Heart Failure

Heart Failure



I. Heart failure is diagnosed clinically, not by echocardiography

HF is defined as a clinical combination of congestive findings and low output findings, corroborated by objective evidence of cardiac dysfunction (elevated BNP, echo evidence of systolic or diastolic dysfunction, or elevated LA pressure).

A. Congestive findings

  • Orthopnea and paroxysmal nocturnal dyspnea (PND), both relatively specific for HF. They result from the increase in venous return during recumbency and the subsequent increase in pulmonary capillary pressure. Many patients present with nocturnal cough or wheezes rather than nocturnal dyspnea; a dry cough during recumbency may be an orthopnea equivalent and should suggest HF. Wheezes result from congestion of the bronchial mucosa and the small airways.

    Orthopnea can be tested by asking the patient to lie supine for 1 to 2 minutes. Beside HF, orthopnea may be seen in obese patients or those with advanced lung disease and flattened diaphragm. Beside HF, PND may be seen in lung disease; whereas HF-related PND is relieved by upright posture, COPD-related PND results from mucus hypersecretion and is relieved by cough and albuterol, and asthma-related PND results from nocturnal bronchospasm and is relieved by albuterol.

  • Exertional dyspnea. This may be a manifestation of a rise in pulmonary capillary pressure during exertion, even without overt pulmonary edema and without significant hypoxemia. The pulmonary venous engorgement stiffens the lungs and reduces vital capacity, leading to dyspnea. Dyspnea may also be a manifestation of an inappropriate rise in cardiac output during exertion, with a subsequent peripheral and respiratory muscle fatigue and reduced pulmonary perfusion (increased pulmonary dead space). Thus, outside florid edema, diuretics do not necessarily improve exertional dyspnea. Dyspnea is frequently described as “chest pressure” and therefore, in HF, especially decompensated HF, chest pressure is not necessarily CAD.
  • Bendopnea, i.e. dyspnea within 15 sec of bending forward (e.g., dyspnea when putting on shoes), is highly specific for a baseline elevation of PCWP or RA pressure, with a further rise upon bending, particularly in the setting of low cardiac output.
  • Quick weight gain or quick weight loss in response to treatment implies volume overload.
  • Crackles/pulmonary edema/pleural effusions. Note that crackles are frequently (80%) absent in patients with decompensated chronic HF, even when pulmonary capillary pressure is severely elevated. The increased lymphatic drainage of alveolar fluid prevents alveolar pulmonary edema.1
  • Increased JVP ≥ 8 cmH2O. JVP exam assesses the height of the internal, or sometimes external, jugular vein pulsations. In patients with a normal JVP, a hepatojugular reflux maneuver may be performed to unveil HF: a positive result is defined as a sustained rise of JVP of ≥ 3 cmH2O following > 10 seconds of pressure on the right upper quadrant during normal breathing (no Valsalva), or a fall of JVP of 4 cmH2O upon release of pressure.24

    JVD means distension of the external jugular vein, which is the visible jugular vein, in an upright position, and usually implies an elevated JVP. When JVD is present, JVP may be assessed on the external jugular vein.

  • S3: in patients > 40 years old, S3 is highly specific (~90%) for an elevated PCWP, but insensitive.5 A loud P2 implies significant pulmonary hypertension, suggestive of left HF in the right context. Normally, P2 is soft and is only heard at the left upper sternal border, where it is split from A2; a loud P2 translates into a diffusely loud and sometimes split S2. S4 is less specific for an elevated PCWP and may be seen with compensated LV dysfunction.
  • Peripheral edema ± ascites; congestive and sometimes painful hepatomegaly (pulsatile hepatomegaly if severe TR); and rarely, congestive splenomegaly in prolonged failure. Over 4 liters of volume overload are required to see peripheral edema. Therefore, peripheral edema is an insensitive finding and may be absent in 60% of patients with elevated PCWP; moreover, in many cases, pulmonary edema results from volume redistribution to the lungs rather than florid volume overload.2,6 Edema is, however, specific for HF in a dyspneic patient.

  • Congestion on X-ray: similarly to the limitations of the pulmonary exam in chronic HF, the increased lymphatic drainage prevents the appearance of alveolar or even interstitial pulmonary edema in the majority of patients with chronically elevated PCWP. Pulmonary vas- cular cephalization, pleural effusion, and perihilar engorgement are the most sensitive findings, but overall, X-ray may not show any congestive finding in ~40% of patients with elevated PCWP.1,2 CT scan is more helpful: interlobular septal thickening is a marker of interstitial pulmonary edema and is highly sensitive and specific for elevated PCWP.
  • B-lines on lung ultrasound: the lung fields may be interrogated with an ultrasound probe placed between the ribs. Lung fluid manifests as white vertical lines (like the beam of a flashlight); this is highly sensitive and specific for pulmonary edema, more than chest X-ray, mainly when ≥ 3 lines are seen per view, bilaterally. ~4 to 6 regions are interrogated per lung (upper and lower lungs). The higher the number of lines and the more they are seen in the upper lungs, the more severe the pulmonary edema. Normal lung contains much air and little water, so no reflection of the ultrasonographic beams occurs and normally no B-line artifact appears.

In general, one to two congestive signs along with at least one congestive symptom are required for the clinical diagnosis of HF. The most reliable congestive findings are orthopnea, elevated JVP, S3, and a recent, quick weight gain. In particular, elevated JVP at rest or with hepatojugular reflux is the most sensitive and specific finding (>80%) for the diagnosis of elevated right- but also left-sided filling pressures (PCWP).24

B. Low-output findings (also known as “cold” signs) correlate with a more advanced HF stage:

  • Severe fatigue.
  • Since the pulse pressure corresponds to stroke volume, a narrow pulse pressure (<25% SBP) or a borderline systolic blood pressure implies a low stroke volume. In fact, a narrow pulse pressure is the physical finding that most reliably predicts a low cardiac output (>85% sensitivity and specificity).1 Occasionally, severe arterial noncompliance prevents pulse pressure from narrowing.
  • Compensatory tachycardia.
  • Pulsus alternans, which refers to an every-other-beat variation in pulse intensity. This is different from pulsus paradoxus, which is a more gradual variation in pulse intensity, seen in tamponade.
  • Cold, clammy extremities.
  • Renal failure, hyponatremia, poor response to diuretics.
  • At an advanced stage: impaired mentation, drowsiness, central hypoventilation with Cheyne–Stokes pattern (hyperpnea alternating with hypopnea/apnea).
  • Abdominal pain may result from functional bowel ischemia or from liver distension.

The functional capacity of HF patients is classified into four classes (NYHA class I: no limitation, can jog or carry > 24 lb up a flight of stairs; class II: symptoms occur with heavy weight carrying or walking two blocks or two flights of stairs; IIIA: symptoms with walking one block; IIIB symptoms with mild activities, such as dressing, showering, short walking; IV: symptoms at rest). This functional classification applies to patients at their most stable cardiac status, outside of HF exacerbations. Frequent HF hospitalizations, however, usually imply a worse functional class and a poor prognosis.

II. After HF is defined clinically, echocardiography is used to differentiate the three major types of HF

A. HF secondary to LV systolic dysfunction, where EF is reduced ≤40%

In order to improve the stroke volume, the LV cavity dilates, but may be of normal size in acute systolic HF, resulting in more hemodynamic compromise (acute MI, acute valvular regurgitation, acute myocarditis).

B. HF secondary to LV diastolic dysfunction, where EF is normal (≥50%), sometimes supranormal, and LV is generally of normal or small size

LV is non-compliant, i.e., its diastolic pressure rises even at a normal preload volume. Diastolic HF is also called HF with preserved EF (HFpEF) and constitutes 40–50% of all HF presentations. HFpEF is diagnosed when the following four features are present (ESC) (Figure 4.1):710

  1. Clinical HF
  2. Normal EF and normal or only mildly increased LV volume
  3. No dynamic MR; and no transient, ischemic systolic dysfunction (on stress echo or echo within 72 hours of decompensation)11
  4. Invasive, echocardiographic, or BNP evidence of elevated LA pressure or diastolic dysfunction

HFpEF may be difficult to diagnose in dyspneic patients with no overt congestion. A combination of BNP and echo features is suggestive of the diagnosis but is insensitive and may miss HFpEF in 20-40% of the cases.12 The gold standard for diagnosis is elevated mean PCWP at rest (>15 mmHg) or with exercise (≥25 mmHg), and right heart catheterization with exercise may be considered in select cases. Alternatively, the diagnosis is established via exertional echo measurement of E/E’ and PA pressure, or assumed on clinical ground, such as the H2FPEF score (Figure 4.2).10,12,13

Patients with HF and EF of 40 to 50% constitute a specific group wherein the LV pressure–volume relationship may be as steep as in diastolic failure, and wherein the diastolic function is more severely impaired than systolic function. HF with EF 40–50% is predominantly a diastolic failure with mild systolic dysfunction: it behaves like HFpEF but its etiologies are closer to systolic HF, hence it is better called HF with mildly reduced EF (HFmEF).8


Figure 4.1 Diagnosis of HFpEF by catheterization, echo, or BNP features, according to ESC criteria.7 A normal or mildly dilated LV is defined as LV end- diastolic volume < 75 ml/m2 or 96 ml/m2, respectively.


