Abstract e264

Top 10 Take-Home Messages e266

Preamble e267

Introduction e268

Definition of HF e272

Epidemiology and Causes of HF e278

Initial and Serial Evaluation e279

Stage A (Patients at Risk for HF) e292

Stage B (Patients With Pre-HF) e295

Stage C HF e297

Stage D (Advanced) HF e334

Patients Hospitalized With Acute Decompensated HF e341

Comorbidities in Patients With HF e351

Special Populations e356

Quality Metrics and Reporting e363

Goals of Care e364

Recommendation for Patient-Reported Outcomes and Evidence Gaps and Future Research Directions e367

References e371

Appendix 1

• Author Relationships With Industry and Other Entities (Relevant) e411

Appendix 2

• Reviewer Relationships With Industry and Other Entities (Comprehensive) e414

Appendix 3

• Appendix for Tables 3 and 4: Suggested Thresholds for Structural Heart Disease and Evidence of Increased Filling Pressures e421

Prevention of HF.

Management strategies in stage C HF, including:

Specific management strategies, including:

Implantable devices.

Left ventricular assist device (LVAD) use in stage D HF.

Clinical congestion can be assessed by various methods, including the presence of jugular venous distention (17), orthopnea (18), bendopnea (19), a square-wave response to the Valsalva maneuver (20), and leg edema (6). On a practical level, clinicians use extent of clinical congestion to guide titration of pharmacological treatments, including doses of diuretics. Observational studies have shown that clinical congestion is an important adverse risk factor in patients with HF (1-6,17). Recently, the PARADIGM-HF (The Efficacy and Safety of LCZ696 Compared to Enalapril on Morbidity and Mortality of Patients With Chronic Heart Failure) investigators showed that, in patients with chronic HFrEF, changes in markers of clinical congestion were associated with QOL as assessed by the Kansas City Cardiomyopathy Questionnaire and also provided prognostic information independently even of natriuretic peptides or the MAGGIC (Meta-analysis Global Group in Chronic Heart Failure) risk score (2). These data highlight the ongoing relevance of clinical congestion ascertained by the history and physical examination.

Some patients with HF progress to an advanced state, a condition that can be treated with specialized interventions such as mechanical circulatory support (MCS) or cardiac transplantation. Such patients should be identified before they progress to a state of extremis, at which point they may succumb to their illness or suffer complications of an intervention as a result of their very advanced state. Several “simple clinical clues” are available to identify advanced HF and should be ascertained via a focused history and physical examination. The recognition that a patient has advanced HF will allow for earlier referral to an advanced HF center, when appropriate, as will be discussed later in this document (see Section 8, “Specialty Referral for Advanced HF”).

Increasingly, familial cardiomyopathy is recognized as a more accurate diagnosis in some patients previously classified as having an idiopathic dilated cardiomyopathy (DCM). A detailed family history may provide the first clue of a genetic basis. A broad array of questions includes whether family members had a weak, enlarged, or thick heart, or HF; muscular dystrophy; a pacemaker or defibrillator; were on a heart transplant list; or died unexpectedly. Periodic updating of the family history in patients with a cardiomyopathy of uncertain origin may lead to a diagnosis of familial cardiomyopathy in the event that a relative subsequently develops a cardiomyopathy or a related complication. A 3-generation family pedigree obtained by genetic health care professionals improved the rate of detection of a familial process as compared with routine care (14). Furthermore, a family history of cardiomyopathy, as determined by a 3-generation pedigree analysis, was associated with findings of gadolinium enhancement on cardiac magnetic resonance imaging (MRI) and increased major adverse cardiac events (13). The possibility of an inherited cardiomyopathy provides the impetus for cascade screening of undiagnosed family members, thereby potentially avoiding preventable adverse events in affected relatives by implementation of GDMT and other management that otherwise would not be initiated.

Certain conditions that cause HF require disease-specific therapies. For example, in amyloid heart disease, whether on the basis of transthyretin (21) or light chain deposition (22), there are specific treatments that otherwise would not be used in patients with HF. Hence, expeditious and accurate diagnosis of such conditions is important. Currently, important delays have been reported in diagnosing amyloid heart disease (16), perhaps not unexpectedly given the wide spectrum of possible clinical presentations (15). Similarly, HF attributable to sarcoidosis, hemochromatosis, hypothyroidism, hyperthyroidism, acromegaly, connective tissue disease, tachycardia-induced cardiomyopathy, or high-output HF from an arteriovenous fistula, among others, requires specific therapeutic approaches. Given that the differential diagnosis of HF is broad, the history and physical examination can provide clues to narrow the number of causes to consider and guide the diagnostic approach to identify such conditions (Table 5).

The history and physical examination help to identify the cause of a clinical deterioration. To determine the cause of a clinical deterioration, the clinician assesses for concurrent illness (e.g., ongoing myocardial ischemia, pulmonary emboli, or systemic infection), initiation of a medication potentially detrimental in the setting of HF (e.g., nonsteroidal anti-inflammatory drugs [NSAIDs]), or the possibility of chronic RV pacing (e.g., a newly implanted pacemaker or medications such as amiodarone that leads to bradycardia and resultant chronic RV pacing), nonadherence to a medication or dietary regimen, and ongoing substance abuse. In addition, an assessment of social determinants of health (e.g., housing stability, food security, available transportation) should be made.

Identifying the specific cause of HF is important, because conditions that cause HF may require disease-specific therapies. Depending on the clinical suspicion, additional diagnostic studies are usually required to diagnose specific causes (Table 6) such as ischemic cardiomyopathy, cardiac amyloidosis, sarcoidosis, hemochromatosis, infectious mechanisms (e.g., HIV, COVID-19, Chagas), hypothyroidism, hyperthyroidism, acromegaly, connective tissue disorders, tachycardia-induced cardiomyopathy, Takotsubo, peripartum cardiomyopathy, cardiotoxicity with cancer therapies, or substance abuse would require specific management in addition to or beyond GDMT (1,2,9-15).

Laboratory evaluation with complete blood count, urinalysis, serum electrolytes, blood urea nitrogen, serum creatinine, glucose, fasting lipid profile, liver function tests, iron studies (serum iron, ferritin, transferrin saturation), and thyroid-stimulating hormone levels provides important information regarding patients’ comorbidities, suitability for and adverse effects of treatments, potential causes or confounders of HF, severity and prognosis of HF, and is usually performed on initial evaluation. Pertinent laboratory tests are repeated with changes in clinical condition or treatments (e.g., to monitor renal function or electrolytes with diuretics).

Electrocardiography is part of the routine evaluation of a patient with HF and provides important information on rhythm, heart rate, QRS morphology and duration, cause, and prognosis of HF. It is repeated when there is a clinical indication, such as a suspicion for arrhythmia, ischemia or myocardial injury, conduction, or other cardiac abnormalities.

Measurement of BNP and NT-proBNP levels in the ambulatory setting for a suspected cardiac cause of dyspnea provides incremental diagnostic value to clinical judgment when the cause of dyspnea is unclear and the physical examination equivocal (1-9). In the emergency setting, BNP and NT-proBNP levels have higher sensitivity than specificity and may be more useful for ruling out HF than ruling in HF. Although lower levels of BNP and NT-proBNP may help exclude the presence of HF, and higher levels have high positive predictive value to diagnose HF, increases in both BNP and NT-proBNP levels have been reported in patients with various cardiac and noncardiac causes (Table 6) (50-53).

and 3. Higher levels of BNP and NT-proBNP are associated with a greater risk for adverse short- and long-term outcomes in patients with HF, including all-cause and cardiovascular death and major cardiovascular events (11,13-19). Studies have shown incremental prognostic value of these biomarkers to standard approaches of CVD risk assessment (11,16). Not all patients may need biomarker measurement for prognostication, especially if they already have advanced HF with established poor prognosis or persistently elevated levels of biomarkers in former settings.

The STOP-HF (St Vincent’s Screening to Prevent Heart Failure) study is a large single-center trial of patients at risk of HF, defined by the presence of hypertension, diabetes, or known vascular disease but without established LV systolic dysfunction or symptomatic HF, who were randomly assigned to screening with BNP testing or usual care (31). Participants in the intervention group with BNP levels ≥50 pg/mL underwent echocardiography and referral to a cardiovascular specialist (31). All patients received coaching by a specialist nurse who provided education on the importance of adherence to medication and healthy lifestyle behaviors (31). BNP-based screening reduced the composite endpoint of incident asymptomatic LV dysfunction with or without newly diagnosed HF. Similarly, accelerated uptitration of renin-angiotensin-aldosterone system (RAAS) antagonists and beta blockers reduced cardiac events in patients with diabetes and elevated NT-proBNP levels but without cardiac disease at baseline (30). Standardized screening for HF remains challenging as a result of the heterogeneity of risk factors across different patient populations. Studies are needed to assess the cost-effectiveness and risks of such screening, as well as its impact on QOL and mortality.

Predischarge BNP and NT-proBNP levels are strong predictors of the risk of death or hospital readmission for HF (14,17,20-29). Although patients in whom levels of BNP or NT-proBNP decreased with treatment had better outcomes than those without any changes or with a biomarker rise (14,23,28,29), targeting a certain threshold, value, or relative change in these biomarker levels during hospitalization has not been shown to be consistently effective in improving outcomes (37-39). Patients in which GDMT leads to a reduction in BNP and NT-proBNP levels represent a population with improved long-term outcomes compared with those with persistently elevated levels despite appropriate treatment (37-39). BNP and NT-proBNP levels and their change could help guide discussions on prognosis as well as adherence to, and optimization of, GDMT.

Inherited dilated, restrictive, and hypertrophic cardiomyopathies have been identified, although 1 gene variant may cause different phenotypes in the same family. The most common pathogenic variants identified are truncations in the large structural protein titin, which have been implicated in DCM (3-5) and also in peripartum or alcoholic cardiomyopathies; however, variants that do not cause disease are also common. Pathogenic variants in lamin A/C can be associated with conduction block and atrial arrhythmias as well as ventricular arrhythmias, which may progress more rapidly than symptoms of HF. Although previously linked with the phenotype of arrhythmogenic RV cardiomyopathy, desmosomal protein variants are now recognized to affect the left ventricle also with or without the right ventricle, and the term arrhythmogenic cardiomyopathy is now preferred for the phenotype of arrhythmias combined with DCM. Filamin-C mutations have been associated with skeletal myopathies and with isolated cardiomyopathy with ventricular arrhythmias. The identification of pathogenic variants associated with increased risk of sudden death may trigger consideration of primary prevention implantable cardioverter-defibrillators (ICDs) even in patients who have LVEF >0.35 or <3 months of guideline-recommended therapies (6). Evidence of desmosomal cardiac disease carries the additional implication of advice to avoid strenuous exercise, which may accelerate ventricular remodeling (7). Genetic confirmation of symptomatic Fabry’s cardiomyopathy is an indication for replacement therapy with the enzyme agalsidase beta, and migalastat was recently approved for this uncommon cardiomyopathy.

The chest x-ray is a useful initial diagnostic test for the evaluation of patients presenting with signs and symptoms of HF because it assesses cardiomegaly, pulmonary venous congestion, and interstitial or alveolar edema and may reveal alternative causes, cardiopulmonary or otherwise, of the patient’s symptoms (1,2). Apart from congestion, other findings on chest x-ray are associated with HF only in the context of clinical presentation. Importantly, cardiomegaly may be absent in acute HF and, although cephalization, interstitial edema, and alveolar edema are modestly specific for HF, these findings are relatively insensitive (2,33). Considering the limited sensitivity and specificity, the chest x-ray should not be used as the only determinant of the specific cause or presence of HF.

TTE provides information regarding cardiac structure and function and identifies abnormalities of myocardium, heart valves, and pericardium. Echocardiography reveals structural and functional information that predicts subsequent risk (34-40). Guidelines provide recommendations for quantification of cardiac structure and function, including LVEF measurements, ventricular dimensions and volumes, evaluation of chamber geometry, and regional wall motion (41). RV size and function, atrial size, and all valves are evaluated for anatomic and flow abnormalities. Guidelines also provide recommendations for diastolic function and estimates of LV filling and left atrial pressure (42). The tricuspid valve regurgitant gradient, coupled with inferior vena cava diameter and its response during respiration, provides estimates of systolic PA pressure and central venous pressure. Indices of myocardial deformation, such as global longitudinal strain, may identify subclinical LV systolic dysfunction, which has been associated with greater risk of developing HF or recurrent HF hospitalizations (38,43-46). Given the widespread availability, lack of ionizing radiation, and wealth of provided information, echocardiography is the preferred initial imaging modality for evaluation of patients with suspected HF. Point-of-care cardiac ultrasound is an evolving tool for assessment of cardiac function and assessment of volume status and pulmonary congestion (47-52).

Serial echocardiograms to assess changes in EF, structural remodeling, and valvular function, although not recommended routinely in stable patients, are useful in various situations. In patients who have an unexplained, significant change in clinical status, echocardiography can provide important information, such as worsening ventricular or valvular function. A subset of patients may also have reverse remodeling, improvement in LVEF, and valvular function in response to evidence-based medical, revascularization, and device therapies, and repeat assessment of LVEF and remodeling is appropriate in those who have received treatments that might have had a significant effect on cardiac structure and function (4-7,53-59). Recovery of function appears more common in those with LV systolic dysfunction occurring in the setting of adverse energetic circumstances (e.g., chronic tachycardia or thyroid disease), dilated cardiomyopathies associated with immune responses (e.g., peripartum cardiomyopathy, acute myocarditis, systemic inflammatory responses), or in those who have undergone revascularization or device-based therapies (60). Reevaluation of EF (>40 days after MI, >90 days after revascularization, >90 days after GDMT) is useful to determine candidacy for implantable cardioverter-defibrillator (ICD) or cardiac resynchronization therapy (CRT). Finally, repeat surveillance of LV function is appropriate in patients exposed to treatments that potentially damage the myocardium, such as chemotherapy.

