonale, or isolated right-sided HF, typically occurs as a consequence of pulmonary arterial hypertension (which can be of various etiologies). Because cor pulmonale is a distinct entity, and substantially less common than left-sided or biventricular HF, it will not be discussed further in this chapter (see Chap. 76).

Incidence

In absolute numbers, incident HF afflicts a large number of individuals. Based on data from 2005 to 2012, investigators from the Atherosclerosis Risk in Communities (ARIC) study estimated that approximately 915,000 individuals in the United States are newly diagnosed with HF annually.⁴⁸ Estimates of HF incidence rates vary markedly, from 1.0 to 32 cases per 1000 person-years (Table 69–2).⁴⁹ Factors contributing to this variation include the population studied and the diagnostic criteria used.⁴⁹ Incidence rates in northern and central Europe appear to be similar to that of the United States, whereas reliable data from the rest of the world are lacking.⁴⁹ In general, incidence rates increase with age and are higher in men than in women and in blacks compared to white individuals.⁴⁹ Older individuals consistently have the highest incidence rates, with an approximate two- to threefold higher incidence for each decade after age 55 years (Table 69–3).^48,50,51,52

Studies of recent trends demonstrate a rise followed by a fall in the incidence of HF in the United States, Canada, and Scotland.^{53,54,55,56,57,58} Advances in the care of coronary artery disease, hypertension, valvular heart disease, arrhythmias, and congenital heart disease over the past half a century have substantially improved cardiovascular mortality.⁵⁹ The consequence is a larger population of individuals at risk for the development of HF. Accordingly, an increasing incidence of HF was observed from the 1970s to 1990s among elderly individuals and in patients following myocardial infarction.⁵⁴ However, since the late 1990s, HF incidence rates have declined in the United States, suggesting that primary prevention approaches for HF, including better recognition and treatment of hypertension and acute myocardial infarction, may be effective. Among Medicare beneficiaries, the incidence rate for HF following an acute myocardial infarction significantly declined by approximately 15%, from 16.1 to 14.2 per 100 person-years, in 1998 to 2010.^55,60 Over a similar time period, from 2000 to 2010, the age- and sex- adjusted HF incidence rate in Olmsted County, Minnesota, declined nearly 40%, from 3.2 to 2.2 per 1000 person-years.⁵⁶ Temporal trends in HF incidence from Ontario, Canada, Scotland, Denmark, and Sweden are congruent with the decline seen in the United States that began in the late 1990s to early 2000s.^57,58,61,62 Whether this trend applies to other parts of the world is largely unknown. Higher incidence rates for HF might be expected in developing countries as a result of “Westernization,” with increasing contributions of modifiable risk factors, such as obesity, diabetes mellitus, and systemic hypertension.⁶³

The proportion of incident HF cases with preserved compared with reduced LVEF appears to be increasing over time.^64,65 Depending on the cutoff value for defining preserved LVEF, the proportion of individuals with HFpEF ranges from 44% to 72%.^41,66 The increasing contribution of HFpEF may be related to several factors. First, there is a greater awareness that HF can occur in individuals with a “normal” ejection fraction, which may increase the frequency with which HFpEF is diagnosed. Second, improved care for patients with acute and chronic ischemic heart disease has decreased the prevalence of reduced LVEF. Third, there is a rising burden of obesity, diabetes mellitus, and physical inactivity, compounded by the aging of the population, all risk factors for HFpEF.^30,67

Prevalence

Approximately 26 million individuals worldwide (5.7 million in the United States) have HF.⁴⁸ In developed countries, the prevalence is approximately 1% to 2% of the overall population.⁴⁹ However, the distribution of individuals with prevalent HF substantially differs by age, affecting approximately 1% of individuals < 65 years old and approximately 10% of individuals ≥ 65 years old.⁴⁹ In the United States, the prevalence of HF is similar between men and women, at 2.3% and 2.2%, respectively, but is more common in black (~3.0%) compared with white (~2.2%) individuals.⁴⁸ Forecasts in the United States estimate a 25% to 46% increase in the prevalence of HF by 2030, with 8 million adults afflicted.^48,68 In the setting of declining incidence rates for HF, the predicted increase in prevalence reflects an aging population as well as anticipated reductions in mortality following the onset of HF.⁴⁹