Figure 4.2 Diagnosis of HFpEF in subtle cases of dyspnea with no overt congestion.810, 12, 13

aThe 4 major diastolic markers on echo are (ASE echo guidelines):

  • -Septal E’<7 cm/s or lateral E’<10 cm/s
  • -E/E’>14
  • -LA volume index>34 ml/m2
  • -TR velocity >2.8 m/s (PA pressure >35 mmHg)
  • -Consider the additional use of severe LVH, as per ESC (LV mass ≥149 g/m2 in men, 122 g/m2 in women)9

bH2FPEF score (a total score of 6 is >90% predictive of HFpEF; a score of 2-5, especially 3-5, suggests possible HFpEF, with a probability of 30-70%):

  • H (Heavy, BMI>30): score of 2 // H (Hypertensive requiring ≥2 drugs): score of 1// F (atrial Fibrillation): score of 3
  • P (Pulmonary hypertension, systolic PA pressure>35 mmHg): score of 1 // E (Elderly>60): score of 1
  • F (Filling pressure with E/E’>9 on echo): score of 1

Patients with EF<40% usually have a variable degree of impairment of diastolic filling and diastolic properties. LV systolic dysfunction is, in fact, a combined systolic and diastolic dysfunction. However, by definition, to use the term diastolic HF, EF must be preserved.

C.HF due to severe valvular disease, in which the ventricular function is initially normal. Ultimately, with most left valvular disorders except MS, left ventricular systolic and diastolic dysfunction results.

III. Two additional types of HF

A.High-output heart failure, particularly obesity, which overlaps with the obese subtype of HFpEF

High-output HF is characterized by elevated left ± right filling pressures, yet a cardiac index that is at the upper limit of normal or elevated (>3.5-4 l/min/m2) (e.g., morbid obesity, anemia, lung disease). EF is often normal but may be mildly reduced. See Chapter 5, Section 3.IX.

B. Predominant or isolated right heart failure

Left HF is often (~55–80%) associated with a secondary right systolic or diastolic failure, manifested as increased RA pressure. In fact, left HF is the most common cause of right HF; back-up pressure overload from the left side and loss of the septal contribution to RV contraction lead to biventricular failure. Therefore, an elevated JVP or an abnormal hepatojugular reflux suggests not only right but also left HF, and predicts an elevated PCWP with a sensitivity of 55–80% and a positive predictive value of 85–95%.14

In some patients, the right HF becomes clinically more prominent than left HF as the lungs become underfilled, attenuating the pul- monary edema. Also, in non-ischemic cardiomyopathy, the RV may be directly involved by the myocardial process and may, in some cases, be more affected than the LV.

A truly isolated RV failure may be seen with ASD shunt, lung disease, or pulmonary vascular disease. It is clinically characterized by peripheral edema and elevated JVP yet clear lungs, and invasively, an elevated RA pressure with either a normal PCWP or a mildly increased PCWP that is equalized to the RA pressure. Pleural effusions may be seen with isolated right HF as the parietal pleura drains in the venae cavae (visceral pleura drains in the pulmonary veins).


I. Systolic HF or HF with reduced EF (HFrEF)


CAD is the most common cause of systolic HF. Systolic HF that is secondary to CAD is called “ischemic cardiomyopathy.” Two ischemic processes may explain LV dysfunction:

  1. Large transmural Q-wave MI.
  2. Hibernating myocardium. Severe coronary artery stenosis with chronic ischemia at rest or recurrent exertional ischemia may cause the myocardium to “shut down,” i.e., hibernate without dying. Hibernation can be reversed with revascularization because the tissue usually remains viable.

LV dysfunction associated with single-vessel angiographic CAD, not involving the left main or the proximal LAD and without evidence of prior infarction, is disproportionate to the severity of CAD and implies a mortality that is similar to non-ischemic cardiomyopathy; this LV dysfunction is better classified as non-ischemic cardiomyopathy.14 Ischemic cardiomyopathy, which has ~1.5-fold higher mortality than non-ischemic cardiomyopathy, requires one of the following: (1) history of a large MI; (2) ≥ two-vessel CAD; or (3) single-vessel CAD that involves the left main or the proximal LAD. More extensive CAD is associated with further increase in mortality.14 Some patients have mixed cardiomyopathies (e.g., LV dysfunction secondary to two-vessel CAD and exaggerated by HTN).

B. Hypertension

Hypertension is the second most common cause of systolic HF. Chronic, severe hypertension leads to diastolic dysfunction initially, followed by systolic dysfunction. Acute blood pressure rise does not usually cause acute LV systolic dysfunction, unless myocardial disease or chronic hypertensive cardiomyopathy is present, as a normal LV tolerates acute pressure overload. Acute blood pressure rise more readily causes pulmonary edema through acute diastolic dysfunction.11

C. Advanced valvular heart disease (MR, AI, AS)

D. Dilated cardiomyopathy (DCM)

  1. DCM is characterized by LV ± RV systolic dysfunction that is not due to ischemia, uncontrolled HTN, or valvular disease. The LV dysfunc- tion is often global, but segmental wall motion abnormalities may be seen and do not necessarily imply ischemic cardiomyopathy.
  2. Causes (see Chapter 5, Section 1):

    • Viral myocarditis. There are several forms of myocarditis: (i) subclinical myocarditis, usually self-limited, although a minority of these cases progress to chronic cardiomyopathy; (ii) mild clinical myocarditis with EF 45–50%, usually recovers within weeks or months; (iii) severe myocarditis, i.e., myocarditis manifesting as HF or significant LV dysfunction, fully reverses in one-third of the cases, improves in another 40%, but may rapidly progress in a minority.16
    • Idiopathic. Viral and idiopathic DCMs are the most common forms of DCM.
    • Toxic: long-standing alcohol abuse with or without thiamine deficiency (reversible); cocaine use; radiation therapy; certain chemo- therapy agents, such as doxorubicin (irreversible), cyclophosphamide and trastuzumab (reversible).
    • Genetic inheritance in 30% of patients, mainly autosomal dominant.
    • HIV cardiomyopathy.
    • Metabolic: hyper- or hypothyroidism.
    • Infiltrative: sarcoidosis, hemochromatosis.
    • Tachycardia-mediated cardiomyopathy.
    • Peripartum cardiomyopathy develops in the last month of pregnancy or within 5 months of delivery, without any identifiable cause, and without pre-existing heart disease. 50–60% of the cases are reversible within 6 months.
    • Takotsubo cardiomyopathy.

The worst prognosis is seen with the following three cardiomyopathies: HIV cardiomyopathy, amyloidosis, and doxorubicin- associated cardiomyopathy. 15

  1. Some DCMs may be reversible:

    • Myocarditis.
    • Alcoholic cardiomyopathy reverses if the alcohol abuse is stopped at an early stage ± thiamine supplemented.
    • Hypertensive cardiomyopathy.
    • Peripartum cardiomyopathy.
    • Tachycardia-mediated cardiomyopathy.
    • Takotsubo cardiomyopathy.
    • Other stress-related cardiomyopathies: sepsis or critical illness-associated cardiomyopathy, and neurogenic cardiomyopathy (following hemorrhagic or ischemic stroke).
    • Cardiomyopathy related to thyroid disorders.

On average, ~35% of recent-onset DCMs (<6 months) that are idiopathic or secondary to myocarditis resolve spontaneously or with medical therapy within 6 months (HF-recovered EF, also called HF-improved EF, defined as EF improvement of ≥10% to >40%). Another 40% have a significant improvement of EF.16 This rate is higher for hypertensive cardiomyopathy and lower for genetic cardiomyopathy. Until they recover, these patients have an increased risk of arrhythmias and sudden death, well described with severe myocarditis. HF medical therapy should be continued over the long term, even after LV function normalizes.

II. HF with preserved EF (HFpEF)

A. Hypertension or obesity with or without LV hypertrophy

Hypertension and obesity are the most common risk factors for diastolic HF. LV hypertrophy is common in diastolic HF; however, LV mass is normal in up to 50% of patients, and LV thickness is normal in up to 25% of patients.19 Diastolic dysfunction may precede LV hypertrophy and is more prevalent than LV hypertrophy.