If TTE is unable to accurately evaluate cardiac structure and function, additional noninvasive imaging modalities are available to clarify the initial diagnosis and to provide information on cardiac structure and function. The choice between these modalities depends on availability, local expertise, patient characteristics, indication, and goal of limiting radiation exposure. CMR provides an accurate and highly reproducible assessment of cardiac volumes, mass, and EF of the left and right ventricles (8-10). CMR provides high anatomic resolution of all aspects of the heart and surrounding structures and is not associated with ionizing radiation, leading to its recommended use in known or suspected congenital heart diseases (11,61). Electrocardiographic-gated cardiac CT can also accurately assess ventricular size, EF, and wall motion abnormalities, but it is accompanied with ionizing radiation (13-15). Radionuclide ventriculography is highly reproducible for measurement of LVEF, although it also exposes the patient to ionizing radiation (12).

CMR provides noninvasive characterization of the myocardium that may provide insights into HF cause (62). Late-gadolinium enhancement, reflecting fibrosis and damaged myocardium, can identify acute and chronic MI (63,64) and identify HF caused by CAD (65,66). Patterns of late-gadolinium enhancement or specific T-1 and T-2 techniques can suggest specific infiltrative and inflammatory cardiomyopathies, such as myocarditis, sarcoidosis, Fabry disease, Chagas disease, noncompaction, iron overload, and amyloidosis (16,20,22,67). T-1 mapping techniques allow for measurement of interstitial space characteristics and extracellular volume fraction and provides diagnostic and prognostic information (19,21-23,68-71). The presence of delayed hyperenhancement has been associated with worse outcomes and can provide risk stratification (72-77). Although registry data show that CMR findings commonly impact patient care management and provide diagnostic information in patients with suspected myocarditis or cardiomyopathy (17,18), a strategy of routine screening with CMR in patients with nonischemic cardiomyopathy was not shown to yield more specific HF causes than a strategy of selective CMR strategy based on echocardiographic and clinical findings in a recent trial (78).

HF is often caused by coronary atherosclerosis (79), and evaluation for ischemic heart disease can help in determining the presence of significant coronary artery disease (CAD). Noninvasive stress imaging with echocardiography or nuclear scintigraphy can be helpful in identifying patients likely to have obstructive CAD (24,25). Invasive or computed tomography coronary angiography can detect and characterize extent of CAD (26,27).

CAD is a leading cause of HF (79) and myocardial ischemia may contribute to new or worsening HF symptoms. Noninvasive testing (i.e., stress echocardiography, SPECT, CMR, or PET) may be considered for detection of myocardial ischemia to help guide coronary revascularization decisions. Multiple nonrandomized, observational studies have reported improved survival with revascularization in patients with viable but dysfunctional myocardium (28,30-32). Despite these observational data, RCTs have not shown that viability imaging improves guidance of revascularization to a reduction of adverse cardiovascular outcomes (80-82). A prespecified viability substudy of the STICH (Surgical Treatment for Ischemic Heart Failure) trial showed that the presence of myocardial viability did not determine the long-term benefit from surgical revascularization in patients with ischemic cardiomyopathy (81,82). Of note, a relatively small number of individuals enrolled in the STICH substudy did not have viability, which may limit the power of the study. Although these data do not support the concept of routine viability assessment before revascularization, myocardial viability is used as one of the tools to inform decisions regarding revascularization in patients with high surgical risk or with complex medical problems.

Repeat noninvasive imaging of cardiac structure and function for routine surveillance is rarely appropriate in the absence of a change in clinical status or treatment interventions (11,83).

Endomyocardial biopsy may be useful when seeking a specific diagnosis that would influence treatment, and biopsy should thus be considered in patients with rapidly progressive clinical HF or worsening ventricular dysfunction that persists despite appropriate medical treatment. Endomyocardial biopsy should also be considered in patients suspected of having acute cardiac rejection status after heart transplantation or having myocardial infiltrative processes. A specific example is to determine treatment for light chain (AL) amyloidosis or transthyretin amyloidosis (5). Additional indications for endomyocardial biopsy include patients with rapidly progressive and unexplained cardiomyopathy and those in whom active myocarditis, especially giant cell myocarditis, is being considered (1).

Right-heart catherization in patients in acute HF. The ESCAPE (Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness) trial found that routine use of PA catheter monitoring for patients with HF did not provide benefit (3). However, invasive hemodynamic evaluation or monitoring can be useful to guide management in carefully selected patients with acute HF who have persistent symptoms despite treatment. This includes patients whose fluid status, perfusion, or systemic or pulmonary vascular resistance is uncertain whose systolic blood pressure (SBP) remains low, or is associated with symptoms, despite initial treatment; whose renal function is worsening with therapy; or who require parenteral vasoactive agents.

There has been no established role for routine or periodic invasive hemodynamic measurements in the management of HF. Most drugs used to treat HF are prescribed on the basis of their ability to improve symptoms or survival rather than their effect on hemodynamic variables. The initial and target doses of these drugs are generally selected on the basis of controlled trial experience rather than changes produced in cardiac output or pulmonary capillary wedge pressure (3,4).

Patients with HF should not undergo routine endomyocardial biopsy because of the risk of complications that include perforation, cardiac tamponade, and thrombus formation, as well as limited diagnostic yield (5,6).

The CHAMPION (CardioMEMS Heart Sensor Allows Monitoring of Pressure to Improve Outcomes in NYHA Class III Heart Failure patients) trial reported a significant 28% reduction of HF-related hospitalizations after 6 months in patients randomized to an implanted PA pressure monitor compared with a control group (1). Patients had to have a HF hospitalization in the previous year and be on stable doses of a beta blocker and angiotensin-converting enzyme inhibitor (ACEi) (or angiotensin (II) receptor blocker [ARB]) if tolerated. The clinical benefit persisted after longer term follow-up and was seen in both subjects with reduced (3) and preserved (2) LVEF. However, CHAMPION was a nonblinded trial, and there was differential contact of study personnel with patients in the treatment arm, raising methodological concerns about the opportunity for bias to have influenced its results (19-21). In the recent GUIDE-HF (Haemodynamic-GUIDEed management of Heart Failure) study, hemodynamic-guided management of patients with NYHA class II to IV heart failure did not significantly reduce the composite endpoint rate of mortality and total HF events (4). The usefulness of noninvasive telemonitoring (11,12,22,23) or remote monitoring of physiological parameters (13-18) (e.g., patient activity, thoracic impedance, heart rate) via implanted electrical devices (ICDs or CRT-Ds) to improve clinical outcomes remains uncertain. Further study of these approaches is needed before they can be recommended for routine clinical care.

Three model-based studies (5-7) have evaluated the cost-effectiveness of wireless PA pressure monitoring using data from the CHAMPION-HF (1) study of the CardioMEMS device. All 3 studies estimated CardioMEMS implantation and monitoring increased survival and quality-adjusted life year (QALY) while increasing costs. Primarily based on differences regarding the expected magnitude of clinical benefit, 2 analyses (5,7) estimated the device provided high value while the third (6) estimated intermediate value. These analyses had several important differences detailed in the evidence tables, including the model duration, QOL data, cost estimates, and assumptions regarding mortality. One analysis (6) found the economic value of CardioMEMS implantation was highly dependent on its effect on mortality and duration of treatment benefit, both of which remain unclear. Cost-effectiveness studies incorporating data from GUIDE-HF (4) have not been published. Additional data regarding clinical outcomes following CardioMEMS implantation will improve estimates of its economic value.

NYHA functional classification is an ordinal, categorical variable (I-IV) that is used to document functional limitation in patients with cardiac disease, including HF (1). In HF, NYHA functional class I includes patients with no limitations in physical activity resulting from their HF. NYHA class II includes patients who are comfortable at rest but have slight symptoms resulting from HF (dyspnea, fatigue, lightheadedness) with ordinary activity. NYHA class III includes patients who are comfortable at rest but have symptoms of HF with less than ordinary activity. NYHA class IV includes patients who are unable to carry out any physical activity without symptoms and have symptoms at rest. NYHA functional classification has been widely used in clinical practice, clinical trials, and clinical practice guidelines to determine candidacy for drug and device therapy. Limitations include its ability to be inconsistently assessed from 1 clinician to another, resulting in poor reproducibility (23).

Many CPET variables have been associated with prognosis in patients with HF (4,5,12,14,16,24). Peak exercise oxygen consumption/oxygen uptake (VO₂) is often used to risk stratify patients and make decisions about timing of advanced HF therapies, including heart transplantation and LVAD. In a landmark article (7), investigators divided patients referred for heart transplantation into groups based on their peak VO₂ (7). Patients with peak VO₂ <14 mL/kg/min were listed for transplant, while those with higher peak VO₂ values were deferred for being too well. Patients with peak VO₂ >14 mL/kg/min who were deferred had 1- and 2-year survival of 94% and 84%, respectively, which was similar to survival after heart transplant. As such, the authors proposed peak VO₂ ≤14 mL/kg/min as a cutoff to distinguish patients who may derive survival benefit from heart transplant (7). Patients tolerating beta blockers may have improved survival with an equivalent VO₂ compared with patients who do not tolerate beta blockers (25,26). For patients on beta blockers, a peak VO₂ ≤12 mL/kg/min has been suggested as a more appropriate cutoff to consider cardiac transplant listing (8).

Objective assessment of exercise capacity with CPET can be useful in the clinical management of patients with HF. Although CPET remains the gold standard measure of exercise capacity, limitations to more widespread use include need for special equipment and trained personnel, which leads to lack of availability at many hospitals and clinics. Furthermore, it is not well tolerated by some patients. The 6-minute walk test is an alternative way to measure exercise capacity that is widely available and well tolerated by patients. It entails walking for 6 minutes on a measured flat course, and patients are allowed to slow down or stop if needed. A systematic review of 14 studies found that the 6-minute walk test results correlated moderately with peak VO₂ levels and were a reliable and valid indicator of functional capacity in patients with HF who did not walk >490 m (8). Distance walked in the 6-minute walk test has been associated with prognosis in HF across multiple studies (9-13,15,16,27). A cutoff of <300 m roughly correlates to patients with NYHA class III to IV symptoms and is associated with worse 3-year survival free of heart transplant (62% versus 82% for those walking ≥300 m) (27).

Dyspnea is a complex symptom that can reflect abnormalities in a number of different systems and can be influenced by psychological and environmental factors. CPET involves having patients perform a treadmill (or stationary bicycle) exercise test, while also performing ventilatory gas exchange measurements (28). CPET enables the comprehensive assessment of multiple physiological measures that can impact exercise capacity and contribute to dyspnea. It provides analysis of gas exchange and yields measures of oxygen uptake (VO₂), carbon dioxide output, and ventilation. These measures can be integrated with standard exercise testing variables, such as heart rate, blood pressure, electrocardiographic findings, and symptoms to provide insights into the physiologic mechanisms underlying a patient’s dyspnea. In particular, CPET can help to distinguish respiratory versus cardiac etiologies of dyspnea. If exercise capacity is diminished but cardiopulmonary responses are normal, other causes of dyspnea, such as metabolic abnormalities and deconditioning, should be considered.

For HF, there are several clinical models to consider that include the spectrum of HF based on EF and clinical setting. For chronic HF, the Seattle Heart Failure Model (2), the Heart Failure Survival score (1), and the MAGGIC score (3) have commonly been used to provide estimates of survival. The MAGGIC predictive model may be quite useful given its derivation and validation across multiple clinical trials and cohorts, including more recent studies. For chronic HFrEF, there are additional models that include other clinical variables, including exercise capacity (7) and natriuretic peptide levels (8). Likewise, for chronic HFpEF there are more specific predictive models for that population derived from clinical trial data (9,10). In acute HF, several clinical models may be used to predict short-term survival (11-13).

Elevated systolic and diastolic blood pressure are major risk factors for the development of symptomatic HF (8,9,32). Many trials have shown that hypertension control reduces the risk of HF (1-7). Although the magnitude of benefit varies with the patient population, target blood pressure reduction, and HF criteria, effective hypertension treatment invariably reduces HF events. In the SPRINT (Systolic Blood Pressure Intervention Trial) trial, control to an SBP goal <120 mm Hg decreased incident HF by 38% and mortality by 23% compared with an SBP goal of <140 mm Hg (6,7). A meta-analysis showed that blood pressure control was associated with an approximately 40% reduction in HF events (5). Therefore, SBP and diastolic blood pressure should be controlled in accordance with published clinical practice guidelines (30).

Multiple RCTs in patients with type 2 diabetes and at risk for, or with established CVD or at high risk for CVD, have shown that SGLT2i prevent HF hospitalizations compared with placebo (10-12). The benefit for reducing HF hospitalizations in these trials predominantly reflects primary prevention of symptomatic HF, because only approximately 10% to 14% of participants in these trials had HF at baseline. The mechanisms for the improvement in HF events have not been clearly elucidated but seem to be independent of glucose lowering. Proposed mechanisms include reductions in plasma volume, cardiac preload and afterload, alterations in cardiac metabolism, reduced arterial stiffness, and interaction with the Na⁺/H⁺ exchanger (33,34). SGLT2i are generally well tolerated, but these agents have not been evaluated in those with severe renal impairment (estimated glomerular filtration rate [eGFR] <25 mL/min/1.73 m²) (35).

Greater adherence to healthy lifestyle habits such as regular physical activity, avoiding obesity, maintaining normal blood pressure and blood glucose, not smoking, and healthy dietary patterns have been associated with a lower lifetime risk of HF and greater preservation of cardiac structure (13-16,27). Healthful eating patterns, particularly those that are based more on consumption of foods derived from plants, such as the Mediterranean, whole grain, plant-based diet and the DASH (Dietary Approaches to Stop Hypertension) diet, are inversely associated with incident HF and may offer some protection against HF development (17-21).

A large-scale unblinded single-center study (STOP-HF [The St Vincent’s Screening to Prevent Heart Failure]) (22) of patients at risk of HF (identified by the presence of hypertension, diabetes, or known vascular disease) but without established LV systolic dysfunction or symptomatic HF at baseline found that screening with BNP testing and then intervening on those with levels of ≥50 pg/mL (performing echocardiography and referral to a cardiovascular specialist) reduced the composite endpoint of asymptomatic LV dysfunction (systolic or diastolic) with or without newly diagnosed HF (22). Similarly, in another small, single-center RCT, accelerated uptitration of RAAS antagonists and beta blockers reduced cardiac events in patients with diabetes and elevated NT-proBNP levels but without cardiac disease at baseline (23).