Cost

The global economic cost of HF is high. It was estimated that 108 billion US dollars were spent in 2012 on HF worldwide, with approximately 60% coming from direct costs (hospital care, medications, physician costs, outpatient care, and follow-up) and 40% from indirect costs (lost productivity, illness benefits, and social support programs).⁶⁹ Higher income countries spend a greater amount on direct costs, whereas low- to middle-income countries spend more on indirect costs. In the United States, annual spending on HF ranges from $30 to $40 billion, with direct costs accounting for nearly 70%.^48,70 Estimates suggest that HF–related costs will double in the United States by 2030.⁶⁸

Hospitalizations

Hospitalizations are a major driver of costs. Approximately 1 million primary hospitalizations for HF occur within the United States each year, with a similar number reported in Europe.⁴⁸ Data from Olmsted County, Minnesota, indicate that over a mean follow-up of 4.7 years 83% of HF patients are hospitalized at least once and 43% are hospitalized four or more times (Fig. 69–3).⁷¹ HF remains the leading cause for hospitalization among Medicare beneficiaries, although the primary hospitalization rate for HF has declined from 1998 to 2008.⁷² This may be, in part, a result of increased intensity of outpatient HF management. However, it is also worth noting that hospitalizations of patients with chronic HF as a secondary diagnosis are stable to increasing, with more than half of all hospitalizations being attributable to noncardiovascular causes, reflecting the rising prevalence of HF and the impact of multiple comorbidities in these patients.^{71,73,74,75,76}

Etiology

A number of etiologies for HF have been defined (Table 69–4). Coronary artery disease is often cited as the most common cause of HF.⁷⁷ Recent data from Norway indicate that 15% to 25% of patients with an acute myocardial infarction develop HF over a 3-year period, with the highest risk in the first 6 months after infarction.⁷⁸ However, there is race- and sex-related variation in the etiologies for HF. For example, in the Health, Aging, and Body Composition study, among white individuals, the population-attributable risks for incident HF were 23.9% and 21.3% for coronary artery disease and hypertension, respectively. Among black individuals, hypertension accounted for 30.1% of incident HF, whereas coronary artery disease accounted for 29.5%.⁷⁹ In the Framingham Heart Study, hypertension, rather than coronary artery disease, was the most common risk factor for HF in both men and women, accounting for approximately 60% of HF cases in women and approximately 40% in men.⁸⁰ Findings from Olmsted County, Minnesota, support sex-related differences in HF risk factors, with coronary artery disease being a major contributor for HF risk in men, whereas hypertension was a greater risk factor in women.⁸¹

The relative contributions of different etiologies for HF also vary geographically. The International Congestive HF Study included nearly 6000 individuals with prevalent HF from South America, Africa, the Middle East, and Asia (Table 69–5).⁸² In South America, coronary artery disease and hypertension were the etiologies for HF in 26% and 21% of cases, respectively. Chagas disease is also an important cause for HF in South America.⁸³ In contrast, in Africa, hypertension was a major etiologic factor for HF in 35% of cases, whereas coronary artery disease accounted for 20%.⁸² In Asia and the Middle East, half of HF cases were attributed to coronary artery disease, whereas hypertension caused 10% to 15% of cases. Rheumatic valvular disease was the etiology for HF in 6% to 9% of cases in Asia and the Middle East, accounting for a larger proportion than seen in North America and Europe.⁸²