Arterial, ventricular, and atrial stiffness are increased as a result of increasing collagen, cytoskeletal proteins, and abnormal calcium homeostasis. This stiffness makes the LV filling pressure, LA pressure, and SBP markedly increase with relatively minor volume overload. The arterial stiffness explains a very striking rise in SBP with exercise, which is an important component of diastolic HF in many patients.

Other risk factors for diastolic HF include age (>65 years), diabetes (30–50% of diastolic HF patients have diabetes), female sex, renal failure, and AF.

The following mechanisms are incriminated (see also Chapter 5):

  1. Abnormal diastolic function: (i) LV does not relax well (reduced active relaxation or tau index), and (ii) LV is not “elastic” enough to further distend after relaxation and accept the diastolic filling (increased diastolic stiffness or beta index). This results in a small and stiff LV cavity with limited preload reserve and, consequently, limited stroke volume reserve and backward PCWP rise.
  2. Other mechanisms are found in many patients with HFpEF. Thus, HFpEF does not always equate with diastolic HF:

    • AF, which directly reduces LV filling. Also, stiff LA syndrome, even without AF, may raise LA pressure.
    • Impaired contractility. A normal EF does not necessarily imply normal contractility. EF being equal to stroke volume divided by end-diastolic volume, EF is affected by changes in loading conditions (smaller LV–> higher EF).
    • Volume overload conditions, which stretch a mildly abnormal LV beyond its compliance point (e.g., high-output states such as obesity or anemia). Mildly dilated LV is most characteristic of the young obese HFpEF (vs. small, stiff LV of elderly HFpEF). Obese high-output HF is part of the same spectrum but has a more dramatic rise in volume and cardiac output than obese HFpEF.
    • Chronotropic incompetence, seen in ~50% of patients, may be a contributor to exercise intolerance.
    • Microvascular dysfunction in middle-age women may be the result but also the cause of diastolic dysfunction in a substantial proportion of them, particularly those without LV concentric remodeling.

B. CAD (ischemia without infarction)

Relaxation being an active process, CAD may contribute to diastolic dysfunction. In fact, in patients with CAD, a rise in LVEDP (diastolic stiffening) is an early hemodynamic manifestation of angina induced by pacing or exercise.20

As importantly, CAD may lead to transient LV systolic failure or dynamic ischemic MR that is mislabeled as diastolic failure, and thus, CAD needs to be considered in all acute heart failure cases that are labeled diastolic failure.

Overall, CAD is a less common etiology of HFpEF than HF with reduced EF, but is still considered the underlying etiology of HFpEF in 25–45% of patients in large HFpEF trials.21,22 As with HF with reduced EF, significant, usually multivessel CAD must be present to be considered the underlying mechanism of HFpEF.

C. Restrictive cardiomyopathy (RCM)

RCM is characterized by:2326

  1. LV is small and stiff with severe diastolic failure. Unlike in dilated cardiomyopathy, the LV cavity is small. Systolic function is relatively preserved and becomes progressively impaired at an advanced stage, yet the LV remains small. The RV may also be stiff and small or may be dilated; RV dilatation/RV systolic dysfunction is common at an advanced stage and is a poor prognostic sign.27
  2. As in any decompensated ventricular failure, LA and RA are markedly dilated and functional TR and MR, sometimes severe, may be seen.
  3. The myocardial thickness is increased ≥14 mm in infiltrative restrictive cardiomyopathies, particularly amyloidosis, where it may reach levels seen with hypertrophic cardiomyopathy (>20 mm) and may occasionally be asymmetric. As opposed to hypertrophic or hypertensive cardiomyopathy, the increase in thickness is due to myocardial infiltration rather than myocardial hypertrophy, explaining the discrepancy between the low voltage on the ECG on the one hand and the thick myocardium on echocardiography on the other hand. As opposed to hypertensive cardiomyopathy, the thickening frequently involves both ventricles, not just the LV.

    Moreover, amyloidosis has the following echo characteristics: a small pericardial effusion, valvular thickening, and a granular, “spar- kling” myocardial texture. The sparkles are bright echo spots corresponding to amyloid deposits; note that the heterogeneous texture of HCM may simulate the sparkled texture of amyloidosis.28

  4. On ECG, two findings are common with the thick amyloid cardiomyopathy, corresponding to replacement of the myocardium by the amyloid material: pseudo-Q waves and low QRS voltage, or at least a QRS voltage that is disproportionate to the thickness of the myocar- dium. LVH voltage criteria are never seen in the limb leads in amyloidosis.

Causes of RCM:

  • Infiltrative disease, most commonly amyloidosis, but also Fabry disease, hemochromatosis, sarcoidosis, and hypereosinophilic syndrome.
  • Scleroderma, lupus, rheumatoid arthritis- Primary idiopathic RCM
  • Radiation heart disease, which leads to a combination of RCM, constrictive pericarditis, valvular heart disease (AS/AI, MR, TR), CAD, and pulmonary fibrosis. All these develop 5–30 years after radiation exposure.

D. Constrictive pericarditis

Constrictive pericarditis mimics the presentation of RCM and RV failure.

III. Right HF

Left HF is the most common cause of right HF. Overall, there are three mechanisms of RV failure, including isolated RV failure:

  1. Pressure overload: pulmonary hypertension secondary to left heart disease, lung disease, PE, or pulmonary vascular disease.
  2. Volume overload: ASD, tricuspid or pulmonic regurgitation. TR is often functional and secondary to RV failure but exaggerates its pro- gression through the extra volume load. TR may also be isolated and the cause of an otherwise unexplained RV failure.
  3. Intrinsic RV dysfunction: ARVD or RV infarct. Acute myocarditis, tachycardia-mediated cardiomyopathy, and idiopathic, HIV, or alcoholic cardiomyopathy usually lead to RV and LV involvement, but one may be more predominantly involved than the other.


I. Echocardiography

As mentioned under Definition, HF is a clinical diagnosis. Echocardiography is done to corroborate the three major categories of HF based on the assessment of EF, valvular function, diastolic dysfunction (low E’), and LA pressure (elevated E/E’).

Schematic illustration of change of LV filling pattern between compensated and decompensated HF.

Figure 4.3 Change of LV filling pattern between compensated and decompensated HF. The E and A waves correspond to the transmitral flow in diastole. The L wave is a marker of elevated LA pressure which attempts to push flow throughout diastole.

  • Echo estimates LA pressure and pulmonary arterial pressure, which, if high, indicate decompensation and some degree of pulmonary congestion. Pulmonary hypertension, per se, may persist after resolution of the pulmonary congestion. In fact, a chronic increase in pulmonary pressure may lead to a reactive pulmonary hypertension that may be slow to resolve after normalization of LA pressure.
  • Echo can suggest a cause, such as ischemia in the case of focal wall motion abnormalities (albeit not very specific), or hypertensive or infiltrative cardiomyopathy in the case of a thickened septum.
  • Echo assesses for other forms of HF:

    • Isolated RV failure with its various causes (pulmonary arterial hypertension, ASD).
    • HF mimickers: constrictive pericarditis, tamponade.


BNP is a peptide synthesized by the myocardium in response to cardiomyocyte stretch. Atrial or ventricular wall stress, especially LV wall stress from volume or pressure overload, is the primary driver of BNP secretion, but ischemia and neurohormones may also directly stimulate BNP gene expression. RV and/or atrial stretch may increase BNP to lesser degrees than LV stretch, as RV and the atria have a smaller mass.

Pro-BNP is the precursor of BNP; it is cleaved into BNP and NT-pro-BNP. NT-pro-BNP value is usually ≈4 x BNP, and more so (>6x) with age>65 or renal failure.

A. BNP > 400–500 pg/ml (or NT-pro-BNP > 1200 pg/ml) is highly suggestive of acute left HF in a patient with acute dyspnea. It correlates with increased LVEDP; however, the exact BNP value does not correlate with the exact degree of LVEDP rise. For a given filling pressure, dilated ventricles secrete more peptide because of the greater wall stress/stretch and chamber mass. This explains why BNP is higher in LV systolic than in LV diastolic failure, and how it may be high despite a normal LVEDP. Also, several factors such as age, sex, and renal failure contribute to the absolute BNP levels.

B. BNP < 100 pg/ml (or NT-pro-BNP < 300 pg/ml) excludes acute left HF as a cause of acute dyspnea. The patient may, however, have asymptomatic or less acutely symptomatic LV dysfunction (exertional symptoms). In fact, patients with diastolic or even moderate systolic dysfunction who do not exhibit overt HF frequently have BNP values that overlap with normal individuals’ BNP (BNP < 50 pg/ml). In one study, most patients with asymptomatic LV systolic dysfunction had BNP levels < 50 pg/ml.39 In another study that focused on LV diastolic dysfunction, a significant proportion (~25%) of patients with echocardiographic diastolic dysfunction had BNP < 50, but this was uncommon in patients with a history of HF, current clinical HF, or severe diastolic abnormalities on echo (the mean BNP in these patients was > 300).40

A BNP cutoff of 35 pg/ml (NT-pro-BNP 125 pg/ml) is a better cutoff to exclude chronic HF as a cause of chronic exertional symptoms (ESC).8,10

C. In the acute setting, BNP levels of 100–500 pg/ml are in the intermediate range. Dyspnea may be due to acute left HF but also to acute pulmonary or pulmonary vascular illness (with RA/RV stretch), or sepsis. BNP may also rise because of RV failure, renal failure, or myocardial ischemia.