Incident HF may be predicted from different models, including those derived from diverse populations (Table 9). The PCP-HF (Pooled Cohort equations to Prevent HF) model provides race- and sex-specific 10-year risk equations from 7 community-based cohorts with at least 12 years of follow-up (29). Predictors of HF included in the race- and sex-specific models were age, blood pressure (treated or untreated), fasting glucose (treated or untreated), body mass index, cholesterol, smoking status, and QRS duration. Models can be applied to the clinical setting of interest, with clinical trial models potentially less generalizable to registry- or population-based models. In addition, predictive models provide the average estimate of risk derived from a population, and individual risk may vary (36). The integration of risk scores into clinical practice have shown improved outcomes. As data generation increases from electronic health records and digital sources, advanced methods with machine learning are expected to proliferate the development of risk prediction models. Machine learning models are often not externally validated, and their performance may vary based on the population and clinical setting (37). Patient populations change over time, and models may need to be recalibrated periodically.

ACEi have been shown to impede maladaptive remodeling after acute MI in patients with reduced LVEF (36,37). In survivors of acute MI with asymptomatic LV dysfunction (LVEF <35%–40%), RCTs have shown that ACEi reduced mortality, HF hospitalizations, and progression to severe HF compared with placebo (2,4). Similarly, in those individuals with asymptomatic LV dysfunction in the SOLVD (Studies of Left Ventricular Systolic Dysfunction) prevention trial, which included approximately 20% without ischemic heart disease, enalapril was associated with reduced HF hospitalization and mortality compared with placebo (1,3).

In multiple RCTs (42), statins have been shown to prevent adverse CAD events in patients with an MI, ACS, and with high cardiovascular risk. These trials have also shown that statin therapy reduces the risk of incident HF (5-9). A meta-analysis of 6 RCTs of >110,000 patients with an ACS showed that intensive statin therapy reduced hospitalizations for HF (5). A subsequent, larger collaborative meta-analysis of up to 17 major primary and secondary prevention RCTs showed that statins reduced HF hospitalization (42). These data support the use of statins to prevent symptomatic HF and cardiovascular events in patients with acute MI or ACS.

Two major trials have compared ARB with ACEi after MI. The VALIANT (Valsartan in Acute Myocardial Infarction) trial, which included approximately 25% of patients with asymptomatic LV dysfunction, showed that the benefits of valsartan on mortality and other adverse cardiovascular outcomes were comparable to captopril (10,38). In the OPTIMAAL (Optimal Trial in Myocardial Infarction with the Angiotensin II Antagonist Losartan) trial, losartan did not meet the noninferiority criteria for mortality compared with captopril (39). It has been hypothesized that the lower dose of losartan (50 mg daily) in the OPTIMAAL trial may have contributed to the greater difference than those seen with valsartan in VALIANT (40). No clinical trials have specifically evaluated ARB in patients with asymptomatic reduced LVEF in the absence of previous MI. Although ARB are alternatives for patients with ACEi-induced angioedema, caution is advised because some patients have also developed angioedema with ARB.

Current evidence supports the use of beta blockers to improve adverse cardiac remodeling and outcomes in patients with asymptomatic reduced LVEF after MI. Among patients with a recent MI and reduced LVEF, carvedilol reduced maladaptive remodeling (41) and reduced mortality compared with placebo (11). Among patients with asymptomatic LV systolic dysfunction in the SOLVD prevention trial (which included 80% with previous MI) and the SAVE (Survival and Ventricular Enlargement) trial, secondary analyses showed that the administration of beta blockers in addition to ACEi reduced mortality and hospitalization (12,13).

The Framingham studies have shown a 60% increased risk of death in patients with asymptomatic low LVEF compared with those with normal LVEF, and almost half of these patients remained free of HF before their death (25). MADIT-II (Multicenter Automatic Defibrillator Implantation Trial II) showed a 31% relative risk reduction in all-cause mortality in patients with post-MI with LVEF ≤30% receiving a prophylactic ICD compared with standard of care (14). These findings provided justification for the broad adoption of ICDs for primary prevention of SCD in the post-MI setting with reduced LVEF, even in the absence of HF symptoms.

Although beta blockers have been shown to improve outcomes in patients with symptomatic HFrEF and in patients with reduced LVEF after MI (11), few data exist regarding the use of beta blockers in asymptomatic patients with depressed LVEF without a history of MI. There is evidence to support the role of beta blockers to prevent adverse LV remodeling in asymptomatic patients with LV systolic dysfunction, including those with nonischemic cause (43). Also, in a post hoc analysis of the SOLVD prevention trial, which included approximately 20% of participants with nonischemic HF cause, beta blockers were associated with a reduction in the risk of death and in death or hospitalization for symptomatic HF in those patients randomized to enalapril, a finding that was not seen in the placebo group (12). Given the long-term benefits of beta blockers to reduce HF hospitalizations in patients with symptomatic HFrEF (44), beta-blocker therapy is recommended to prevent symptomatic HF in patients with reduced LVEF.

Thiazolidinediones have been associated with fluid retention and increased rates of HF in RCTs of patients with type 2 diabetes who were predominantly free of symptomatic HF at baseline (47-49). In a smaller RCT of patients with more severely symptomatic HFrEF, pioglitazone was associated with increased rates of HF hospitalization compared with placebo (50). In patients with more mild symptoms (NYHA class I to II) but with depressed LVEF (15), rosiglitazone was associated with more fluid-related events, including worsening edema and need for increased HF medications (15). Given the evidence, thiazolidinediones should be avoided in patients with reduced LVEF.

Nondihydropiridine calcium channel blockers diltiazem and verapamil are myocardial depressants and generally not tolerated in HF. In previous studies of patients with HF or reduced LVEF after acute MI, diltiazem was associated with increased risk of HF (16,17), although in a smaller study of patients with nonischemic cardiomyopathy, diltiazem had no impact on mortality (45). Verapamil had no impact on survival or major cardiovascular events after acute MI (46). Although not specifically tested in asymptomatic patients with low LVEF, nondihydropyridine calcium channel blockers may be harmful in this population because of their negative inotropic effects.

In a meta-analysis of 30 RCTs, multidisciplinary interventions reduced hospital admission and all-cause mortality (1). In a separate meta-analysis of 22 RCTs, specialized multidisciplinary team follow-up was associated with reduced HF hospitalizations and all-cause hospitalizations (2). In a recent meta-analysis of 22 RCTs, multidisciplinary interventions that included a pharmacist reduced HF hospitalizations (3). In a recent Cochrane systematic review and meta-analysis of 43 RCTs, both case management (i.e., active management of complex patients by case managers working in integrated care systems) and multidisciplinary interventions (i.e., coordinated multidisciplinary health care interventions and communications) were shown to reduce all-cause mortality, all-cause readmission, and readmission for HF (4).

Meta-analyses of RCTs have shown that interventions focused on improving HF self-care significantly reduce the risk of HF-related hospitalization (2,5-8), all-cause hospitalization (2,8,9), and all-cause mortality (6,9), as well as improve QOL (5). Interventions that aim to improve self-care knowledge and skill (2,5,8), and those that focus on enhancing medication adherence (9) or reinforce self-care with structured telephone support (6,7), are effective in patients with HF. There is uncertainty whether mobile health–delivered educational interventions improve self-care in patients with HF (39). In a single RCT involving rural patients with HF, an educational intervention was shown to improve knowledge and self-care (40) but did not significantly decrease the combined endpoint of cardiac death or HF hospitalization (41). In a recent pragmatic trial, a transitional care services program that included self-care education improved discharge preparedness, quality of transition, and QOL but did not significantly improve clinical outcomes compared with usual care (42).

In propensity-adjusted models, influenza vaccination was associated with a significant reduction in all-cause mortality among participants in PARADIGM-HF (Prospective Comparison of ARNi with ACEi to Determine Impact on Global Mortality and Morbidity in Heart Failure) (14). In adjusted models, influenza vaccination was associated with significant reductions in all-cause mortality and cardiovascular mortality (12) in 1 registry study and was associated with significant reductions in all-cause mortality and the composite of all-cause mortality and cardiovascular hospitalizations in another large cohort study (11). In a self-controlled case series study of patients with HF, influenza vaccination was associated with a significantly lower risk of cardiovascular, respiratory, and all-cause hospitalization (43). In a meta-analysis of 16 studies of patients with CVD, influenza vaccination was associated with a lower risk of all-cause, cardiovascular mortality, and major adverse cardiovascular events compared with control patients (15). In the Cardiovascular Health Study, pneumococcal vaccination was associated with significant reductions in incident HF, all-cause mortality, and cardiovascular mortality (16). Patients with HF are uniquely susceptible to poor outcomes in the setting of SARS-CoV-2 infection (44-47) and should be vaccinated against COVID-19 (10).

Many health and social factors are associated with poor HF self-care (36,37) (Table 11) but have also been linked to poor clinical outcomes and fundamentally change how education and support must be delivered. Depression is a risk factor for poor self-care (40), rehospitalization (17), and all-cause mortality (18) among patients with HF. Interventions that focus on improving HF self-care have been reported to be effective among patients with moderate/severe depression with reductions in hospitalization and mortality risk (5). Nonrandomized studies have provided evidence of a link between social isolation and mortality in patients with HF (19,20). In a recent meta-analysis of 29 cohort studies, frailty was associated with an increased risk of all-cause mortality and hospitalization (23). Frailty also has been shown to impair self-care among elderly patients with HF (24). A recent meta-analysis of observational studies revealed social isolation to be common among adults with HF (i.e., 37%) and associated with a 55% greater risk of HF-related rehospitalization (21). Poor social support also has been shown in nonrandomized studies to be associated with lower HF self-care (22). A recent meta-analysis of observational studies showed that inadequate/marginal health literacy is common among adults with HF (i.e., 24%) and associated independently with the risk of mortality and hospitalization (25). Low literacy also is associated with poor HF self-care, as most interventions depend on both literacy and health literacy/numeracy (26).

A registered dietitian- or nurse-coached intervention with 2 to 3 g/d sodium restriction improved NYHA functional class and leg edema in patients with HFrEF (1). In a nonrandomized study (>2.5 g/d versus <2.5 g/d), lower dietary sodium was associated with worse all-cause mortality in patients with HFrEF (2). In small RCTs, aggressive sodium restriction (0.8 g/d) during hospitalization for acute decompensated HF has not reduced weight, congestion, diuretic use, rehospitalization, or all-cause mortality in patients with HFrEF (24) or in patients with HFpEF (25). A recent pilot RCT N=27) showed that providing patients with 1.5 g/d sodium meals can reduce urinary sodium and improve QOL but not improve clinical outcomes (3). Another recent pilot RCT (N=38) of 1.5 versus 2.3 g/d sodium resulted in sodium intake and improvement in BNP levels and QOL in the 1.5 g/d sodium intake arm (5); the full trial is due to be completed in 2022. A third pilot RCT (N=66) of home-delivered 1.5 g/d meals showed favorable but nonsignificant trends toward improvement in clinical status and readmission rates (6). Moreover, results from RCTs have shown that reducing dietary sodium is difficult to achieve in patients with HF, even with prepared meals (3) or home visits (26).

Evidence from RCTs indicates that exercise training improves functional status, exercise performance, and QOL in patients with HFrEF and HFpEF. In HF-ACTION, the largest randomized trial with exercise training in patients with HF (1), 2331 patients with LVEF ≤35% (NYHA class II and III) were randomized to usual care versus supervised exercise training plus usual care. There were modest reductions in all-cause mortality and hospitalization rates that did not reach significance by primary analysis but, after prespecified adjustment, were associated with reductions in cardiovascular mortality or HF hospitalizations (1). Many RCTs of exercise training in HF have been conducted, but the statistical power of most was low (2-5,9-13). Meta-analyses suggest that exercise training is associated with improvement in functional capacity, exercise duration, health-related QOL, and reduction in HF hospitalizations in patients with HFrEF as well as HFpEF (2-6,8,11,14,15). Most studies and meta-analyses have not shown significant changes in all-cause mortality (2,12,14-22), except for a few showing mortality benefit with longer follow-up (6,7). Other benefits of exercise training include improved endothelial function, blunted catecholamine spillover, increased peripheral oxygen extraction, and improvement in peak oxygen consumption (2-5,8,10-12,21).

A formal cardiac rehabilitation program usually includes a medical evaluation, education regarding the importance of medical adherence, dietary recommendations, psychosocial support, and an exercise training and physical activity counseling program. Exercise-based cardiac rehabilitation has been associated with an improvement in functional capacity, exercise tolerance, the rate of overall and HF-specific hospitalization, and improved QOL (3,4,6,7,11,16,17). In a diverse population of older patients who were hospitalized for acute decompensated HF, an early, transitional, tailored, progressive rehabilitation intervention that included multiple physical-function domains (strength, balance, mobility, and endurance) initiated during, or early after hospitalization for HF, and continued after discharge, resulted in greater improvement in physical function than usual care (9).

Controlled trials with diuretics showed their effects to increase urinary sodium excretion, decrease physical signs of fluid retention, and improve symptoms, QOL, and exercise tolerance (1-5). Recent data from the nonrandomized OPTIMIZE-HF (Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure) registry revealed reduced 30-day all-cause mortality and hospitalization for HF with diuretic use compared with no diuretic use after hospital discharge for HF (9). The most commonly used loop diuretic for the treatment of HF is furosemide, but some patients respond more favorably to other agents in this category (e.g., bumetanide, torsemide), potentially because of their increased oral bioavailability (10-12). In outpatients with HF, diuretic therapy is commonly initiated with low doses, and the dose is increased until urine output increases and weight decreases, generally by 0.5 to 1.0 kg daily. Patients may become unresponsive to high doses of diuretic drugs if they consume large amounts of dietary sodium, are taking agents that can block the effects of diuretics (e.g., NSAIDs), or have significant impairment of renal function or perfusion.