Although coronary artery disease and hypertension may be the leading causes of HF worldwide, individual patients frequently have multiple comorbidities that contribute to the syndrome of HF.^67,84 Cardiometabolic disorders such as obesity and diabetes mellitus are increasing in prevalence, with greater contributions to HF incidence.^67,84 Although modifiable cardiometabolic risk factors account for the majority of causes of HF, less common causes include dilated cardiomyopathy, hypertrophic cardiomyopathy, hemochromatosis, amyloidosis, and human immunodeficiency virus (HIV) infection, among others (see Table 69–4). An increasingly recognized cause of HF is cardiotoxic medications, such as some cancer chemotherapies (eg, anthracyclines and novel biologics).⁸⁵

Risk Factors

HF appears to cluster within families. For instance, parental HF has been associated with greater left ventricular mass, internal dimension, systolic dysfunction, and incident HF in offspring.⁸⁶ Nonetheless, no robust genetic variant has, as yet, been identified for “common” HF, or that caused by typical risk factors such as hypertension, diabetes mellitus, and coronary artery disease.⁸⁷ Some of the familial risk may be mediated through modifiable risk factors, particularly those present in middle age, which contribute substantially to the development of HF later in life. Modifiable risk factors include blood pressure, body mass, cholesterol, diet, glucose, physical activity, and smoking. The greatest cardiovascular impact of blood pressure lowering in hypertensive individuals is on the reduction in risk of HF, even more so than stroke or atherosclerotic coronary artery disease.⁸⁸ Higher body mass index and obesity are associated with increased risk of HF, and weight loss reduces this risk.^89,90,91 Alcohol, when chronically consumed in large amounts, is a recognized etiology for cardiomyopathy. However, recent observations from the ARIC study suggest that light-to-moderate alcohol consumption (seven or fewer drinks per week) lowered the risk of HF in men and, to a lesser extent, in women.⁹² Similar findings were reported from the Norwegian Nord-Trøndelag Health Study.⁹³ Hyperglycemia and insulin resistance, even in the absence of overt diabetes mellitus, have been associated with increased risk for incident HF.^94,95 Diabetes mellitus is also an established risk factor for incident HF.^96,97,98,99 However, glucose-lowering therapies for diabetes mellitus have demonstrated mixed results regarding the risk of HF. A number of medications for the treatment of diabetes mellitus have been associated with increased risk of HF, such as the thiazolidinediones (rosiglitazone and pioglitazone) and the dipeptidyl peptidase-4 inhibitor (saxagliptin).^100,101 In contrast, glucose lowering with the sodium-glucose cotransporter 2 inhibitor empagliflozin was recently demonstrated to reduce the risk of HF among individuals with prevalent diabetes mellitus.¹⁰² Higher amounts of physical activity have been associated with reduced risk of incident HF.^103,104 Further, among middle-aged to elderly individuals, smoking is an established risk factor for incident HF.^97,98

The AHA recognized the importance of modifiable risk factors to cardiovascular risk and developed a concept known as the Simple 7 (blood pressure, body mass, cholesterol, diet, glucose, physical activity, and smoking). In the ARIC study, individuals with high Simple 7 scores (indicating more healthy behaviors) were at lower lifetime risk for incident HF (14%), compared with individuals with average (27%) or poor (49%) scores.¹⁰⁵ These findings were corroborated by recent observations from the Framingham Heart Study.¹⁰⁶ Data from the Physicians’ Health Study suggest that optimal modifiable risk factor profiles can lower the lifetime risk of HF by as much as 50%.¹⁰⁷