In the chronic setting, BNP levels of 35–100 pg/ml are in the intermediate range (up to 240 pg/ml if AF).10 Chronic dyspnea may be due to HF or non- cardiac disease.

D. Evolution and follow-up of BNP levels. In most patients with acute HF, BNP decreases with diuresis and the reduction of LV filling pressure but may not return to normal, as the LV remains dilated even after LVEDP is normalized (BNP may be as high as several hundred pg/ml, but is not usually higher than 1000 pg/ml). This residual BNP correlates with the degree of underlying LV, RV and atrial remodeling.41,42 Even when euvolemia is achieved and LVEDP normalized, a dilated LV will continue to generate an elevated BNP (called residual, dry BNP). Also, the reduction in BNP may be delayed because of reduced clearance, especially in renal failure. Thus, daily BNP measurements are not warranted; when clinical euvolemia is established, a pre-discharge BNP level may be determined and used in follow-up.42 A persistently elevated BNP upon discharge is a predictor of HF readmission and has a negative prognostic value. In one analysis, a 30–50% BNP or NT-pro-BNP reduction during the hospital stay strongly predicted a reduction in hospitalizations, beyond clinical variables, suggesting that BNP may be used as a target before HF discharge.43

BNP levels may be followed in stable outpatients to detect early congestion, but this has pitfalls, because BNP has an intra-individual biological variability of up to 25–50% unrelated to the development of symptoms.42 Conversely, a rise in BNP of over 25–50% may herald a clinical decompensation and suggests drug titration. Several studies and a meta-analysis have shown the utility of a BNP-guided outpatient drug titration in patients < 75 years old with systolic HF (to a target BNP < 100– 300). This strategy reduced HF hospitalizations and improved survival in the meta-analysis (class IIa recommendation).44,45 In those studies, BNP was checked every 1–3 months. ACE-I, β-blocker, and aldosterone antagonist titration was prioritized over diuretic titration in the absence of clinical congestion. The utility of BNP-guided therapy in older patients and in patients with diastolic dysfunction is unclear; overzealous therapy to reduce BNP in these patients may be associated with more side effects (dizziness, fatigue).

In critically ill patients with severe sepsis or shock, BNP levels poorly correlate with left- or right-sided filling pressures; BNP levels higher than 1000 may be found with normal PCWP and rather correlate with renal failure in this setting.47


A completely normal ECG almost excludes the diagnosis of HF, particularly systolic HF and acute HF.8

Q waves are the single most important finding and point toward ischemic cardiomyopathy. Other abnormalities are non-specific but have prognostic values: LVH with strain pattern, ST–T abnormalities, RBBB or LBBB, AF and various arrhythmias.

  • Advanced AV block is more characteristic of myocarditis and infiltrative disease than other cardiomyopathies.
  • The association of a large QRS voltage in the precordial leads with a low QRS voltage in the limb leads suggests dilated LV with low EF.
  • LBBB with QRS > 130 ms suggests LBBB-aggravated cardiomyopathy and has therapeutic implication (→benefit from CRT).

IV. Coronary angiography and other ischemic tests

Coronary angiography is recommended in the setting of angina (class I) or acute severe presentation. Otherwise, it is reserved for HFrEF with high pre-test probability of CAD (ESC guidelines class IIb).8

Alternatively, any one of the following studies is performed (ESC class IIb for stress testing, class IIa for CT):

  • Nuclear rest and stress perfusion scan, looking for segmental defects suggestive of CAD. However, nuclear defects are common in non- ischemic cardiomyopathy (patchy scar tissue). On the other hand, up to 30% of patients with multivessel CAD have balanced ischemia without any apparent defect.48,49 Thus, nuclear testing is not sensitive or specific to diagnose CAD as the cause of cardiomyopathy.
  • Stress echocardiography: a biphasic response (i.e., an initial increase in contractility at low doses of dobutamine followed by a decrease in contractility at higher levels of stress) suggests that LV is viable and able to increase its contraction but gets ischemic with high stress. A uniphasic response with a progressive increase in contractility from lower to higher doses suggests non-ischemic cardiomyopathy, or ischemic cardiomyopathy from a coronary stenosis that has been revascularized, but the myocardium has not recovered yet or full recovery is prevented by a severe subendocardial scar. The absence of any increase in contractility suggests the absence of myocardial reserve: LV dysfunction could be ischemic or non-ischemic and is possibly irreversible (non-viable), but viability and the potential benefit from revascularization should not be ruled out based purely on this finding.
  • Coronary CT angiography may be used to rule out CAD, mainly when the probability of ischemic HF is low or intermediate.

V. Diastolic stress testing

Chronic dyspnea that is associated with a normal systolic function on echo and no clinical signs of HF may be falsely considered a dyspnea of non-cardiac origin. These patients may have normal pulmonary capillary pressure and PA pressure at rest. However, exertion and the increase in venous return into a stiff LV increases their left-sided filling pressures. This is a masked form of LV diastolic dysfunction that is unveiled with exercise. Thus, “diastolic stress testing” is valuable in patients with chronic unexplained dyspnea, especially if they have borderline BNP or echo findings at rest (e.g., low E’, large LA, or LVH, but normal E/E’ ratio). It consists of reassessing LA pressure (E/E’ ratio>14) and TR velocity (>3.4 m/s) at peak exercise. In addition, stress testing allows the assessment of ischemia, dynamic MR, and the objective functional limitation.10,12,50 Ideally a semi-supine bicycle test with imaging during exercise, or else a treadmill or upright bicycle with imaging within 2 min of peak exercise, is recommended. As an alternative to exercise, passive leg raising may be used (~40% of the eventual PCWP rise already occurs upon passive leg raising).

Some patients with diastolic dysfunction but normal left-sided filling pressure at rest truly have a non-cardiac cause of exertional dyspnea. Diastolic stress testing helps sort out whether dyspnea is due to diastolic dysfunction or a non-cardiac cause.

Diastolic stress testing may also be performed invasively, using a Swan catheter and a supine bicycle.

VI. Endomyocardial biopsy

Endomyocardial biopsy is indicated when severe, progressive myocarditis is suspected, in which case fulminant lymphocytic myocarditis (reversible) and hypersensitivity eosinophilic myocarditis (treatable and reversible) need to be distinguished from the downhill progressive giant-cell myocarditis.51 Fulminant myocarditis often recovers in a few weeks but requires aggressive early support, while giant-cell myocarditis may necessitate cardiac transplantation with immunosuppressive therapy and LV assist device in the interim period. Thus, a biopsy is indicated in the case of unexplained, acute HF (<2 weeks) with hemodynamic compromise; or new HF (<3 months) with either failure to adequately respond to 1–2 weeks of usual care, or intractable ventricular arrhythmias or advanced AV block. It is also indicated when drug reaction is suspected but HF is not quickly improving after drug withdrawal (biopsy establishes the diagnosis of eosinophilic myocarditis and allows steroid therapy).

A biopsy may also be performed to diagnose infiltrative disease: (i) if cardiac AL amyloidosis is suspected on serum protein immunofixation but abdominal pad biopsy is inconclusive; (ii) if sarcoidosis is suspected without a typical pulmonary involvement.

VII. Cardiac MRI

Normally, gadolinium does not penetrate the myocardial cells and briefly and mildly penetrates the myocardial circulation during first-pass MRI perfusion imaging (myocardial cells are tightly packed with no significant interstitium). Late gadolinium enhancement (LGE) on T1 usu- ally implies necrotic or fibrotic tissue, to which gadolinium has a high affinity. Cardiac MRI has four major applications in HF: (i) assessment of LGE patterns that are specific for some cardiomyopathies (subendocardial or transmural LGE implies ischemic cardiomyopathy, subepi- cardial or midwall LGE in a non-coronary distribution implies non-ischemic cardiomyopathy); (ii) diagnosis of myocarditis in unexplained cardiomyopathy or unexplained, large troponin elevation; (iii) diagnosis of infiltrative cardiomyopathy vs. hypertrophic cardiomyopathy in a patient with thick myocardium (global subendocardial LGE suggests amyloidosis, patchy midwall LGE suggests HCM, patchy inferolateral LGE suggests Fabry);52 (iv) viability assessment in ischemic cardiomyopathy.