Diuretic resistance can be overcome in several ways, including escalation of loop diuretic dose, intravenous administration of diuretics (bolus or continuous infusion) (6), or combination of different diuretic classes (13-16). The use of a thiazide or thiazide-like diuretic (e.g., metolazone) in combination with a loop diuretic inhibits compensatory distal tubular sodium reabsorption, leading to enhanced natriuresis. However, in a propensity-score matched analysis in patients with hospitalized HF, the addition of metolazone to loop diuretics was found to increase the risk for hypokalemia, hyponatremia, worsening renal function, and mortality, whereas use of higher doses of loop diuretics was not found to adversely affect survival (17). Although randomized data comparing the 2 diuretic strategies are limited, the DOSE (Diuretic Optimization Strategies Evaluation) trial lends support for the use of high-dose intravenous loop diuretics (18).

An ARNi is composed of an ARB and an inhibitor of neprilysin, an enzyme that degrades natriuretic peptides, bradykinin, adrenomedullin, and other vasoactive peptides. In PARADIGM-HF (Prospective Comparison of ARNi with ACEi to Determine Impact on Global Mortality and Morbidity in Heart Failure), an RCT that compared the first approved ARNi, sacubitril-valsartan, with enalapril in symptomatic patients with HFrEF tolerating an adequate dose of either ACEi or ARB, sacubitril-valsartan significantly reduced the composite endpoint of cardiovascular death or HF hospitalization by 20% relative to enalapril (1). The benefit was observed to a similar extent for death and HF hospitalization and was consistent across prespecified subgroups (1). Use of an ARNi is more frequently associated with symptomatic hypotension and a comparable incidence of angioedema when compared with enalapril (1). Sacubitril-valsartan has been approved for patients with symptomatic HF. HF effects and potential off-target effects may be complex with inhibition of the neprilysin enzyme, which has multiple biological targets. Trial data have included ACEi/ARB-naïve patients before ARNi initiation (53% in the PIONEER-HF [Comparison of Sacubitril-Valsartan versus Enalapril on Effect on NT-proBNP in Patients Stabilized from an Acute Heart Failure Episode] trial and 24% in the TRANSITION [Comparison of Pre- and Post-discharge Initiation of Sacubitril/Valsartan Therapy in HFrEF Patients After an Acute Decompensation Event] trial) and have shown similar efficacy and safety in treatment-naïve patients (2,3). The PIONEER-HF trial showed that ARNi reduced NT-proBNP levels in patients hospitalized for acute decompensated HF without increased rates of adverse events (worsening renal function, hyperkalemia, symptomatic hypotension, angioedema) when compared with enalapril (3). Additional outcome analyses suggested reduction in all-cause mortality and rehospitalization for HF but were only hypothesis-generating as exploratory study endpoints. In the open-label TRANSITION trial, patients with HFrEF hospitalized with worsening HF were randomized to start ARNi either before or after discharge (2). Safety outcomes were similar for both arms, suggesting that early initiation may simplify management (rather than initiating and uptitrating ACEi first and then switching to ARNi) (2). ARNi should be initiated de novo in patients hospitalized with acute HFrEF before discharge in the absence of contraindications. ARNi may be initiated de novo in patients with chronic symptomatic HFrEF to simplify management, although data are limited. The PARADISE-MI (Prospective ARNi vs ACE Inhibitor Trial to DetermIne Superiority in Reducing Heart Failure Events After MI) trial (40) will provide information on whether sacubitril-valsartan will significantly reduce the rate of cardiovascular death, HF hospitalization or outpatient HF requiring treatment in patients after acute MI, with LVEF ≤40% and/or pulmonary congestion, and 1 of 8 additional risk-enhancing factors like AF, previous MI, diabetes, compared with the ACEi ramipril; and whether the safety and tolerability of sacubitril-valsartan was comparable to that of ramipril. Thus, at the present time, the efficacy of ARNi in patients with LV dysfunction, and HF in the early post-MI period, remains uncertain.

ACEi reduce morbidity and mortality in HFrEF. RCTs clearly establish the benefits of ACE inhibition in patients with mild, moderate, or severe symptoms of HF and in patients with or without CAD (6-11). Data suggest that there are no differences among available ACEi in their effects on symptoms or survival (12). ACEi should be started at low doses and titrated upward to doses shown to reduce the risk of cardiovascular events in clinical trials. ACEi can produce angioedema and should be given with caution to patients with low systemic blood pressures, renal insufficiency, or elevated serum potassium (>5.0 mEq/L). If maximal doses are not tolerated, intermediate doses should be tried; abrupt withdrawal of ACE inhibition can lead to clinical deterioration and should be avoided. Although the use of an ARNi in lieu of an ACEi for HFrEF has been found to be superior, for those patients for whom ARNi is inappropriate, continued use of an ACEi for all classes of HFrEF remains strongly advised.

ARB have been shown to reduce mortality and HF hospitalizations in patients with HFrEF in large RCTs (14-16). Long-term treatment with ARB in patients with HFrEF produces hemodynamic, neurohormonal, and clinical effects consistent with those expected after interference with the renin-angiotensin system (17,18). Unlike ACEi, ARB do not inhibit kininase and are associated with a much lower incidence of cough and angioedema, although kininase inhibition by ACEi may produce beneficial vasodilatory effects. Patients who are intolerant to ACEi because of cough or angioedema should be started on an ARB. ARB should be started at low doses and titrated upward, with an attempt to use doses shown to reduce the risk of cardiovascular events in clinical trials. ARB should be given with caution to patients with low systemic blood pressure, renal insufficiency, or elevated serum potassium (>5.0 mEq/L). Although ARB are alternatives for patients with ACEi-induced angioedema, caution is advised because some patients have also developed angioedema with ARB. For those patients for whom an ACEi or ARNi is inappropriate, use of an ARB remains advised.

Several cost-effectiveness analyses consistently found that ACEi therapy provides high value for patients with chronic HF. A model-based analysis, using generic ACEi costs, found ACEi therapy was high value (19). Previous analyses also found ACEi therapy was high value despite previously higher ACEi costs (19,21,22,24,25). This includes a trial-based analysis of SOLVD (Studies of Left Ventricular Dysfunction) that modeled long-term outcomes (21). Previous analyses included a range of clinical scenarios including asymptomatic LV dysfunction (24) and LV dysfunction after MI (25), with ACEi therapy providing high value in each. There are limited data on the cost-effectiveness of ARBs from 2 clinical trials—a within-trial analysis of Val-HeFT (Valsartan Heart Failure Trial) (23) and an analysis of the ELITE (Evaluation of Losartan in the Elderly) study (20)—which both suggested ARB therapy is high value. The high value of ARB therapy is also supported by its similar efficacy as ACEi therapy and the low-cost generic availability for both medication classes.

Patients with chronic stable HFrEF who tolerate ACEi and ARB should be switched to ARNi. In patients with mild-to-moderate HF who were able to tolerate both a target dose of enalapril (10 mg twice daily) and then subsequently an ARNi (sacubitril-valsartan; 200 mg twice daily, with the ARB component equivalent to valsartan 160 mg), hospitalizations and mortality were significantly decreased with the valsartan-sacubitril compound compared with enalapril (1). Another RCT and meta-analysis showed improvement in LV remodeling parameters with ARNi compared with enalapril (4,5).

Multiple model-based analyses evaluated the economic value of ARNi therapy compared with ACEi therapy using the results of PARADIGM-HF (26-29,41). Three high-quality analyses (26,28,29) consistently found costs per QALY <$60,000, which provides high value according to the benchmarks adopted for the current clinical practice guideline. These results were robust to the range of sacubitril-valsartan costs currently seen in care. These results were sensitive to the estimated mortality reduction and duration of treatment effectiveness. ARNi would need to maintain effectiveness beyond the PARADIGM-HF study period (mean, 27 months) to be considered high value (29). If clinical benefit were limited to 27 months, ARNi would be intermediate value. One additional analysis, based on the PIONEER-HF trial, found that inpatient initiation of ARNi was also high value compared with delayed initiation postdischarge (27).

Oral neprilysin inhibitors, used in combination with ACEi, can lead to angioedema, and concomitant use is contraindicated and should be avoided. A medication that represented a neprilysin inhibitor and an ACEi—omapatrilat—was studied in hypertension and HF, but its development was terminated because of an unacceptable incidence of angioedema (30,31) and associated significant morbidity. This adverse effect was thought to occur because ACEi and neprilysin break down bradykinin, which can directly or indirectly cause angioedema (31,32). An ARNi should not be administered within 36 hours of switching from or to an ACEi.

Omapatrilat, a neprilysin inhibitor (as well as an ACEi and aminopeptidase P inhibitor), was associated with a higher frequency of angioedema than that seen with enalapril in an RCT of patients with HFrEF (30). In a very large RCT of hypertensive patients, omapatrilat was associated with a 3-fold increased risk of angioedema compared with enalapril (31). Black patients and patients who smoked were particularly at risk. The high incidence of angioedema ultimately led to cessation of the clinical development of omapatrilat (33,34). Because of these observations, angioedema was an exclusion criterion in the first large trial assessing ARNi therapy in patients with hypertension (35) and then in the large trial that showed clinical benefit of ARNi therapy in HFrEF (1). The rates of angioedema were numerically higher in patients treated with ARNi than in patients treated with ACEi in PARADIGM-HF, although this difference did not reach significance (1). ARNi therapy should not be administered in patients with a history of angioedema because of the concern that it will increase the risk of a recurrence of angioedema.

Angioedema attributable to ACEi is thought to result from defective degradation of the vasoactive peptides bradykinin, des-Arg9-BK (a metabolite of bradykinin), and substance P (36,37). ACEi should not be administered to patients with any history of angioedema, but ARB do not interfere as directly with bradykinin metabolism and have been associated with low rates of angioedema (38,39).

Three beta blockers have been shown to be effective in reducing the risk of death in patients with HFrEF: bisoprolol, sustained-release metoprolol (succinate), and carvedilol (1-3). The favorable findings with these 3 agents, however, should not be considered a beta-blocker class effect in HFrEF. Other beta blockers are not included in this recommendation for use (13-15). Even when asymptomatic, or when symptoms are mild or improve with other therapies, beta-blocker therapy is important and should not be delayed until symptoms return or disease progression is documented (16). Data show that beta blockers can be safely initiated before hospital discharge, provided patients are clinically stabilized and do not require intravenous inotropic therapy for HF (17). If a contraindication or intolerance are noted, they should be documented, and the patient restarted on beta-blocker therapy in the future, so long as an absolute contraindication is not present. Even if symptoms or LVEF improve, long-term treatment with beta blockers and use of target doses should be maintained to reduce the risk of progression in LV dysfunction or major cardiovascular events (18,19). Abrupt withdrawal of beta-blocker therapy can lead to clinical deterioration and should be avoided unless indicated (18).

Multiple analyses have shown the high value of beta-blocker therapy among HF patients. A model-based analysis, using generic beta-blocker costs, found beta-blocker therapy was high value (4). These results were consistent with earlier model-based cost-effectiveness analyses (5-7) and a trial-based economic analysis of the U.S. Carvedilol Heart Failure (CHF) Trials Program (8). Each of these studies also found treatment with a beta blocker was high value despite using previously higher beta-blocker costs.

Clinical trials taken on MRA together—RALES (Randomized Aldactone Evaluation Study) (1) randomized highly symptomatic patients with LVEF ≤35%; EPHESUS (Eplerenone Post–Acute Myocardial Infarction Heart Failure Efficacy and Survival Study) (2) randomized patients post-MI with LVEF ≤40%; and EMPHASIS-HF (Eplerenone in Mild Patients Hospitalization and Survival Study in Heart Failure) (3) randomized patients with mild symptoms and LVEF ≤30%—suggest a benefit of MRA across the spectrum of HFrEF, inclusive of a wide range of etiologies and disease severities. Initiation in the ambulatory or hospital setting is appropriate (12). The starting dose of spironolactone and eplerenone is 25 mg orally daily, increased to 50 mg daily orally after a month; for eGFR 31 to 49 mL/min/1.73 m², dosing should be reduced by half. Regular checks of serum potassium levels and renal function should be performed according to clinical status, approximately 1 week, then 4 weeks, then every 6 months after initiating or intensifying MRA, with more frequent testing for clinical instability. We elected to remove the 2013 recommendation “Aldosterone receptor antagonists are recommended to reduce morbidity and mortality following an acute MI in patients who have LVEF of 40% or less who develop symptoms of HF or who have a history of diabetes mellitus, unless contraindicated” because the new recommendation covers the spectrum of symptomatic patients with HF.

The economic value of MRA therapy was assessed by both RCTs (RALES [5] and EPHESUS [6,7]) and a model-based analysis (4). The model-based analysis used generic MRA costs and found therapy was high value with a cost per QALY of under $1000 (4). The earlier trial-based economic analyses of MRAs from RALES and EPHESUS also found MRA therapy was high value despite using previously higher MRA costs (5-7).

Spironolactone and eplerenone are partially excreted through the kidneys, raising concerns about safety when eGFR is ≤30 mL/min/1.73 m² (10,11). Spironolactone and eplerenone decrease renal potassium excretion, raising the risk of hyperkalemia, particularly when MRA is initiated at serum potassium ≥5.0 mEq/L and continued ≥5.5 mEq/L. The incidence of clinically significant hyperkalemia events was <1% in EPHESUS and EMPHASIS-HF, without a significant difference between eplerenone and placebo (2,3); however, in the closely monitored setting of a RCT with enrollment of younger patients with fewer multiple chronic conditions than seen in the general HFrEF population, safety may be overstated. Observational data have raised concerns about less favorable outcomes of MRA use for HFrEF during usual care (8,9). Coadministration of MRA with ACEi or ARB mildly increases the risk of hyperkalemia. Hyperkalemia risk was lower with ARNi in patients with chronic HF in the PARADIGM-HF trial (13) but not different in patients with HF who were decompensated in the PIONEER-HF trial (14) when compared with ACEi. Diarrhea causing dehydration or loop diuretic therapy interruption, because of worsening renal function or hyperkalemia, should be a consideration for temporarily holding the MRA. The development of worsening renal function or hyperkalemia is often a reflection of acute clinical change or progressive disease, prompting careful evaluation of the entire medical regimen and other causes of hyperkalemia, in addition to holding the MRA. The efficacy of the use of potassium binders (e.g., patiromer, sodium zirconium cyclosilicate) to improve outcomes by facilitating continuation of MRA is uncertain (15,16) and is addressed in Section 7.3.6, “Other Drug Treatment”.