Modifiable risk factors and comorbidities contribute to the risk of HF, regardless of LVEF. However, the relative risk associated with risk factors for HF with a preserved versus a reduced LVEF may differ. In the Framingham Heart Study, a history of hypertension, diastolic dysfunction, atrial fibrillation, female sex, and lower ratio of forced expiratory volume in 1 second to forced vital capacity were associated with greater risk for symptomatic HFpEF compared with HFrEF.^76,108 Patients with HFpEF often demonstrate exaggerated blood pressure responses to exercise, increased vascular stiffness, chronotropic incompetence, and increased baroreceptor sensitivity, and are exquisitely sensitive to excess preload reduction. In this context, atrial fibrillation appears to be a particularly strong risk factor for HFpEF.¹⁰⁹ In the Framingham Heart Study, prior myocardial infarction, left bundle branch block, asymptomatic left ventricular systolic dysfunction, higher serum creatinine, and lower total blood hemoglobin concentration were significantly associated with greater risk for HFrEF.^76,108

Asymptomatic abnormalities of cardiac structure and function are associated with increased risk for the development of HF. Several population-based studies from the United States, Europe, and Australia have examined the prevalence of asymptomatic left ventricular systolic dysfunction, typically defined by LVEF, with estimates ranging from 1% to 12.9% (pooled 4.8%) (Fig. 69–4).^{110,111,112,113,114,115,116,117,118,119,120} The risk of incident HF associated with asymptomatic left ventricular systolic dysfunction compared with preserved left ventricular function has been reported from the Cardiovascular Health Study, Framingham Heart Study, and Multi-Ethnic Study of Atherosclerosis, with hazard ratios of 1.60 (95% confidence interval [CI], 1.35-1.91), 4.7 (95% CI, 2.7-8.1), and 8.69 (95% CI, 4.90-15.45), respectively.^110,111,114 In a meta-analysis of older (≥ 60 years) adults, the median prevalence of asymptomatic left ventricular diastolic dysfunction was 36%.¹²¹ The presence, severity, and progression of diastolic dysfunction have been demonstrated in both clinical and community-based populations to be associated with increased risk for incident HF.^{108,122,123,124} In the Framingham Heart Study, left ventricular hypertrophy was associated with increased risk of incident HF, with eccentric hypertrophy portending a greater risk for HFrEF, whereas concentric hypertrophy was associated with greater risk for HFpEF.^125,126 Increased arterial stiffness and vascular remodeling have also been associated with greater risk of HF.^127,128,129

Lifetime Risk and Risk Prediction Models

The lifetime risk for the development of HF is high. In the Framingham Heart Study, the overall lifetime risk among white men and women was 20%.¹³⁰ Even among individuals without a history of myocardial infarction, the lifetime risk of HF was substantial, estimated at one in nine men and one in six women.¹³⁰ Estimates for lifetime risk using data pooled from the Chicago Heart Association, ARIC, and Cardiovascular Health Studies corroborated the high lifetime risks of HF observed in the Framingham Heart Study and extended the findings to black individuals.¹³¹ The lifetime risks for HF ranged from 30% to 42%, 20% to 29%, 32% to 39%, and 24% to 46% in white men, black men, white women, and black women, respectively.¹³¹ The lifetime risk of HF also differed according to the presence of modifiable risk factors in the ARIC study, as previously mentioned.¹⁰⁵

Heart failure risk prediction models from community cohorts have been developed (Table 69–6). Factors included in the risk prediction models vary by study, but typically include age, sex, heart rate, systolic blood pressure, body mass index, and diabetes mellitus. Other factors, such as smoking status, serum creatinine, coronary heart disease, antihypertensive medication use, left ventricular hypertrophy, and valvular disease, are included to varying extents. The c-statistics for model performance are good, ranging from 0.73 to 0.87.^{96,97,98,132,133} Further attention has been given to the use of circulating biomarkers for risk prediction. Although a number of candidate biomarkers have been studied, the natriuretic peptides (eg, B-type natriuretic peptide [BNP] and N-terminal pro-BNP [NT-proBNP]) are the most robust biomarker predictors of incident HF after adjusting for clinical factors.^{97,98,134,135} It is also noteworthy that prediction algorithms that can quantify risk of developing cardiovascular disease incidence in general seem to work reasonably well for predicting incidence of HF.¹³⁶