I. Treat the underlying etiology: target BP and CAD

HTN should be treated to a goal < 130/80 mmHg. In systolic HF, HTN may get “normalized” or reduced as a result of the low cardiac output; HF drugs are provided regardless of blood pressure, up to SBP of 90-100 mmHg, as long as clinically tolerated. Revascularization with CABG or PCI should be considered in patients with ischemic cardiomyopathy, i.e., CAD extensive enough to explain the cardiomyopathy, rather than incidental single-vessel CAD (as explained under Causes of heart failure). Revascularization should also be considered in patients with diastolic HF and extensive CAD.

II. Value of revascularization in ischemic cardiomyopathy: STICH trial

The STICH trial randomized patients with ischemic cardiomyopathy and EF ≤ 35% to revascularization with CABG vs. medical therapy.53,54 Most of these patients had three-vessel CAD, often with proximal LAD disease, and a prior history of MI; patients were not excluded based on viability testing. The trial excluded patients with significant (CCS 3 or 4) angina and consisted mainly of patients with LV dysfunction, no angina, and moderate HF (class II–III for the most part). At 5 years, CABG was associated with a strong trend towards mortality reduction (~8% per year vs. 6.5% per year, p = 0.12), which became significant at 10 years, and a significant reduction of each of the following endpoints: cardiovascular mortality and hospitalization for HF or cardiovascular causes. In addition, the benefit was higher in the as-treated analysis and the early hazard with CABG was relatively low in the trial (3.6 % mortality at 30 days).

Thus, overall, this trial supports CABG in ischemic cardiomyopathy, particularly 3-vessel CAD with the lowest EF <27%. Yet the benefit from CABG did not appear until 2 years of follow-up and the benefit was not dramatic, which implies that CABG is not urgent in patients with ischemic cardiomyopathy and is only given a class IIb indication for HF purpose (ESC). It is unclear whether the benefit extends to sicker patients with more advanced clinical HF: they have higher surgical risk and may not live long enough to derive the late CABG benefit.

III. Role of viability testing and ischemic testing

In the STICH trial, ~50% of patients underwent viability testing (with echo or SPECT) and were randomized to CABG vs. medical therapy regardless of the result. CABG was beneficial regardless of viability, and viability did not provide any discriminatory effect.55,56 Viability did predict eventual EF improvement; in fact, EF improved only in patients with viability, but to a similar magnitude with either CABG or medical therapy. Importantly, CABG reduced mortality independently of EF improvement, more so in patients with the most extensive disease and the lowest EF.57 There was no association between EF reversal and subsequent mortality, or between viability and 10-year mortality.56 This suggests that in the chronic ischemic setting, revascularization for ischemic LV dysfunction is mildly beneficial regardless of viability testing, and regardless of whether EF eventually improves (viability testing is not recommended in ESC guidelines).8

A. Definition of global and regional viability; viability tests

Viability implies that a dysfunctional myocardium is still metabolically active but is hibernating from persistent severe ischemia; or stunned after a transient episode of severe ischemia, even if perfusion is re-established. Those processes are called hibernation and stunning, respectively. A dysfunctional myocardial area is considered viable if:5862

  • The regional dysfunction significantly improves after revascularization by at least one point (from akinesis to hypokinesis, or from hypoki- nesis to normal). This is called regional viability.


  • The overall EF significantly improves (>5%). This is called global viability .

Thus, viability can only be confirmed in retrospect, after revascularization is performed. In general, 70% of patients with chronic ischemic LV dysfunction have significant improvement in LV function after revascularization (81% in STICH trial). Four imaging modalities have been used to assess viability:

  • Thallium myocardial uptake at rest and at 4 hours and 24 hours post-rest injection (rest-redistribution imaging). Also, technetium uptake at rest (or after nitrates administration) may be used to assess viability. An uptake ≥ 50% at rest or after thallium redistribution implies that the regional myocardium is viable.63 Also, an absolute change of 12% during redistribution implies regional viability even if the uptake remains < 50%.55,64
  • Low-dose dobutamine echocardiography evaluates for an improvement of myocardial contraction in response to low doses of dobu- tamine (<10 mcg/kg/min): from dyskinesis or akinesis to hypokinesis, or hypokinesis to normal contraction (a change from dyskinesis to akinesis is not considered improvement). This is called myocardial contractile reserve. A decline in myocardial contraction at subsequent higher doses reflects ischemia; this biphasic response is highly predictive of myocardial recovery and viability. Conversely, a uniphasic response (progressive increment of contractility at higher doses of dobutamine) suggests that the coronary flow is not compromised and thus, revascularization cannot provide any benefit; myocardial dysfunction is non-ischemic, or ischemic with subendocardial scarring that persisted after blood flow was re-established. The absence of any response may imply non-viability; yet, 25–30% of viable segments do not augment with dobutamine. Nuclear studies are more sensitive for viability, albeit less specific.
  • PET scan assesses metabolic uptake of a glucose analog tracer (metabolic viability). It can also assess perfusion with N13 or Rb82 tracer. A mismatch between a low perfusion and a high metabolic uptake implies viability. A reduction in both perfusion and metabolic uptake implies irreversible injury. PET has the highest sensitivity (~90%) but the lowest specificity (~58%) for viability assessment.
  • Cardiac MRI: late myocardial hyperenhancement 10 minutes after gadolinium injection is a sign of scarring and highly predicts non-viability when it involves > 50% of the transmural myocardial thickness. Transmural involvement <25% highly predicts viability, while 25-50% involvement is less specific but usually viable.

    Also, several ECG, echocardiographic, and angiographic features are helpful, to a limited extent, in assessing viability:

  1. The lack of Q waves implies less scar and necrosis and strongly predicts recovery.65,66 However, the presence of Q waves does not rule out viability, as Q waves may be seen with a hibernating myocardium.65
  2. As compared to akinesis, hypokinesis does not necessarily imply a higher likelihood of recovery with revascularization. After revascu- larization, akinesis and dyskinesis frequently improve, while occasionally hypokinesis does not improve. The hypokinetic wall may be tethered by adjacent akinetic walls or extensive subendocardial necrosis and scarring.67 Yet, hypokinesis implies the presence of viable tissue, and thus revascularization of a hypokinetic wall may prevent further deterioration of LV function and is more readily performed.
  3. A thin and bright myocardium ≤ 5.5 mm on echo may imply a scarred, non-viable myocardium. Yet, if it is not extensive, it does not preclude a revascularization benefit to the surrounding walls. Also, one study showed that 18% of thin myocardial regions have a limited scar burden by MRI and recover myocardial thickness and contractility with revascularization, with Q-wave disappearance; thus, a thin myocardium does not preclude revascularization.68
  4. Severe LV dilatation (end-diastolic volume ≥ twice normal, which often corresponds to a diastolic diameter ≥ 70 mm) usually implies extensive scarring and negative remodeling that would not be reversed with revascularization and predicts a poor prognosis after revascularization.69
  5. When a dysfunctional segment is supplied by an occluded artery, good collateral flow predicts myocardial recovery.70 In one study, the presence of grade 2 or 3 collaterals on coronary angiography, i.e., good collaterals that reconstitute partially or fully the epicardial vessel, predicted regional and global myocardial recovery in 81% of the cases.71

B. Variations in the definition of global viability

In the STICH trial, global LV viability has been defined as viability of ≥ 65% of all myocardial segments (11 out of 17 segments), counting both dysfunctional segments and segments with normal contractility.55,72,73 Conversely, most studies have defined global viability with a focus on the dysfunctional segments. When the viable dysfunctional segments constitute ≥ 25% of the LV, global viability is considered present and a global improvement of EF is expected (e.g., 4 dysfunctional myocardial segments are viable).

C. Limitations of viability testing

Viability testing has the following limitations:

  • Viability tests have positive and negative predictive values of ~75–80%, for both regional and global viability. This is particularly the case with thallium or technetium nuclear viability testing (sensitivity ~80%, specificity ~60%) and dobutamine echocardiography (sensitivity 65%, specificity 85%), which were used in the STICH trial.58,75 The assessment of the extent of transmural scarring by MRI is more valid (sensitivity and negative predictive value 90%, positive predictive value 60-80%, depending on the transmural cutoff used).
  • The definition is not uniform. In a patient with anteroapical infarction, regional viability addresses the viability of the anteroapical segments to decide whether LAD revascularization is beneficial; conversely, global viability addresses whether there are ≥4 dysfunctional segments that are viable.
  • Even without any viability and any contractile and EF improvement, revascularization improves symptom status, HF, and mortality.76 Revascularization may improve LV remodeling, prevent further deterioration of LV function and further MI, preserve infarct border zone integrity, and reduce arrhythmogenesis.
  • A STICH analysis has shown that patients with the most severe ischemic cardiomyopathy derive the most benefit from CABG (up to 29% mortality reduction); this was defined as a combination of 2 of the following 3 factors: 3-vessel CAD, EF<27%, dilated LV with end-systolic volume index >79 ml/m2. 56,57 This extends to many patients with limited viability, as their cardiomyopathy is more severe than the viable group.