In the DAPA-HF and EMPEROR-Reduced trials, SGLT2i compared with placebo reduced the composite of cardiovascular death or HF hospitalization by approximately 25% (1,2,9). The benefit in reduction of HF hospitalization was greater (30%) in both trials (9). Risk of cardiovascular death was significantly lowered (18%) with dapagliflozin, as was risk of all-cause mortality (17%). Although no significant cardiovascular mortality benefit was observed with empagliflozin in a meta-analysis of DAPA-HF and EMPEROR-Reduced trials, SGLT2i therapy was associated with a reduction in all-cause mortality and cardiovascular death (9). The benefits in both trials were seen irrespective of baseline diabetes status. Furthermore, serious renal outcomes were less frequent, and the rate of decline in eGFR was slower in patients treated with SGLT2i (1,2,9). In the SOLOIST-WHF (Effect of Sotagliflozin on Cardiovascular Events in Patients With Type 2 Diabetes And Worsening Heart Failure) trial, patients with diabetes and HF hospitalization (79%: LVEF, <50%) were enrolled before discharge or within 3 days of discharge. Sotagliflozin, a dual inhibitor of sodium-glucose co-transporters 1 and 2, reduced the combined endpoint of cardiovascular death, HF hospitalization, or urgent HF visits by 33% (10) but has not been approved by the U.S. Food and Drug Administration (FDA) as of 2021. Although SGLT2i increased risk for genital infections, they were otherwise well tolerated in the trials. As the use of SGLT2i is translated into clinical practice, caution is warranted for euglycemic ketoacidosis, genital and soft tissue infections, and adjustment of diuretics, if needed, to prevent volume depletion (11).

Two model-based analyses evaluated the economic value of dapagliflozin therapy compared with usual care based on the results of the DAPA-HF trial (3,4). Both analyses found costs per QALY between $60,000 and $90,000, which is consistent with intermediate value according to the benchmarks adopted for the current guideline. The results were most sensitive to the magnitude of cardiovascular mortality reduction, with a ≥8% reduction in cardiovascular mortality necessary for a cost per QALY below $150,000 in 1 study (3). There are a wide range of costs currently seen with dapagliflozin. These 2 analyses estimated a cost per QALY below $50,000 with annual dapagliflozin costs of $3240 (43% reduction from main analysis) and $2500 (40% reduction from main analysis), respectively (3,4). A smaller reduction in drug cost would lead to a cost per QALY of under $60,000, the threshold for high value in this guideline.

In a large-scale trial that compared the vasodilator combination with placebo, the use of hydralazine and isosorbide dinitrate reduced mortality in patients with HF treated with digoxin and diuretics but not an ACEi or beta blocker (4). However, in 2 other trials that compared the vasodilator combination with an ACEi, the ACEi produced more favorable effects on survival (6,7). A post hoc retrospective analysis of these vasodilator trials showed particular efficacy of isosorbide dinitrate and hydralazine in the African American cohort (1). In a subsequent trial, which was limited to patients self-identified as African American, the addition of a fixed-dose combination of hydralazine and isosorbide dinitrate to standard therapy with an ACEi or ARB, a beta blocker, and MRA offered significant benefit (2). Thus, the combination of hydralazine and isosorbide dinitrate is appropriate for African Americans with HFrEF who remain symptomatic despite concomitant use of ACEi (or ARB), beta blockers, and MRA. There are insufficient data for concomitant use with ARNi.

The economic value of hydralazine and isosorbide nitrate therapy was assessed by the A-HeFT trial (3). This analysis found hydralazine and isosorbide dinitrate increased survival and reduced health care costs over the 12.8-month trial. Extrapolating beyond the trial, the analysis found hydralazine and isosorbide dinitrate remained high value over a lifetime with a cost per life-year <$60,000 despite conservative assumptions regarding the durability of therapy effectiveness and previously higher hydralazine and isosorbide dinitrate costs.

It is unclear if a benefit of hydralazine-isosorbide dinitrate (suggested in a trial before the use of ACEi) (4) exists for non–African Americans with HFrEF. Despite the lack of data with the vasodilator combination in patients who are intolerant of ACEi or ARB, especially those with renal insufficiency, the combined use of hydralazine and isosorbide dinitrate might be considered as a therapeutic option in such patients. Although the potential benefit is unknown and has not been shown in recent observational datasets (5), in V-HeFT I, the use of hydralazine and isosorbide dinitrate reduced mortality in patients with HF treated with digoxin and diuretics, compared with placebo (4). If patients are unable to tolerate first-line agents, such as ARNi, ACEi, or ARB, because of drug intolerance, hypotension, or renal insufficiency, referral to a HF specialist can provide guidance for further management, and the use of hydralazine and isosorbide dinitrate in these patients might be considered.

Supplementation with omega-3 PUFA has been evaluated as an adjunctive therapy for CVD and HF (14). The GISSI-HF (Effect of n-3 polyunsaturated fatty acids in patients with chronic heart failure) trial showed a reduction in death among post-MI patients taking 1 g of omega-3 PUFA (850–882 mg of eicosapentaenoic acid [EPA] and docosahexaenoic acid [DHA] as ethyl esters in the ratio of 1:1.2) (10). A post hoc subgroup analysis revealed that this reduction in mortality and SCD was concentrated in the approximately 2000 patients with reduced LVEF (10). The GISSI-HF investigators randomized symptomatic patients with HF to 1 g daily of omega-3 PUFA (850–882 mg of EPA-DHA) or placebo. Death from any cause was reduced from 29% with placebo to 27% in those treated with omega-3 PUFA (2). The outcome of death or admission to hospital for a cardiovascular event was also significantly reduced. The REDUCE-IT trial randomized patients with established CVD or diabetes with risk factors to 2 g of icosapent ethyl (a highly purified EPA) twice daily or placebo and showed a reduced risk for the composite of cardiovascular death, nonfatal MI, nonfatal stroke, coronary revascularization, or unstable angina (3). In reported studies, omega-3 PUFA therapy has been well tolerated. Recent studies have reported that in patients with cardiovascular risk treated with omega-3 fatty acid, there may be a dose-related risk of AF (3,15,16).

Hyperkalemia is common in HF as a result of the syndrome itself, comorbidities (diabetes, CKD), and use of RAASi, and can increase the risk for ventricular arrhythmias and mortality (11). Hyperkalemia results in dose reductions or discontinuation of RAASi, compromising their cardiorenal benefit in HF (12). Two newer gastrointestinal potassium binders—patiromer (RLY5016) and sodium zirconium cyclosilicate (SZC)—remove potassium by exchanging cations (calcium for patiromer, and sodium and hydrogen for SZC), leading to increased fecal excretion. Both agents have been FDA approved for treatment of hyperkalemia for patients receiving RAASi. In the PEARL-HF (Evaluation of the efficacy and safety of RLY5016, a polymeric potassium binder in patients with chronic heart failure) trial, patiromer led to lower potassium levels, less hyperkalemia, and a higher proportion of patients able to increase spironolactone dose to 50 mg daily compared with placebo (5). The HARMONIZE (Hyperkalemia Randomized Intervention Multidose ZS-9 Maintenance) trial included 94 patients (out of 258 total) with HF (87 of whom entered the double-blind phase) (6,13). The SZC groups achieved lower potassium levels overall compared with placebo, and a higher proportion maintained normokalemia (potassium levels, <5.1 mEq/L). Whether patiromer or SZC improve clinical outcomes is under investigation. Adverse effects for the newer potassium binders include hypomagnesemia (for patiromer) and edema (for SZC).

In several retrospective analyses, the risk of thromboembolic events was not lower in patients with HF taking warfarin than in patients not treated with antithrombotic drugs (17-19). The use of warfarin was associated with a reduction in major cardiovascular events and death in patients with HF in some studies but not in others (20-22). An RCT that compared the outcome of patients with HFrEF assigned to aspirin, warfarin, or clopidogrel found that no therapy was superior (7). Another trial that compared aspirin with warfarin in patients with reduced LVEF, sinus rhythm, and no cardioembolic source showed no difference in either the primary outcome of death, stroke, or intracerebral hemorrhage, and no difference in the combined outcome of death, ischemic stroke, intracerebral hemorrhage, MI, or HF hospitalization (8). There was a significant increase in major bleeding with warfarin. A trial of rivaroxaban in patients with HFrEF, CAD, and normal sinus rhythm showed no difference in mortality, MI, and stroke compared with placebo (9). Therefore, there is no evidence of benefit for anticoagulation in HF patients without a specific indication (e.g., VTE, AF, a previous thromboembolic event, or a cardioembolic source).

Second-generation dihydropyridine calcium channel blockers, including amlodipine and felodipine, have greater selectivity for calcium channels in vascular smooth muscle cells and less myocardial depressant activity. By reducing peripheral vasoconstriction and LV afterload, calcium channel blockers were thought to have a potential role in the management of chronic HF. The PRAISE-1 (Prospective Randomized Amlodipine Survival Evaluation-1) study showed a reduction in mortality in the subgroup of patients with nonischemic cardiomyopathy who received amlodipine (1). However, in the PRAISE-2 (Prospective Randomized Amlodipine Survival Evaluation 2) trial, which enrolled only patients with nonischemic cardiomyopathy, no survival benefit was observed, indicating the limitations of conclusions derived from subgroup analyses (29). However, dihydropyridine calcium channel blockers may be used for treatment of hypertension in patients who have elevated blood pressure despite optimization of GDMT.

Many nutritional supplements and hormonal therapies have been proposed for the treatment of HF (3-9,30,31). Ultimately, most studies are limited by small sample sizes, surrogate endpoints, or nonrandomized design (32,33). In addition, adverse effects and drug-nutraceutical interactions remain unresolved. There is a lack of evidence of benefit from vitamin D (3-5), thiamine (34-36), carnitine (37), and taurine (38,39) and potential harm from vitamin E (6,7). The largest RCT of coenzyme Q10—Q-SYMBIO (Coenzyme Q10 as adjunctive treatment of chronic heart failure with focus on SYMptoms, BIomarker status [Brain-Natriuretic Peptide], and long-term Outcome [hospitalisations/mortality])—showed no changes in NYHA functional classification at 16 weeks, although the incidence of major adverse cardiovascular events at 2 years was significantly reduced (hazard ratio, 0.50; 95% CI, 0.32-0.80; P=0.003) (8). Despite these findings, concerns about slow recruitment in this trial have tempered enthusiasm for coenzyme Q10 supplementation in clinical practice (9,31). Hormonal therapies have been proposed for the treatment of HF, but trials have shown a neutral effect of testosterone (40,41), growth hormone (30,42), and thyroid hormone (43-45) in HF outcomes.

Nondihydropyridine calcium channel blockers—diltiazem and verapamil—are myocardial depressants and generally not well tolerated in HF. Verapamil had no impact of survival or major cardiac events post-MI, including in those patients with HFrEF after acute MI (10). In patients with nonischemic cardiomyopathy, diltiazem had no impact on mortality (13) but, in HFrEF after acute MI, diltiazem was associated with a higher risk of recurrent HF (11,12).

In the CAST (Cardiac Arrhythmia Suppression) trial, patients with asymptomatic ventricular arrhythmias post-MI on the class IC antiarrhythmics encainide or flecainide had increased mortality (14). The applicability of CAST to patients without recent MI or to other class I antiarrhythmic drugs is uncertain, but class IC antiarrhythmic agents are generally avoided in patients with structural heart disease. In ANDROMEDA (Antiarrhythmic Trial with Dronedarone in Moderate to Severe CHF Evaluating Morbidity Decrease Study), for the class III antiarrhythmic dronedarone, patients with HFrEF who were hospitalized had increased mortality (16). In the SWORD (Survival With ORal D-sotalol) trial of the class III antiarrhythmic sotalol, patients with HF post-MI had increased mortality (15). However, SWORD was published in 1996, and whether sotalol would be harmful in the current era of GDMT and ICDs is uncertain; sotalol may be used for refractory atrial-ventricular arrhythmias with close monitoring for decompensation. Amiodarone (46,47) and dofetilide (48,49) are the only antiarrhythmic agents with neutral effects on mortality in clinical trials of patients with HFrEF. Class IA antiarrhythmic agents such as quinidine and class IB agents such as mexiletine have not been studied and may be indicated for the management of refractory ventricular arrhythmias in the context of the individual patient’s risk benefit calculus and in conjunction with electrophysiology consultation.

Thiazolidinediones increase insulin sensitivity by activating nuclear peroxisome proliferator-activated receptor gamma (PPAR-γ). Expressed in virtually all tissues, PPAR-γ also regulates sodium reabsorption in the collecting ducts of the kidney. In observational cohort studies (17), meta-analysis (18), and clinical trials (19-21), thiazolidinediones have been associated with increased incidence of fluid retention and HF events in those patients with (19,21) or without (18,20) a previous history of HF.

DPP-4 is a cell-surface enzyme that deactivates several peptides include glucose-dependent insulinotropic polypeptide and glucagon-like peptide 1. DPP-4 inhibitors affect glucose regulation through multiple mechanisms, including enhancement of glucose-dependent insulin secretion, slowed gastric emptying, and reduction of postprandial glucagon and of food intake. The impact of DPP-4 inhibitors on cardiovascular outcomes in patients with diabetes and high cardiovascular risk has been assessed in multiple RCTs. Saxagliptin increased the risk of hospitalization for HF (22), as did alogliptin in a post hoc analysis including only patients with no HF history (23,50), but sitagliptin (51,52) and linagliptin (53-55) did not; these findings may have been a result of baseline differences in the use of metformin, thiazolidinediones, and insulin, which also affect HF risk. The FDA recommends discontinuation specifically of saxagliptin and alogliptin in patients who develop HF (56), and whether the risk of worsening HF is a class effect of DPP-4 inhibitors is unclear.

NSAIDs inhibit the synthesis of renal prostaglandins, which mediate vasodilation in the kidneys and directly inhibit sodium resorption in the thick ascending loop of Henle and collecting tubule. Hence, NSAIDs can cause sodium and water retention and blunt the effects of diuretics. Several observational cohort studies have revealed increased morbidity and mortality in patients with HF using either nonselective or selective NSAIDs (25-28).