Outcomes in Prevalent Heart Failure

Morbidity

Once HF develops, morbidity and mortality are high. The adverse impact of HF on quality of life is well established, with substantial reductions in both physical function and mental health compared with the general population.^137,138,139 The impairment in quality of life experienced by individuals with HF is comparable to, or worse than, patients on chronic dialysis and those with chronic lung disease.^137,138 Quality of life is similar between patients with HFrEF and HFpEF, and declines with worse NYHA classification.^137,138,140 The syndrome of HF is complex with many systemic effects, as patients are at increased risk of stroke, atrial fibrillation, diabetes, chronic kidney disease, and venous thromboembolism.^{141,142,143,144,145,146} Recent attention has also focused on the increased burden of cognitive impairment among patients with HF, specifically with HF as a risk factor for dementia and with cognitive impairment as a risk factor for adverse outcomes among patients with HF.^147,148

The most commonly used end point for HF–related morbidity is hospitalizations, which are frequent for both cardiovascular and noncardiovascular causes (see Fig. 69–3).^{48,71,75,149,150,151,152,153} For HF hospitalizations, length of stay varies substantially worldwide, with median times ranging from 4 to 20 days.¹⁵⁴ Prior hospitalization for HF is a strong predictor of future HF hospitalization, with 90-day readmission rates approaching 30%.^154,155

Mortality

Survival among individuals with symptomatic HF (AHA/ACC stages C and D) is generally poor, with approximately 50% alive at 5 years following diagnosis.^56,156 However, there is variation in this risk according to disease severity and other factors. Death risk is directly proportional to NYHA class among patients with both preserved and reduced ejection fraction and in hospital and outpatient settings. The elderly are at particularly high risk, with 1-year mortality among Medicare beneficiaries approaching 30%.⁷² In meta-analyses, the mortality in ambulatory patients with HFrEF (deaths per 1000 person-years 159; 95% CI, 154-165) exceeds that of patients with HFpEF (deaths per 1000 person-years 146; 95% CI, 138-154).^157,158 The highest risk period appears to follow a hospitalization for HF, with 28-day and 1-year case fatality rates of 10% and 30%, respectively.^159,160,161 Indeed in the post–HF hospitalization setting, mortality may be similar between patients with HF with reduced and preserved ejection fraction.^64,162,163 Other factors reported to influence post–HF survival include blood pressure, right ventricular dysfunction and pulmonary hypertension, diabetes, anemia, chronic kidney disease, chronic obstructive pulmonary disease, and sleep apnea.^{46,164,165,166,167}

The association between body mass index and prognosis in HF is particularly noteworthy. Obesity, a prominent risk factor for the development of HF, paradoxically appears to be related to better outcomes among individuals with prevalent HF.^{168,169,170,171,172,173} Thus, in the setting of acute or chronic HF, individuals with low body mass index are at increased risk of adverse cardiovascular outcomes and death compared with overweight or obese individuals. The obesity paradox has been consistently observed across a number of HF populations and does not appear to be solely attributable to cardiac cachexia, although the mechanisms underlying the relationship between body mass and composition, HF, and outcomes are not completely understood.^{168,169,170,171,172,173} It has also been proposed that the obesity paradox seen in observational studies may not be biologically driven, but rather a statistical phenomenon caused by unmeasured confounders.¹⁷⁴