D. Role of ischemic testing; difference between ischemia and viability

The concepts of ischemia and viability are different, and ischemia is a stronger marker of prognosis and functional myocardial recovery than the mere presence of viability.7780 On nuclear imaging, viability implies that at rest, the dysfunctional territory uptakes significant nuclear agent. Ischemia implies that at stress, the nuclear defect of the dysfunctional territory is significantly more intense or more extensive than at rest; or significant ST-segment deviations occur. An ischemic territory is viable; conversely, a viable territory may not be ischemic. The lack of ischemia means that the myocardial territory is already receiving enough blood supply, and should be able to recover its function without revascularization if truly viable, which questions the value of revascularization (Figure 4.4).81

Nonetheless, the assessment of ischemia in LV dysfunction with significant baseline defects is cumbersome, as fixed wall motion abnormality on echo or fixed defects on nuclear imaging may very well represent severe ischemia with hibernation, particularly if the nuclear uptake is ≥ 50%, or no Q wave is present, or the patient has angina or ST deviation; up to 75% of these defects reverse with revascularization.82,83 This lessens the predictive value of ischemic testing in LV dysfunction; in STICH stress substudy, ischemia did not predict prognosis or response to CABG.84

E. Indications for revascularization in ischemic LV dysfunction

Based on the STICH trial, global revascularization is mildly useful in chronic ischemic cardiomyopathy regardless of viability and of eventual EF improvement. Viability may be used to guide the specifics of regional revascularization, particularly with PCI, but not to exclude the overall revascularization (e.g., may not revascularize one non-viable territory, such as RCA CTO). On the other hand, regardless of viability, revascularization may not be beneficial in patients with a recent STEMI or Q-wave infarct that is 1–28 days old and responsible for the low EF (those patients were excluded from the STICH trial). As shown in the OAT trial, patients with acute MI who do not receive culprit artery reperfusion in the first 24 hours of presentation, and have an akinetic wall, a totally occluded culprit artery, and no residual angina, do not benefit from revascularization.85

IV. Drugs that affect survival in EF<40%

Three compensatory mechanisms occur in HF and are ultimately harmful.

  1. LV remodeling is the process of LV dilatation, LV change in geometry, and LV eccentric hypertrophy that attempt to increase the stroke volume of a hypocontractile myocardium. This process may be useful to a certain degree early on. Over time, however, as the LV undergoes progressive remodeling, it becomes less elliptical and more spherical, progressively more dilated, thin, and fibrotic, with increased wall stress (afterload); all this ultimately decreases the stroke volume.

    LV end-systolic volume is the best measurement of LV remodeling. EF, often considered a contractility index, is affected by preload and afterload and is, in fact, more a remodeling index than a contractility index (Figure 4.5 ).

  2. Increased activity of the sympathetic system increases cardiac contractility but ultimately exhausts the myocardium, makes it less responsive to catecholamines, and promotes apoptosis.
  3. Increased activity of the renin–angiotensin–aldosterone system (RAAS) elicits vasoconstriction and increases blood volume. This aims to maintain the blood pressure and the kidney perfusion, but is deleterious to the LV. In addition, angiotensin II directly acts on harmful myocardial AT1 receptors that promote cellular growth, LV dilatation, and LV fibrosis and increase the release of myocardial norepinephrine.

    Six main treatments improve mortality. They reduce LV remodeling, promote reverse remodeling, and, eventually, increase EF even if myocardial contractility is unchanged:

    Schematic illustration of (Top) The worsening of a resting defect is indicative of ischemia.(Bottom) A fixed defect suggests the lack of ischemia.

    Figure 4.4 (Top) The worsening of a resting defect is indicative of ischemia.(Bottom) A fixed defect suggests the lack of ischemia. Yet, if the patient has angina and no Q waves, or if the nuclear uptake is ≥ 50%, this fixed defect may indicate hibernation and ischemia.

    Schematic illustration of case of chronic HF with LV dilatation and mild increase in PCWP.

    Figure 4.5 Case of chronic HF with LV dilatation and mild increase in PCWP.

Left figure corresponds to LV volume and pressure before diuresis and before drug therapy. After diuresis, RAAS blockade, and β-blockade (right figure), LV volume decreases and LV morphology changes from a sphere to an ellipse. Stroke volume increases because of reduced afterload (smaller LV), reduced functional MR, and reduced RV–LV interdependence. Diuresis is beneficial, even though PCWP is only mildly elevated at baseline.

  • Note that afterload correlates with LV size and is not simply equivalent to aortic pressure (Laplace law):
  • Afterload = wall stress = LV radius x LV systolic pressure/2x LV thickness
  • EDV, end-diastolic volume; SV, stroke volume.

  1. Angiotensin-converting-enzyme inhibitors (ACE-Is) or angiotensin receptor blockers (ARBs). ACE-Is and ARBs improve HF outcomes not only through afterload reduction but also through blocking the harmful RAAS effect on volume status and on the AT1 myocardial receptor. In fact, afterload reduction with amlodipine or with α-blockers does not improve HF outcomes.
  2. β-Blockers blunt the high sympathetic tone of HF.
  3. Aldosterone antagonists.
  4. Neprilysin inhibitor combined with ARB.
  5. Sodium-glucose cotransporter-2 (SGLT-2) inhibitors.
  6. Hydralazine and nitrate combination. This vasodilator combination reduces mortality and symptoms in black patients and reduces symptoms in white patients.

V. Specifics of drugs that affect survival

ACE-Is/ARBs and β-blockers are indicated in all cases of systolic LV dysfunction with EF<40%, including asymptomatic patients.

A. ACE-I or ARB*

  1. ACE-Is and ARBs increase cardiac output, reduce LV remodeling, and improve EF. They reduce mortality by ~20% and reduce HF hospi- talizations by ~30%.8689
  2. ARB is an alternative to ACE-I when cough or angioedema occurs with ACE-I. There is a small cross-risk of angioedema with ARB (<5%), and thus extreme caution is advised when substituting ARB for ACE-I in a patient who had angioedema with ACE-I; this is not a contrain- dication to ARB (CHARM-Alternative). The main ARBs studied in HF are candesartan and valsartan.9093 When used as an alternative to ACE-I therapy, ARB therapy reduces mortality in systolic HF (CHARM alternative trial and ValHeFT substudy).
  3. Avoid or be cautious in cases of:

    • Hypotension with SBP < 90 mmHg or with symptoms of low output (dizzy, obtunded, oliguric). A low SBP of 90 mmHg may be well tolerated in HF because it helps unload the LV; dizziness upon ACE-I initiation often improves with time.
    • Elevation of creatinine of over 50% within 1–2 weeks of ACE-I/ARB initiation. In advanced HF, the worsening of renal function may be related to the low BP/low kidney perfusion state aggravated by ACE-I initiation, which impairs renal autoregulation (Figure 23.1). Yet, a rise in creatinine of over 25% and/or over 0.3 mg/dl early after initiation of ACE-I is not associated with a loss of benefit with continued ACE-I therapy, and renal function largely recovers on follow-up; in fact, in 2 studies, patients with early worsening of renal function appeared to derive the largest benefit from continued ACE-I therapy, partly because renal deterioration is a reflection of severe HF and baseline renal hypoperfusion.94,95 If creatinine rises ≥ 50% or ≥ 0.5 mg/dl, reduce other BP-lowering drugs, ensure the patient is not using NSAIDs, assess for excessive diuresis and consider reducing the diuretic dose. Along with that, hold the ACE-I for a few days to allow renal recovery then restart half the initial dose. In fact, one study suggests that even ACE-I-induced creatinine elevations ≥ 50% are not associated with increased mortality.94 Only if creatinine rises ≥ 100% ACE-I is stopped.8

      ACE-I/ARB is avoided in acute kidney injury, as it may acutely worsen renal perfusion. Short of that, it may be initiated or uptitrated acutely in decompensated HF, wherein it improves diuretic response and possibly even renal perfusion.96,97

    • K >5 mEq/l at baseline or >5.5 mEq/l with therapy. If K rises to 5.5-5.9 upon ACE-I initiation without a severe rise in creatinine, discontinue potassium supplements, then, if needed, reduce the dose of the aldosterone antagonist and the ACE-I.