The use of these specific medications for HFrEF should involve initiation at low-starting doses, uptitration at specified intervals as tolerated, and achieving-maintaining the target doses shown to be effective in major clinical trials. Every effort should be made by clinicians to achieve and maintain the clinical trial–defined target doses (Table 13) of guideline-directed medications, as long as they are well tolerated by the patient. Patients should be monitored for changes in heart rate, blood pressure, electrolytes, renal function, and symptoms during this uptitration period. Planned uptitration of a HF medication should be delayed until any adverse effects observed with lower doses have resolved. When such a strategy is used for dose titration, most patients (approximately 70%–85%) enrolled in clinical trials who received these medications were able to tolerate short-, intermediate-, and long-term treatment with these agents and achieve and maintain the trial defined target dose (1-9,11-14). Repeated attempts at uptitration can result in optimization, even if initial attempts may fail. In patients with HFrEF, beta blockers provide dose-dependent improvements in LVEF, reduction in HF hospitalizations, and reduction in all-cause mortality (17). Trials of lower versus higher dose of ACEi and ARB have shown lower risk of cardiovascular death or HF hospitalization with higher doses, with similar safety and tolerability (17,18).

Initiation and titration should be individualized and optimized without delay according to patient’s symptoms, vital signs, functional status, tolerance, renal function, electrolytes, comorbidities, specific cause of HF, and ability of follow-up. In patients with HFrEF, simultaneous initiation or sequencing, and order of guideline-directed medications are usually individualized according to patient’s symptoms, vital signs, functional status, tolerance, renal function, electrolytes, comorbidities, specific cause of HF, and ability of follow-up, and does not necessarily need to be done according to the sequence of trial publications and should not be delayed.

Although the primary outcome in SHIFT was a composite of hospitalization and cardiovascular death, the greatest benefit was a reduction in HF hospitalization. SHIFT included patients with HFrEF and LVEF ≤35% who were in sinus rhythm with a resting heart rate of ≥70 bpm. Participants were predominantly NYHA class II and III. Participants had been hospitalized for HF in the preceding 12 months and were on stable GDMT for 4 weeks before initiation of ivabradine therapy (1-4). The target of ivabradine is heart rate, and the benefit of ivabradine results from a reduction in heart rate. However, only 25% of patients studied in SHIFT were on optimal doses of beta-blocker therapy. Given the well-proven mortality benefits of beta-blocker therapy, these agents should be initiated and uptitrated to target doses, as tolerated, before assessing the resting heart rate for consideration of ivabradine initiation (5,6).

Digoxin is usually initiated at a low dose because higher doses are rarely required in the management of HF and are potentially detrimental. Two retrospective analyses of large-scale clinical trials have shown a linear relationship between mortality and digoxin serum concentration in patients with AF and at risk for stroke, including those with HF, and in patients with HF. The risk of death was independently associated with serum digoxin concentration, with a significantly higher risk observed in those with concentrations ≥1.2 ng/mL and ≥1.6 ng/mL (8,9). The benefit of digoxin in patients with HF remains controversial. GDMT is expected to be optimized before considering the addition of digoxin. Clinical worsening after withdrawal of digoxin has been shown (10). Therapy with digoxin may either be continued in the absence of a contraindication or discontinued with caution (11). Therapy with digoxin is commonly initiated and maintained at a dose of 0.125 to 0.25 mg daily. Low doses (0.125 mg daily or every other day) should be used initially if the patient is >70 years of age, has impaired renal function, or has a low lean body mass. Higher doses (e.g., digoxin 0.375 to 0.50 mg daily) are rarely used or needed in the management of patients with HF.

Patients with HFrEF in the VICTORIA trial had LVEF <45%, NYHA class II to IV, were on GDMT, with elevated natriuretic peptides (BNP ≥300 pg/mL or NT-proBNP ≥1000 pg/mL if in sinus rhythm; higher cutoffs with AF), and recent HF worsening (hospitalized within 6 months or recently received intravenous diuretic therapy without hospitalization). Patients on long-acting nitrates, with SBP <100 mm Hg, or eGFR <15 mL/min/1.73 m² were excluded (1). Over a median follow-up of 10.8 months, the primary outcome, cardiovascular death or HF hospitalization, occurred in 35.5% with vericiguat compared with 38.5% with placebo (HR, 0.90; P=0.019). All-cause mortality occurred in 20.3% in the vericiguat group and 21.2% in the placebo group (HR, 0.95; 95% CI, 0.84-1.07; P=0.38) and composite of any-cause death or HF hospitalization was also lower in the vericiguat group versus placebo group (HR, 0.90; 95% CI, 0.83–0.98; P=0.02). The relative risk reduction of 10% in the primary outcome was lower than expected, even in a higher risk population. Although not statistically significant, symptomatic hypotension (9.1% versus 7.9%; P=0.12) and syncope (4.0% versus 3.5%; P=0.30) were numerically higher in the vericiguat group versus placebo. There was heterogeneity by subgroup analysis, and patients in the highest quartile of NT-proBNP subgroup (NT proBNP level >5314 pg/mL) did not have benefit from vericiguat when compared with placebo.

ICDs were first assessed in patients who had been resuscitated from a cardiac arrest. In AVID (Antiarrhythmics versus Implantable Defibrillators trial), CASH (Cardiac Arrest Study Hamburg), and CIDS (Canadian Implantable Defibrillator StudyS), benefit was observed in those who were randomized to ICDs (1-3). Extension of benefit was then shown in other patient populations that were at perceived risk of SCD. In the first MADIT (Multicenter Automated Defibrillator Implantation Trial) trial, patients with previous MI, LVEF ≤35% with nonsustained VT had a mortality benefit with ICD (4). Similar populations in MUSTT (Multicenter UnSustained Tachycardia Trial) also showed benefit (5). In MADIT-II (Multicenter Automatic Defibrillator Implantation Trial II), patients with no arrhythmia qualifier but with previous MIs and LVEF ≤30% derived benefit from ICD (6). The DEFINITE (Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation) study included only nonischemic patients with LVEF ≤35% and frequent premature ventricular contractions (PVCs) or nonsustained ventricular tachycardia (VT) (7). There was a trend to mortality benefit, but it ultimately did not achieve significance. In SCD-HEFT (Sudden Cardiac Death in Heart Failure Trial), patients with ischemic and nonischemic cardiomyopathy, LVEF ≤35%, and HF class II to III showed benefit with an ICD compared with either amiodarone or placebo (8). More recently, the DANISH (Defibrillator Implantation in Patients with Nonischemic Systolic Heart Failure) trial enrolled patients with nonischemic cardiomyopathy and LVEF ≤35% to ICD or standard care (9). There was no reduction in the primary endpoint of total mortality, but there was a reduction in SCD risk. In the DANISH trial, 58% of patients in each limb received CRT, possibly mitigating the benefit of an ICD.

Economic outcomes of ICD implantation for primary prevention of SCD were assessed in 3 RCTs (MADIT-I [13], MADIT-II [15], and SCD-HeFT [12]), 1 observational study (10), and 3 simulation models (11,14,42), all of which had generally consistent results. All studies reported increased survival and life expectancy and higher lifetime costs of medical care with an ICD than without an ICD. The incremental cost-effectiveness ratios were generally <$60,000 per year of life added by an ICD, which provides high value according to the benchmarks adopted for the current guideline. The value provided by an ICD was consistently high when life expectancy was projected to increase by >1.4 years (14). In contrast, when survival was not increased by ICD implantation, as in the coronary artery bypass graft (CABG) Patch trial (43), the ICD did not provide value, because the higher costs were unaccompanied by a gain in life expectancy (14).

The MADIT-II trial randomized patients with previous MI and LVEF <30%, without any limitation of HF class, to ICDs or not (6). Thirty-seven percent of the patients were in class I congestive heart failure (CHF). Mortality was reduced with an ICD.

Most of the relevant data for the guidelines of CRT in HF come from seminal trials published from 2002 to 2010. The first of these was the MIRACLE (Multicenter InSync Randomized Clinical Evaluation) trial, which took patients with LVEF ≤35%, moderate to severe HF, and QRS duration ≥130 ms (16). There was a benefit in the 6-minute walk test, QOL, functional HF classification, and LVEF. The COMPANION (Comparison of Medical Therapy, Pacing and Defibrillation in Heart Failure) trial, which enrolled NYHA class III to IV patients with QRS ≥120 ms, included 3 arms: GDMT, CRT-D, and CRT pacemaker (CRT-P) (17). The primary endpoint of death or hospitalization was decreased with CRT-P and CRT-D. The CARE-HF (Cardiac Resynchronization Heart Failure) trial included a similar group with NYHA class III to IV, LVEF ≤35%, QRS >120 ms, and showed a significant reduction in primary and endpoint of death or hospitalization (18). In the REVERSE (Resynchronization Reverses Remodeling in Systolic Left Ventricular Dysfunction) trial, patients with NYHA class I to II and LVEF ≤40% were randomized to CRT-D on for 1 year and CRT-D off for 1 year or vice versa (19). A HF composite endpoint was less common when CRT was activated. MADIT-CRT enrolled NYHA class I and II HF with LVEF ≤30% and QRS ≥130 ms and compared CRT-D with ICD (20). The primary endpoint of death or HF was reduced by CRT-D. The RAFT (Resynchronization-Defibrillation for Ambulatory Heart Failure) trial randomized patients with NYHA class II to III HF, LVEF ≤30%, QRS >120 ms, or paced QRS ≥200 ms and compared CRT-D with ICD (21). Again, there was a reduction in the primary endpoint of death or HF hospitalization.

The economic value of CRT has been evaluated by 3 RCTs (COMPANION [22], MADIT-CRT [26], and REVERSE [23]), 2 model-based analyses (25,27), and 1 observational study (24). These analyses consistently found CRT increased survival and QOL in addition to increasing health care costs. However, the economic value of CRT likely varies as a result of the shown variation in treatment effect (26). Among populations with larger expected mortality reduction and improvement in QOL, such as patients with a LBBB with QRS duration >150 ms, the cost per QALY is <$60,000 (22,26,27). Among other populations expected to have smaller treatment benefit, the economic value is more uncertain. However, a model-based analysis of patients with NYHA class I to II found the incremental cost-effectiveness ratio remained <$150,000 per QALY with even small reductions in all-cause mortality (27). Therefore, CRT likely provides at least intermediate value for patients with other guideline-indicated recommendations in which CRT is expected to reduce mortality.

Subgroup analysis of the previously mentioned trials has informed us of the predictors of benefit, including longer QRS duration, and LBBB versus non-LBBB (28). The most benefit was gained with wider QRS durations and with LBBB. This was true in COMPANION, CARE-HF, MADIT-CRT, REVERSE, and RAFT (17,29-32). A QRS duration >150 ms was also a predictor of response, and in those with non-LBBB, a prolonged PR predicted benefit in MADIT-CRT but not in REVERSE (33).

Extension of benefit to those with LVEF between 35% and 50% has been seen. In the BLOCK-HF (Biventricular versus Right Ventricular Pacing in Heart Failure) trial, patients with NYHA class I to III HF, LVEF ≤50%, and atrioventricular block randomized to RV pacing or CRT, there was benefit to CRT in reduction in the primary outcome of death, urgent HF visit, or 15% increase in LV end systolic volume (34).

In the previously mentioned CRT trials, there was some benefit for those with LBBB and QRS durations between 120 and 149, but not as much benefit as those with LBBB ≥150 ms (17,28-32).

Several trials have included patients with AF. In the MUSTIC AF (Multisite Stimulation in Cardiomyopathies) (44), RAFT (45), and the SPARE (Spanish Atrial Fibrillation and Resynchronization) (46) trials, there were benefits in patients with AF, while in COMPANION (47), AF attenuated the benefit of CRT. In the PAVE (Post AV Nodal Ablation Evaluation) study, patients with NYHA class II to III, mean LVEF of 46%, and AF undergoing atrioventricular node ablation, CRT improved the 6-minute walk test and LVEF compared with those who were RV paced (35).

In patients in whom there is an expected high burden of ventricular pacing, especially if >40%, CRT may be used to reduce mortality, reduce hospitalizations, and improve symptoms and QOL (35,48).

Identification of specific arrhythmogenic genetic variants such as LMNA/C, desmosomal proteins, phospholamban, and Filamin-C carry implications for implantation of ICDs for primary prevention of sudden death even in patients who have LVEF >35%, or <3 months of GDMT. Most patients with LMNA/C cardiomyopathy will progress to cardiac transplantation, sometimes precipitated by refractory arrhythmias more than by pump failure (36-38,49).

Subgroup analysis of the CRT RCTs has shown that patients with LVEFs ≤35%, non-LBBB, and QRS duration of 120 to 149 ms and NYHA class III to ambulatory class IV did not derive as much benefit as those with LBBB ≥120 ms (17,28-32).

The MADIT-CRT trial included NYHA class I (and class II) patients with ischemic heart disease, LVEF ≤30%, and QRS >130 ms (39). Patients with nonischemic cardiomyopathy were enrolled if they had NYHA class II HF.

Extension of benefit to patients with narrow QRS has been attempted but has generally failed. In the RETHINQ (Cardiac Resynchronization Therapy in Patients with Heart Failure and Narrow QRS) trial, patients with QRS duration <130 ms were randomized to CRT or not (40). There was no benefit from CRT, but subgroup analysis showed there was a benefit with QRS durations between 120 and 130 ms. In the ECHO-CRT (Echocardiography Guided Cardiac Resynchronization Therapy) trial, patients with NYHA class III to IV HF, LVEF ≤35% and a QRS duration ≤130 ms, and mechanical dysynchrony on echocardiography underwent randomization to CRT (50). There was no benefit to CRT in this trial. And in the LESSER-EARTH (Evaluation of Resynchronization Therapy for Heart Failure) trial, patients with severe LV dysfunction and QRS <120 ms derived no benefit from CRT (51). The NARROW-CRT (Narrow QRS Ischemic Patients Treated With Cardiac Resynchronization Therapy) was the only trial that showed a benefit in a clinical composite score in patients with an indication for an ICD and QRS <120 ms (52).