Despite HF portending a poor prognosis with high mortality, risk prediction for the individual patient remains challenging.¹⁷⁵ Survival may differ according to etiology of HF, with ischemic heart disease, particularly in the setting of reduced LVEF, consistently reported to have the lowest survival compared with nonischemic etiologies.^77,176 To aid prognostication, risk prediction models for mortality in HF, such as the Seattle Heart Failure Model and Heart Failure Survival Score, have been derived and validated largely from cohorts of individuals with reduced ejection fraction and advanced HF (Table 69–7).^177,178 In a recent meta-analysis, these scores were found to be fair to modest in their discriminatory capacity for predicting death in HF, with c-statistics ranging from 0.56 to 0.81.¹⁷⁹ More recently, another model for predicting mortality in chronic ambulatory HF was developed by the Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) using data from 39,372 patients largely enrolled in clinical trials.¹⁸⁰ A scoring system was developed based on 13 predictors, including age, LVEF, NYHA class, serum creatinine, diabetes, β-blocker use, systolic blood pressure, body mass, time since diagnosis, current smoking, chronic obstructive pulmonary disease, male sex, and angiotensin-converting enzyme inhibitor or angiotensin receptor blockers use. The advantages of the MAGGIC model are that it spans the spectrum of LVEF and uses readily available clinical predictors. The disadvantage is that it was largely derived from clinical trial populations.

The mode of death among HF patients is not always related to HF or even cardiovascular causes, with noncardiovascular causes accounting for 34% and 54% of deaths among HF patients in the Framingham Heart Study and Olmsted County, Minnesota, study respectively.^56,181 Noncardiovascular causes of death are particularly frequent among patients with HFpEF.^56,181 Cardiovascular causes of death in chronic HF include sudden death (36.4%), worsening of pump failure (27.5%), or death caused by a vascular event such as myocardial infarction (5.3%) or stroke (4.7%).¹⁸²

A full discussion of HF therapies is described in Chapters 70, 71, 72,73. Tremendous advances in the treatment of HF have been made over the past several decades. More attention has been given to HFrEF than to HFpEF, despite a similar proportion of patients of each type. Many landmark trials have defined the therapeutic approach to HFrEF (Fig. 69–5).^{15,16,17,183,184,185,186,187,188,189,190,191,192,193,194,195,196,197,198,199,200,201,202,203,204,205} In contrast, the few phase III randomized clinical trials in HFpEF have been neutral. With the increasing percentage of cases of HFpEF, it is reasonable to anticipate a larger number of future clinical trials in this population. A critical factor in trial design and interpretation will be to ensure the HF population studied meets standardized criteria for diagnosing HF both with regard to inclusion criteria and event adjudication. Additionally, matching clinical trial populations with those identified from community-based studies remains an important aim.

The prevalence of HF is forecast to increase substantially over the next 10 to 15 years as a result of changing risk factor profiles, Westernization of developing countries, reductions in mortality, and the aging of the population.^48,49 Modifiable risk factors underlie the majority of cases of HF; therefore, HF prevention strategies may greatly reduce the global burden of HF. Similarly, improved adherence to evidence-based guidelines for patients with chronic HFrEF may limit morbidity and reduce the risk of death.^{24,33,152,206} However, efficacious therapies for acute decompensated HF and HFpEF are lacking. Because HF is a heterogeneous syndrome with considerable interindividual variability, detailed phenotyping and personalized medicine that includes genomics, proteomics, and metabolomics may refine HF diagnosis, etiologic classification, prognostication, and therapeutic approaches.^207,208,209

Heart failure is a major public health problem. Despite tremendous advances in the prevention and treatment of cardiovascular disease, HF remains highly prevalent, morbid, and lethal. It is also a complex syndrome that can be difficult to diagnose, often because of comorbidities, such as obesity, chronic obstructive pulmonary disease, and anemia, that are themselves associated with dyspnea. Although recent data indicate a possible decline in the incidence rates of HF in developed countries, an overall increase in the number of individuals with HF is anticipated in the coming years. In a large number of cases, HF may be preventable as a result of modifiable risk factors. The morbidity and mortality associated with HF remain high, with a substantial proportion of patients undergoing repeated hospitalizations and succumbing to death as a result of cardiovascular and noncardiovascular causes. Adherence to existing evidence-based therapies for HFrEF may further limit the burden of this syndrome. HFpEF is becoming the dominant form of HF over time; however, efficacious therapies for this condition, as we all as for acute decompensated HF, are currently lacking.