  1. Attempt titration every 5 days to the optimal dose used in randomized trials, which is approximately one-half of the maximal dose. In comparison with the low dose, the intermediate or high dose further reduces HF hospitalizations by 24% but does not significantly improve mortality (ATLAS trial, where lisinopril 2.5–5 mg was compared with 30–40 mg). Thus, even if a high dose is not reached, it is expected that the low dose will lead to similar benefit on mortality.99

    Examples of doses:

    • Lisinopril (ACE-I): start 5 mg Qday and try to reach 20–40 mg Qday. It is renally eliminated and the effect is doubled in patients with renal failure, wherein the starting dose may be 2.5 mg Qday.
    • Enalapril (ACE-I): start 2.5 mg BID and reach 10 mg BID (also renally eliminated).
    • Candesartan (ARB): 4 mg Qday; goal, 32 mg Qday (biliary elimination).
    • Valsartan (ARB): 40 mg BID; goal, 160 mg BID (biliary elimination).

  2. Significant, sometimes drastic, improvement of exercise duration (20–50%) and functional class is seen with ACE-I.90 Symptomatic improvement may occur within 48 hours of ACE-I initiation but is generally delayed weeks to months. Abrupt withdrawal of ACE-I can lead to clinical deterioration and should be avoided, except in decompensated HF with low output and hypotension, or severe deterioration of renal function.
  3. In CHARM-Added trial, the addition of ARB to ACE-I in class II–III HF already receiving a β-blocker reduced both HF hospitalizations and cardiovascular death in comparison to ACE-I alone, without an effect on total mortality.91 Yet in patients receiving ACE-I, preference should be given to adding an aldosterone antagonist rather than ARB, as an aldosterone antagonist is associated with a larger reduction of HF hospitalizations and cardiovascular death, and a striking reduction of total mortality. Because the triple combination ACE-I/ARB/aldosterone antagonist is contraindicated (risk of hyperkalemia), the role of the ARB/ACE-I combination is limited in HF.

B. β -Blockers

  1. β-Blockers used to be contraindicated in HF, and in fact they may worsen HF initially, especially in the first 2–3 weeks after therapy initiation and/or up-titration. This is due to their initial negative inotropic effect. But when used over the long term, β-blockers:

    • Improve contractility, EF (by 5–15%), and HF symptoms, and reduce mortality by 35–60% and HF hospitalizations by 40%.100102
    • Reverse the deleterious and apoptotic effects of catecholamines on the heart. The reduction in energy demands reduces apoptosis.
    • Reverse remodeling and reduce cardiac chamber size by reducing LV wall stress.
    • Prevent arrhythmias; and reduce heart rate, which, by itself, is beneficial in HFrEF

  2. The three agents that have been shown to improve survival and outcomes in HF are:

    1. Long-acting metoprolol (metoprolol succinate): metoprolol is a selective β1-blocker, but loses selectivity with high doses > 100 mg. Long-acting metoprolol has a more sustained effect than metoprolol tartrate; this limits the daily fluctuations of the β-blocker effect and the β-blocker withdrawal effect that occurs between the doses of metoprolol tartrate.

      • Dose: Start 12.5 or 25 mg Qday and titrate to 200 mg Qday

    2. Carvedilol: carvedilol is a non-selective β1-blocker, β2-blocker, and α1-blocker (vasodilatory effect), with additional antioxidant proper- ties. Despite its α-blocker effect, it is as well tolerated as metoprolol in patients with borderline BP. The non-selective blockade of adrenergic receptors is advantageous. While metoprolol upregulates β-receptor density towards normal levels, carvedilol maintains a low density of these receptors. Moreover, selective β1-blockade with metoprolol may enhance the ino- and chronotropic response to β2-adrenergic stimulation, an untoward effect.103 Thus, carvedilol is a more potent antiadrenergic agent than metoprolol, as mani- fested by the more significant blunting of heart rate response.104 With its more comprehensive blockade of all adrenergic receptors, carvedilol improves myocardial function, EF, and cardiac hemodynamics such as stroke volume, PA pressure, and PCWP, more than other β-blockers and reverses remodeling more effectively.103,104 Because of the α-blocking effect, carvedilol acts as a moderate vasodilator acutely, but with long-term treatment the vasodilator activity is no longer prominent, as tolerance to the α-blocking effect occurs. This transient α-blocker effect is useful, as it allows an early improvement in stroke volume and LV remodeling before the long-term effect of β-blockade kicks in.105

      • Dose: start 3.125 mg BID; titrate to the goal of 25 mg BID if patient’s weight < 85 kg, 50 mg BID if weight > 85 kg.
      • A long-acting formulation of carvedilol, Coreg CR, is available. It is given once daily. Coreg CR 10 mg is equivalent to carvedilol 3.125 mg BID. Start Coreg CR 10 mg Qday and titrate up to 80 mg Qday.

    3. Bisoprolol, a selective β1-blocker. Start 1.25 mg Qday; goal, 10 mg Qday.

  3. Keys to β-blocker therapy:

    1. Start “low and slow.” Patients must be euvolemic and stable. During hospitalization for HF, low-dose β-blockade is started before discharge, after the patient has stabilized.
    2. Contraindications: overt HF, SBP < 90 mmHg or symptomatic hypotension, bradyarrhythmia (bradycardia < 55 bpm, any second- or third-degree AV block, PR > 0.24 s), bronchospasm.
    3. Double the dose every 2 weeks and monitor for worsening of dyspnea, edema, weight gain, bradycardia, and hypotension. If the patient is off the β-blocker for over 1 week for any reason, or after an episode of cardiogenic shock, restart at the lowest dose and re-titrate.
    4. If edema increases or HF develops within a week of titration, increase the diuretic dose; if this is not effective, decrease the β-blocker dose and titrate more slowly.

      In case of severe fatigue, decrease the dose of β-blocker (and sometimes the diuretic) or titrate more slowly. Most often, fatigue resolves spontaneously in a few weeks.

      In case of bradycardia (symptomatic bradycardia < 55 bpm, asymptomatic bradycardia < 50 bpm, or AV block), try to decrease the dosage of digoxin and amiodarone first, then, if needed, reduce the β-blocker.

    5. It is not necessary to reach an optimal dose of an ACE-I before starting a β-blocker. In fact, it is preferred to start low doses of both and titrate up in an alternating fashion. The combination of a low β-blocker dose and a low ACE-I dose produces a greater symptomatic improvement and mortality reduction than an increase in the ACE-I dose.

      Order of therapy: start low-dose ACE-I (lisinopril ~5 mg), then start low-dose β-blocker and titrate β-blocker every 2 weeks. Spironolactone then SGLT2 inhibitor may be sequentially initiated while the patient is still on small doses of ACE-I and β-blocker, before full uptitration of the latter. After reaching the maximally tolerated dose of β-blocker, and if it is possible from the blood pressure and renal standpoints, up-titrate the ACE-I at 1-week intervals.

    6. If the patient has been on a β-blocker for more than a few weeks and is hospitalized with HF decompensation, the β-blocker should not generally be withheld. In low-output HF decompensation with borderline BP, the β-blocker dosage may be reduced. In full-blown shock, or when inotropes are considered, the β-blocker is withheld. Discontinuation of β-blocker therapy is independently associated with increased short- and long- term mortality and should be avoided if possible.109 If interrupted, β-blocker is restarted at the lowest dose after the patient stabilizes, before hospital discharge, then re-titrated.
    7. How about EF 40-50%? Major β-blockers trials mainly included patients with EF<35-40%. Yet, two small studies that included any EF and a meta-analysis of β-blockers in HF showed that EF 40-50% derives the same mortality benefit as lower EF.110

      How about AF? Although no specific benefit was seen in AF subgroups on post-hoc analyses, major trials included patients with AF at baseline (~15%) and thus, AF patients should receive β-blockers.111

    8. Should β-blockers be titrated to achieve a specific target heart rate (HR)? The improvement in outcomes with β-blocker therapy is incremental with the following three factors, in order of importance: (1) dose achieved; (2) baseline severity of HF; (3) HR achieved and HR reduction.108,110,115,116 The 3 factors are interactive. The HR achieved with β-blocker therapy depends on the overall HF severity and catecholamine level, rather than just the β-blocker dose. A more severe or decompensated HF is associated with a higher HR yet derives a larger mortality benefit even from a small β-blocker dose. Thus, β-blockers are not titrated to achieve a specific target HR, although a low HR (~60–70 bpm) is desirable, is associated with a reduced mortality, and implies better HF control through both β-blockade and adjunctive HF therapy. Part of the reason metoprolol XL achieved similar relative benefit vs placebo at low and high doses is the similar HR achieved with both doses in the trial. Either a high β-blocker dose or a low HR (<70 bpm) should at least be achieved, and it is crucial to keep pushing the β-blocker dose even if the patient becomes asymptomatic. Even if HR remains > 70 bpm, the mortality is low when a high dose of β-blocker is reached.115 The relationship between lower heart rate and survival is only true in sinus rhythm, not AF.111

      While a high HR is essentially secondary to HF, it may also, by itself, perpetuate systolic HF through perpetuating energetic and metabolic inefficiencies. A trial of a pure sinus node blocker (ivabradine) showed that pure HR reduction was associated with reduced HF hospitalizations, confirming a direct role of HR in the pathophysiology of systolic HF.117

C. Aldosterone receptor antagonists (spironolactone, eplerenone)

  1. The high aldosterone levels seen in HF not only induce renal sodium retention but directly act on the myocardial and arterial aldoster- one receptors, leading to myocardial and arterial fibrosis and baroreceptor dysfunction (sympathetic augmentation). In fact, the small doses of aldosterone antagonists studied in HF only have a mild or no diuretic effect (RALES trial).118 The benefit in HF mainly results from: (i) blockade of myocardial and baroreceptor aldosterone receptors; (ii) increase in potassium (antiarrhythmic effect); (iii) reduction of the tubular resistance to loop diuretics.119 Aldosterone antagonists reduce HF mortality by 30%, HF hospitalization by 30–40%, and improve NYHA functional class (RALES trial).118,120,121

    Aldosterone acts at the DNA level and induces the genomic synthesis of distal tubular Na channels and basal Na/K pumps that absorb Na and secrete K and H. The aldosterone receptor blockers inhibit this synthesis rather than directly block Na/K pumps, and therefore have a slow onset of action of 2–3 days and a slow offset (3–4 days), similar to aldosterone and other steroids.