Subgroup analysis of the CRT trials has shown no benefit for those with LVEF ≤35%, non-LBBB 120 to 149, and NYHA class I-II HF (17,28-32).

The 1-year survival is a standard inclusion for ICD and CRT trials (1-9,16-21).

CABG has been shown to improve outcomes in patients with left main or left main equivalent disease and HF (1,4,10-14). Long-term follow-up shows a reduction in all-cause, cardiovascular, and HF hospitalizations and in all-cause and cardiovascular mortality in patients with LV dysfunction who receive CABG and GDMT compared with GDMT alone (2,7). The long-term survival benefit is greater in those with more advanced ischemic cardiomyopathy (lower EF or 3-vessel disease) and diminishes with increasing age (5,7). CABG also improves QOL compared with GDMT alone (3). An RCT of CABG combined with surgical ventricular remodeling compared with CABG alone did not show a reduction in death or hospitalization, or improvement in symptoms with surgical ventricular remodeling (15). Surgical ventricular remodeling performed at the time of CABG may be useful in patients with intractable HF, large thrombus, or persistent arrhythmias resulting from well-defined aneurysm or scar, if other therapies are ineffective or contraindicated (15,16).

VHD is a significant cause of HF. In patients with HF, management of VHD should be performed by a multidisciplinary team with expertise in HF and VHD, in accordance with the VHD guidelines (15). Cardiologists with expertise in the management of HF are integral to the multidisciplinary team and to guiding the optimization of GDMT in patients with HF and coexisting valve disease. Severe aortic stenosis, aortic regurgitation, MR, and tricuspid regurgitation are associated with adverse outcomes and require timely assessment, optimization of medical therapies, and consideration of surgical or transcatheter interventions accordingly to prevent worsening of HF and other adverse outcomes (1-10,12-20,22-35).

GDMT, including RAAS inhibition, beta blockers, and biventricular pacing, improves MR and LV dimensions in patients with HFrEF and secondary MR, particularly MR that is proportionate to LV dilatation (1-5,12,13,17). In a small RCT, sacubitril-valsartan resulted in a significant reduction in effective regurgitant area and in regurgitant volume when compared with valsartan. The COAPT trial showed a mortality benefit with TEER in patients with severe secondary MR, LVEF between 20% and 50%, LV end-systolic diameter ≤70 mm, PA systolic pressure ≤70 mm Hg, and persistent symptoms (NYHA class II to IV) while on optimal GDMT (28), and these criteria apply when considering TEER. A cardiologist with expertise in the management of HF is integral to shared decision-making for valve intervention and should guide optimization of GDMT to ensure that medical options for HF and secondary MR have been effectively applied for an appropriate time period and exhausted before considering intervention.

EMPEROR-Preserved (Empagliflozin Outcome Trial in Patients with Chronic Heart Failure with Preserved Ejection Fraction) showed a significant benefit of the SGLT2i, empagliflozin, in patients with symptomatic HF, with LVEF >40% and elevated natriuretic peptides (1). The 21% reduction in the primary composite endpoint of time to HF hospitalization or cardiovascular death was driven mostly by a significant 29% reduction in time to HF hospitalization (nonsignificant lower cardiovascular death [HR, 0.91; 95% CI, 0.76–1.0]), with no benefit on all-cause mortality. Empagliflozin also resulted in a significant reduction in total HF hospitalizations, decrease in the slope of the eGFR decline, and a modest improvement in QOL at 52 weeks. Of note, the benefit was similar irrespective of the presence or absence of diabetes at baseline. In a subgroup of 1983 patients with LVEF 41% to 49% in EMPEROR-Preserved, empagliflozin, a SGLT2i, reduced the risk of the primary composite endpoint of cardiovascular death or hospitalization for HF (1). Although the benefit in the primary endpoint did not have a significant interaction by LVEF subgroups (41%–49%, 50%–<60%, and >60%) (1), in a subgroup analysis by EF, there was a signal for lower benefit on the primary composite endpoint, first and recurrent hospitalizations for HF at higher LVEFs >62.5% (10).

Post hoc and subsets of analyses of HFrEF trials that included HFmrEF (LVEF 41%–49%) have suggested benefit from use of GDMT for HFrEF (i.e., beta blockers, ARNi, ACEi or ARB, and spironolactone) (2,3,5-8). The BBmeta-HF (Beta-blockers in Heart Failure Collaborative Group) performed a meta-analysis of 11 HF trials; in a subgroup of 575 patients with LVEF 40% to 49% in sinus rhythm, beta blockers reduced the primary outcome of all-cause and cardiovascular mortality (2). A subgroup analysis of the PARAGON-HF (Prospective Comparison of ARNi with ARB Global Outcomes in HF with Preserved Ejection Fraction) trial for patients with LVEF 45% to 57% (lower range of EFs in the trial) suggested benefit of sacubitril-valsartan versus valsartan alone (rate ratio, 0.78; 95% CI, 0.64-0.95) (3). In a subgroup of 1322 patients with LVEF 41% to 49% in a post hoc analysis of pooled data from the CHARM (Candesartan in Heart failure-Assessment of Reduction in Mortality and morbidity) trials, candesartan reduced risk of cardiovascular death and HF hospitalization, the risk of first HF hospitalization, and the risk of recurrent HF hospitalization (5). In a subgroup of 520 patients with LVEF 44% to 49% in a post hoc analysis of TOPCAT (Treatment of Preserved Cardiac Function Heart Failure With an Aldosterone Antagonist), spironolactone reduced the risk of the primary composite endpoint of cardiovascular death, HF hospitalization, or resuscitated sudden death, which was mostly caused by a reduction in cardiovascular mortality with spironolactone and among patients enrolled in North and South America (6). Spironolactone is preferred among HFmrEF patients with poorly controlled hypertension given previous evidence supporting its use for blood pressure management (1). Continuation of GDMT for patients with improved HFrEF and HFmrEF is important to reduce risk of recrudescent HF (4). Meta-analyses report diverse findings with neurohormonal antagonism in patients with HFmrEF, specifying benefit in certain subgroups, underlining the heterogeneity of this phenotype (2,9). Patients with HFmrEF should have repeat evaluation of LVEF to determine the trajectory of their disease process and should undergo testing as clinically indicated to diagnose conditions warranting disease-specific therapy (e.g., CAD, sarcoidosis, amyloidosis).

In an open-label RCT (1), phased withdrawal of HF medications in patients with previous DCM—who were now asymptomatic, whose LVEF had improved from <40% to ≥50%, whose left ventricular end-diastolic volume (LVEDV) had normalized, and who had an NT-proBNP concentration <250 ng/L—resulted in relapse of cardiomyopathy and HF in 40% of the patients within 6 months. Relapse was defined by at least 1 of these: 1) a reduction in LVEF by >10% and <50%; 2) an increase in LVEDV by >10% and to higher than the normal range; 3) a 2-fold rise in NT-proBNP concentration and to >400 ng/L; or 4) clinical evidence of HF. Treatment was withdrawn successfully in only 50% of patients (1). Secondary analyses showed worsening Kansas City Cardiomyopathy Questionnaire scores, a substantial reduction in LVEF, and nonsignificant increases in NT-proBNP and LV volumes with withdrawal of HF medications.

The role of blood pressure control is well established for the prevention of HF, as well as for reduction of other cardiovascular events and HF mortality in patients without prevalent baseline HF (1-3,21-24). The SPRINT (Systolic Blood Pressure Intervention) trial and meta-analyses established that more intensive blood pressure control in patients with high cardiovascular risk significantly reduces HF and other cardiovascular outcomes (2,3,25). In recent clinical practice guidelines for hypertension, blood pressure targets in HFpEF are extrapolated from those for treatment of patients with hypertension in general (26). However, the optimal blood pressure goal and antihypertensive regimens are not known for patients with HFpEF. RAAS antagonists including ACEi, ARB, MRA, and possibly ARNi, could be first-line agents given experience with their use in HFpEF trials (8,10,16,20,27,28). Beta blockers may be used to treat hypertension in patients with a history of MI (27), symptomatic CAD, or AF with rapid ventricular response. These effects need to be balanced with the potential contribution of chronotropic incompetence to exercise intolerance in some patients (29).

EMPEROR-Preserved (Empagliflozin Outcome Trial in Patients with Chronic Heart Failure with Preserved Ejection Fraction) showed a significant benefit of the SGLT2i, empagliflozin, in symptomatic patients with HF with LVEF >40% and elevated natriuretic peptides (30). The 21% reduction in the primary composite endpoint of time to HF hospitalization or cardiovascular death was driven mostly by a significant 29% reduction in time to HF hospitalization (nonsignificant lower cardiovascular death [HR, 0.91; 95% CI, 0.76-1.0]), with no benefit on all-cause mortality. Empagliflozin also resulted in a significant reduction in total HF hospitalizations, decrease in the slope of the eGFR decline, and a modest improvement in QOL at 52 weeks. Of note, the benefit was similar irrespective of the presence or absence of diabetes at baseline. Although the benefit in the primary endpoint did not have a significant interaction by LVEF subgroups (<50%, 50%–<60%, and >60%) (30), in a subgroup analysis by EF, there was a signal for lower benefit on the primary composite endpoint, first and recurrent HF hospitalizations at higher LVEFs >62.5% (31).

Large, randomized clinical trial data are unavailable to specifically guide therapy in patients with HFpEF and AF. Currently, the comprehensive care of AF can be extrapolated from the clinical practice guidelines for AF, with individualization of strategies for rate or rhythm control in patients with HFpEF (see also Section 10.2, “Management of Atrial Fibrillation (AF) in HF,” for HF specific recommendations for AF). Although beta blockers and nondihydropyridine calcium channel blockers are often considered as first-line agents for heart rate control in patients with HFpEF, a recent smaller open-label trial, RATE-AF in elderly patients with AF and symptoms of HF (most with preserved LVEF), compared the use of the beta blocker, bisoprolol, to digoxin (32). At 6 months, the primary endpoint of QOL was similar between the 2 groups. However, several secondary QOL endpoints, functional capacity, and reduction in NT-proBNP favored digoxin at 12 months. There was a similar heart rate reduction in both groups. Of note, more adverse events such as higher rates of dizziness, lethargy, and hypotension occurred with beta blockers than digoxin. The comprehensive care of AF is beyond the scope of these guidelines. AF-specific care recommendations can be found in separate ACC/AHA clinical practice guidelines (33,34).

MRAs improve diastolic function in patients with HFpEF (35). The TOPCAT trial investigated the effects of spironolactone in patients with HFpEF. The small reduction (HR, 0.89) in the composite of death, aborted cardiac death, and HF hospitalization was not statistically significant, although HF hospitalization was reduced (HR, 0.83); adverse effects of hyperkalemia and increasing creatinine levels were more common in the treatment group (5). A post hoc analysis (6) showed efficacy in the Americas (HR 0.83) but not in Russia-Georgia (HR 1.10). A sample of the Russia-Georgia population in the active treatment arm had nondetectable levels of a spironolactone metabolite. Post hoc analyses have limitations, but they suggest a possibility of benefit in appropriately selected patients with symptomatic HFpEF (LVEF ≥45%, elevated BNP level or HF admission within 1 year, eGFR >30 mL/min/1.73 m², creatinine <2.5 mg/dL, and potassium <5.0 mEq/L). Furthermore, another post hoc analysis suggested that the potential efficacy of spironolactone was greatest at the lower end of the LVEF spectrum (7). Careful monitoring of potassium, renal function, and diuretic dosing at initiation and follow-up are key to minimizing the risk of hyperkalemia and worsening renal function.

Although RAAS inhibition strategies have been successful in the treatment of HFrEF, and RAAS activation is suggested in HFpEF (36,37), clinical trials with RAAS inhibition have not showed much benefit in patients HFpEF. In the CHARM-Preserved (Candesartan in patients with chronic HF and preserved left-ventricular ejection fraction) trial, patients with LVEF >40% were randomized to an ARB, candesartan, or to placebo (38). The primary endpoint (cardiovascular death or HF hospitalization) was not significantly different between the 2 groups (HR, 0.89; 95% CI, 0.77–1.03, P=0.118; covariate-adjusted HR, 0.86; P=0.051). Cardiovascular mortality was identical in the 2 groups; HF hospitalizations were lower in the candesartan arm, with borderline statistical significance on the covariate-adjusted analysis only (HR, 0.84; 95% CI, 0.70–1.00; P=0.047; unadjusted P=0.072). The number of individuals hospitalized for HF (reported by the investigator) was lower in the candesartan group than placebo (230 versus 279; P=0.017). A post hoc analysis of the CHARM trials showed that improvement in outcomes with candesartan was greater at the lower end the LVEF spectrum (39). In a meta-analysis of 7694 patients with HFpEF in 4 trials evaluating ARB, there was no signal for benefit on cardiovascular mortality (HR, 1.02), all-cause mortality (HR, 1.02), or HF hospitalization (HR, 0.92; 95% CI, 0.83–1.02) (40,41).

In the PARAMOUNT-HF (Prospective Comparison of ARNi With ARB on Management of Heart Failure With Preserved Ejection Fraction) trial, a phase II RCT in patients with HFpEF (LVEF ≥45%), sacubitril-valsartan resulted in a lower level of NT-proBNP after 12 weeks of treatment compared with the ARB, valsartan (42). In the PARAGON-HF (Prospective Comparison of Angiotensin Receptor Neprilysin Inhibitor With Angiotensin Receptor Blocker Global Outcomes in Heart Failure and Preserved Left Ventricular Ejection Fraction) trial, in 4822 patients with HFpEF (LVEF ≥45%, HF admission within 9 months or elevated natriuretic peptide levels, and eGFR ≥30 mL/min/m²), sacubitril-valsartan compared with valsartan did not achieve a significant reduction in the primary composite endpoint of cardiovascular death or total (first and recurrent) HF hospitalizations (rate ratio, 0.87; 95% CI, 0.75-1.01; P=0.06) (10). Given the primary outcome was not met, other analyses are exploratory. There was no benefit of sacubitril-valsartan on cardiovascular death (HR, 0.95) or total mortality (HR, 0.97). There was a signal of benefit for the ARNi for HF hospitalizations (rate ratio, 0.85; 95% CI, 0.72–1.00; P=0.056). The occurrence of hyperkalemia and the composite outcome of decline in renal function favored sacubitril-valsartan, but it was associated with a higher incidence of hypotension and angioedema. In prespecified subgroup analyses, a differential effect by LVEF and sex was noted. A benefit of sacubitril-valsartan compared with valsartan was observed in patients with LVEF below the median (45%–57%; rate ratio, 0.78; 95% CI, 0.64–0.95), and in women (rate ratio, 0.73; 95% CI, 0.59–0.90) (10,43,44).

Nitrate therapy can reduce pulmonary congestion and improve exercise tolerance in patients with HFrEF. However, the NEAT-HFpEF (Nitrate’s Effect on Activity Tolerance in Heart Failure With Preserved Ejection Fraction) trial (45) randomized 110 patients with EF ≥50% on stable HF therapy, not including nitrates, and with activity limited by dyspnea, fatigue, or chest pain, to either isosorbide mononitrate or placebo and found no beneficial effects on activity levels, QOL, exercise tolerance, or NT-proBNP levels. Although the routine use of nitrates in patients with HFpEF does not appear beneficial, patients with HFpEF and symptomatic CAD may still receive symptomatic relief with nitrates. Phosphodiesterase-5 inhibition augments the nitric oxide system by upregulating cGMP activity. The RELAX (Phosphodiesterase-5 Inhibition to Improve Clinical Status and Exercise Capacity in Heart Failure with Preserved Ejection Fraction) trial (13) randomized 216 patients with EF ≥50% on stable HF therapy and with reduced exercise tolerance (peak observed VO₂, <60% of predicted) to phosphodiesterase-5 inhibition with sildenafil or placebo. This study did not show improvement in oxygen consumption or exercise tolerance.

Diagnosis of ATTR-CM requires a high index of suspicion. LV thickening (wall thickness ≥14 mm) along with fatigue, dyspnea, or edema should trigger consideration of ATTR-CM, especially with discordance between wall thickness on echocardiogram and QRS voltage on ECG (10), or other findings such as apical sparing of LV longitudinal strain impairment on echocardiography and diffuse late-gadolinium enhancement on cardiac MRI. ATTR-CM is prevalent in severe aortic stenosis (1), HFpEF (2), carpal tunnel syndrome (3), lumbar spinal stenosis (4), and autonomic or sensory polyneuropathy (5). Practically, screening for the presence of a monoclonal light chain and technetium pyrophosphate (^99mTc-PYP) scan can be ordered at the same time for convenience, but the results of the ^99mTc-PYP scan are interpreted only on the context of a negative monoclonal light chain screen. ^99mTc-PYP scans may be positive even in AL amyloidosis (7) and, thus, a bone scintigraphy scan alone, without concomitant testing for light chains, cannot distinguish ATTR-CM from AL-CM. Serum free light chain (FLC) concentration and serum and urine immunofixation electrophoresis (IFE) are assessed to rule out AL-CM. IFE is preferred because serum plasma electrophoresis and urine plasma electrophoresis are less sensitive. Together, measurement of serum IFE, urine IFE, and serum FLC is >99% sensitive for AL amyloidosis (6,11).

The use of ^99mTc bone-avid compounds for bone scintigraphy allows for noninvasive diagnosis of ATTR-CM (7). ^99mTc compounds include PYP, 3,3-diphosphono-1,2-propanodicarboxylic acid, and hydromethylene diphosphonate, and PYP is used in the United States. In the absence of a light-chain abnormality, the ^99mTc-PYP scan is diagnostic of ATTR-CM if there is grade 2/3 cardiac uptake or an H/CL ratio of >1.5. In fact, the presence of grade 2/3 cardiac uptake in the absence of a monoclonal protein in serum or urine has a very high specificity and positive predictive value for ATTR-CM (7). SPECT is assessed in all positive scans to confirm that uptake represents myocardial retention of the tracer and not blood pool or rib uptake signal (12).

If ATTR-CM is identified, then genetic sequencing of the TTR gene will determine if the patient has a pathological variant (ATTRv) or wild-type (ATTRwt) disease (12). Differentiating ATTRv from ATTRwt is important because confirmation of ATTRv would trigger genetic counseling and potential screening of family members and therapies, inotersen and patisiran, which are presently approved only for ATTRv with polyneuropathy (13,14).

Tafamidis is currently the only therapy to improve cardiovascular outcomes in ATTR-CM (1). Tafamidis binds the thyroxin-binding site of TTR. In the ATTR-ACT (Safety and Efficacy of Tafamidis in Patients With Transthyretin Cardiomyopathy) clinical trial, a randomized trial of patients with ATTRwt-CM or ATTRv-CM and NYHA class I to III symptoms, tafamidis had lower all-cause mortality (29.5% versus 42.9%) and lower cardiovascular-related hospitalization (0.48 versus 0.70 per year) after 30 months (1). There was a higher rate of cardiovascular-related hospitalizations in patients with NYHA class III HF, potentially attributable to longer survival during a more severe period of disease. Given that tafamidis prevents but does not reverse amyloid deposition, tafamidis is expected to have greater benefit when administered early in the disease course. As the survival curves separate after 18 months, patients for whom noncardiac disease is not expected to limit survival should be selected. Benefit has not been observed in patients with class IV symptoms, severe aortic stenosis, or impaired renal function (eGFR <25 mL·min⁻¹·1.73 m⁻² body surface area). Tafamidis is available in 2 formulations: tafamidis meglumine is available in 20-mg capsules; and the FDA-approved dose is 80 mg (4 capsules) once daily. Tafamidis is also available in 61-mg capsules; the FDA-approved dose for this new formulation is 61 mg once daily.

One model-based analyses used the results of the ATTR-ACT study (1) to evaluate the cost-effectiveness of chronic tafamidis compared with no amyloidosis-specific therapy among patients with wild-type or variant transthyretin amyloidosis and NYHA class I to III HF (2). With assumptions that tafamidis remained effective beyond the clinical trial duration, they estimated tafamidis increased average survival by 1.97 years and QALY by 1.29. Despite these large clinical benefits, tafamidis (with an annual cost of $225,000) had an incremental cost-effectiveness ratio >$180,000 per QALY gained, the benchmark used by this guideline for low value. The cost of tafamidis would need to decrease by approximately 80% for it to be intermediate value with a cost per QALY <$180,000.

Intracardiac thrombosis occurs in approximately one-third of patients with cardiac amyloidosis, in some cases in the absence of diagnosed AF (3,4,12) and regardless of CHA₂DS₂-VASc score (13). The use of anticoagulation reduced the risk of intracardiac thrombi in a retrospective study (4). The choice of direct oral anticoagulants (DOAC) versus warfarin has not been studied in patients with ATTR, nor has the role of left atrial appendage closure devices. The risk of anticoagulation on bleeding risk in patients with ATTR-CM and AF has not been established. However, although patients with AL amyloidosis may have acquired hemostatic abnormalities, including coagulation factor deficiencies, hyperfibrinolysis, and platelet dysfunction, TTR amyloidosis is not associated with hemostatic defects.

Clinical indicators of advanced HF that should trigger possible referral to an advanced HF specialist are shown in Table 18 (1,2,12-14). Timely referral for review and consideration of advanced HF therapies is crucial to achieve optimal patient outcomes (15-17). Acronyms such as I-Need-Help…

• I, Intravenous inotropes

• N, New York Heart Association (NYHA) class IIIB to IV or persistently elevated natriuretic peptides

• E, End-organ dysfunction

• E, EF ≤35%

• D, Defibrillator shocks

• H, Hospitalizations >1

• E, Edema despite escalating diuretics

• L, Low systolic BP ≤90, high heart rate

• P, Prognostic medication; progressive intolerance or down-titration of GDMT

In a registry study of hyponatremia in acute decompensated HF, fluid restriction only improved hyponatremia marginally (1). A registered dietitian-guided fluid and sodium restriction intervention improved NYHA functional classification and leg edema in patients with HFrEF who were not in stage D HF (2), and fluid restriction improved QOL in a pilot RCT of patients with HFrEF and HFpEF (NYHA class I to IV) (3). In a meta-analysis of RCTs on fluid restriction in HF in general, restricted fluid intake compared with free fluid consumption did not result in reduced hospitalization or mortality rates, changes in thirst, the duration of intravenous diuretic use, serum creatinine, or serum sodium levels (4). The validity of a previous trial supporting clinical benefits of fluid restriction in HF is in serious question (12).

More prolonged use of inotropes as “bridge” therapy for those awaiting either heart transplantation or MCS may have benefit in reducing pulmonary hypertension and maintaining end-organ perfusion beyond initial stabilization of patients (1-4).

The use of inotropes for palliation does carry with it risks for arrhythmias and catheter-related infections, although the presence of an ICD does decrease the mortality associated with arrhythmias. This risk should be shared with patients if there is planned use of inotropes in a patient without an ICD, or in whom the preference is to deactivate the ICD for palliative purposes. The rate of inappropriate shocks for sinus tachycardia is relatively low, and the concomitant use of beta blockers may help in these patients. Patients may elect to have their shocking devices deactivated, especially if they receive numerous shocks (14,15).

With the currently available inotropic agents, the benefit of hemodynamic support and stabilization may be compromised by increased myocardial oxygen demand and increased arrhythmic burden. As newer agents are developed, more options may not have these known risks. There are investigational inotropic agents that may provide more options for the management of patients with HF and represent different classes of agents (16).

Durable LVADs should be considered in selected patients with NHYA class IV symptoms who are deemed dependent on IV inotropes or temporary MCS. The magnitude of the survival benefit for durable LVAD support in advanced NYHA class IV patients has progressively improved, with a 2-year survival >80% in recent trials with newer generation LVADs, which approaches the early survival after cardiac transplantation (2). The 2020 INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support) report showed that 87.6% of recent durable LVAD recipients were categorized as INTERMACS 1 to 3 before their implant surgery (10). It also showed improved mean survival, >4 years for the destination LVAD cohort, and >5 years for bridge-to-transplant patients. Durable LVAD support has also achieved impressive functional improvement and QOL improvement in multiple trials (2,7,31), although patients remain tethered to external electrical power supplies via a percutaneous lead can limit this improvement. Most patients require rehospitalization within the first year post-implant. These factors emphasize the need for a thorough evaluation and patient education before the decision to proceed with the treatment. Appropriate patient selection benefits from review by a multidisciplinary team that typically includes an HF cardiologist, surgeon, social worker, nurse, pharmacist, dietician, and a palliative medicine specialist.

Durable MCS should be considered in patients with NHYA class IV symptoms despite optimal medical therapy or those deemed dependent on IV inotropes. Destination therapy MCS provides considerable survival advantage in addition to improvement in functional status and health-related QOL (1,7,12,32,33). There is no clear 1-risk model to assess patient risk for complications, but factors such as elevated central venous pressure, pulmonary hypertension, and coagulopathy have been linked to poorer outcomes (15,34-36). In patients who are unable to tolerate anticoagulation after repeated challenges, implantation of a durable MCS is associated with excess morbidity; incidents of pump thrombosis, hemolysis, and ischemic neurologic events have been linked to subtherapeutic international normalized ratios (37-41). In addition, implantation of MCS in patients with INTERMACS profile of 1 has been associated with poorer outcome, while those ambulatory patients with profiles 5 to 7 might be too well to have large significant benefit, depending on their symptom burden (19). For patients who are initially considered to be transplant ineligible because of pulmonary hypertension, obesity, overall frailty, or other reasons, MCS can provide time to reverse or modify these conditions (35,42-44). Continuing and uptitrating GDMT in patients with durable MCS is recommended (45).

Multiple studies evaluated the cost-effectiveness of ventricular assist device implantation for advanced HF between 2012 and 2017 (20,21,23). They consistently found device implantation was of low economic value, with incremental cost-effectiveness ratios of $200,000 per QALY gained compared with medical therapy alone among patients who potentially underwent subsequent heart transplant and those who were ineligible for heart transplant. In these studies, costs after implantation remained high given high rates of complication and rehospitalization. However, these studies used earlier estimates of post-implant outcomes and complication-related costs that have generally improved over time with better care and newer devices (46-48). Additionally, limited recent data suggest improvement in health care costs and intermediate economic value with LVAD among patients with advanced HF who are either eligible or ineligible for subsequent heart transplant (22,24). The improvement may result from lower complication rates, increased survival, lower implant costs, and higher estimated QOL. However, given the conflicting data and limited analyses of contemporary data, the current value of LVAD therapy is uncertain.

Temporary MCS can help stabilize patients and allow time for decisions about the appropriateness of transitions to definitive management, such as durable MCS as a bridge or destination therapy, stabilization until cardiac transplantation or, in the case of improvement and recovery, suitability for device removal (45). These patients often present in cardiogenic shock that cannot be managed solely with IV inotropes and in whom other organ function is at risk. Temporary MCS is also appropriate for use to allow patients to engage in decision-making for durable MCS or transplantation and for determination of recovery of neurologic status.

Cardiac transplantation is the established treatment for eligible patients with stage D HF refractory to GDMT, device, and surgical optimization. The survival of adult recipients who received a transplantation between 2011 and 2013 at 1, 3, and 5 years is 90.3%, 84.7%, and 79.6%, respectively (2). Conditional survival now approaches 15 years (1). Cardiac transplantation also improves functional status and health-related QOL (12-15). Good outcomes can be achieved in patients not only with HF that is primarily cardiovascular in origin, including reversible pulmonary hypertension (16), congenital heart disease (17), and hypertrophic cardiomyopathy (18), but also in patients with systemic conditions complicated by HF, such as muscular dystrophy (19), sarcoidosis (20), and amyloidosis (21). CPET can refine candidate prognosis and selection (22-28). Appropriate patient selection should include integration of comorbidity burden, caretaker status, and goals of care. The listing criteria, evaluation, and management of patients undergoing cardiac transplantation are described by the International Society for Heart and Lung Transplantation (29). The United Network of Organ Sharing Heart Transplant Allocation Policy was revised in 2018 with a broader geographic sharing policy and a 6-tiered system to better prioritize more unstable patients and minimize waitlist mortality (5-11).