  2. An aldosterone receptor antagonist is indicated in chronic HF with EF ≤ 35% and NYHA classes III–IV.118 Aldosterone antagonist therapy has also reduced mortality in class II systolic HF (EF ≤ 35%) with elevated BNP or recent cardiovascular hospitalization in the last 6 months, and is recommended for these patients (class I recommendation).120

    It is also indicated after a recent MI (within 30 days) when EF is < 40% and any degree of clinical HF or diabetes is present.121

  3. Contraindications: creatinine > 2 mg/dl or GFR < 30 ml/min; K > 5 mEq/l. In the RALES trial, patients with mild renal failure (GFR 30–60 ml/ min) had a higher risk of worsening renal function and hyperkalemia with spironolactone in comparison to those with normal renal function, but derived the largest absolute mortality reduction. Thus, patients with mild renal failure are appropriate candidates for this therapy if well monitored.122
  4. Dosage of spironolactone: 12.5–25.0 mg daily; careful up-titration to 37.5–50.0 mg daily may be tried in patients with refractory HF or persistent hypokalemia. If gynecomastia develops, eplerenone may be used instead of spironolactone. Potassium and creatinine should be checked within 3 days and again at 1 week after therapy initiation, then Q2–4wk for the next 3 months. Typically, all potassium supple- ments need to be stopped at the initiation of the aldosterone antagonist; a minority of patients will continue to require a potassium sup- plement, but this is restarted only after blood testing. In real-world registries, hyperkalemia is common with aldosterone antagonists (up to 20%), therefore justifying careful K monitoring. In fact, these agents should be avoided if K monitoring is not possible (compliance issues), and the patient should be instructed about interruption of therapy should diarrhea or dehydration occur (renal failure). If K rises to 5-5.4 with spironolactone, maintain the same dose; if K rises to 5.5-5.9, hold it for a few days then resume half the dose; if K ≥6, discontinue it.
  5. RAAS activation leads to a marked elevation of aldosterone levels in both HF and cirrhosis, exaggerated by the fact that aldosterone catabolism, a hepatic process, is impaired. In cirrhosis, a high dose of spironolactone (100–400 mg), higher than the one studied in HF, has been shown to induce more natriuresis than a loop diuretic. In fact, the avid distal sodium reabsorption induced by hyperaldosteronism makes loop diuretics ineffective in 50% of cirrhotic patients.123 This high dose has not been largely studied in HF because of the combined therapy with ACE-I, but illustrates the potential value of spironolactone in patients with combined heart and liver failure and the possible role of higher spironolactone doses in patients resistant to loop diuretics.123 In ATHENA-HF trial, a 4-day course of spironolactone 100 mg in acute HF, on top of furosemide, did not improve urine output, weight loss, and NT-pro-BNP but was safe (trial limitations: treatment was brief and patients were furosemide-responsive).124
  6. Patiromer is a K binder that removes K at the gastrointestinal level, hence lowering K by ~ 1 mEq/l. It allows the continuation of the same ACE-I and aldosterone antagonist doses when K is 5.5-5.9; and even when K is ≥6, after 1 week of interruption.125

D. Angiotensin receptor-Neprilysin inhibitor (sacubitril–valsartan combination)

Neprilysin is a peptidase that degrades natriuretic peptides, bradykinin, and adromedullin. Neprilysin inhibition increases the level of these substance via a “double negative” effect, counteracting vasoconstriction, sodium retention, and LV remodeling. In fact, natriuretic peptides reduce maladaptive hypertrophy and fibroblast proliferation in the heart and kidneys. In a trial of patients with HF and EF<40%, NYHA II–IV, and BNP > 150 or a prior HF hospitalization in the last year, valsartan+ neprilysin inhibitor (combined in one molecule, not just one pill) reduced the 2-year mortality by ~3% in comparison with ACE-I (relative risk reduction 16%).126 Hospitalization for HF was also reduced by ~3%. Angiotensin–neprilysin inhibitor was associated with more symptomatic hypotension yet less renal dysfunction and hyperkalemia than ACE-I, and slower renal decline over the long term: natriuretic peptides selectively vasodilate the afferent arteriole, thereby increasing renal flow and offsetting the hypotensive effect and the efferent dilatation.127 This benefit was seen despite a background therapy that included β-blockers and aldosterone antagonists (PARADIGM-HF trial). Most patients had class II symptoms, and the benefit was seen in stable and relatively low-risk patients, which argues that even early-stage HF patients who are stable on ACE-I are better switched to this therapy; the indication is more compelling in patients with recent or current HF hospitalization. In-hospital initiation of angiotensin-neprilysin inhibitor dramatically reduces early HF rehospitalizations by 42%, compared to in-hospital ACE-I followed by a later switch to angiotensin-neprilysin inhibitor at 8 weeks (PIONEER-HF trial).128 Yet, its selective use in advanced class IV HF with mean SBP ~110 did not improve outcomes or NT-pro-BNP vs. ARB (LIFE trial); and its initiation in acute MI with low EF was not superior to ACE-I (PARADISE-MI).

In order to prevent angioedema, the recommended washout period between an ACE inhibitor and sacubitril–valsartan is 36 hours. A prior history of angioedema with ACE-I contraindicates the use of sacubitril–valsartan, but a prior history of cough is not a contraindication.

Of note, neprilysin inhibitor is the only HF therapy that increases BNP, a consequence of its direct effect. This affects the diagnostic value of BNP, at least during therapy initiation, but BNP remains useful for monitoring in respect to the new baseline. Conversely, NT-pro- BNP is not directly affected, as it is not a derivative of BNP, but a byproduct of the same precursor (pro-BNP).

E. Hydralazine–nitrate combination

In the V-HeFT-I trial, the hydralazine–nitrate combination reduced mortality by 34% in comparison to placebo or α-blocker therapy; this was the first trial ever to show a mortality reduction with vasodilator therapy in HF.129 The mortality reduction was mainly driven by the benefit in black patients. In the V-HeFT-II trial of hydralazine–nitrate vs. ACE-I, the hydralazine–nitrate combination reduced mortality as much as ACE-I in black patients, and was superior to enalapril in terms of exercise tolerance and EF improvement.130132 The hydralazine– nitrate combination was also beneficial in white patients, in whom it was superior to enalapril in terms of exercise tolerance and EF improvement, although the mortality was lower with enalapril. The reduction in hospitalization with hydralazine–nitrate was similar to enalapril and similar between white and black patients.132 In the A-HeFT trial of class III–IV black patients already receiving ACE-I and β-blocker therapy, the hydralazine–nitrate combination therapy strikingly reduced mortality by an additional 43% and reduced HF hos- pitalizations by 40%.133

Therefore, the combination is indicated as an additional therapy to ACE-I/ARB and β-blocker in black patients with a functional class III–IV (class I recommendation). It may be considered in non-black patients with persistent symptoms or persistent HTN (>140/90 mmHg); nitrates, per se, may be used for decongestion and symptom relief in all races (CHAMPION trial algorithm). The combination may also be used instead of ACE-I or ARB in case of intolerance to ACE-I/ARB, i.e., renal failure occurs with ACE-I/ARB and does not improve by reducing the dose of furosemide, or hyperkalemia occurs with ACE-I/ARB and does not improve with potassium restriction (class IIa recommendation regardless of race).

The benefit is partly related to the afterload and preload reduction. More importantly, nitrates are metabolized into nitric oxide (NO), while hydralazine prevents the oxidation of NO through its antioxidant properties, allowing to maintain the NO effect. ACE-I also prevents NO oxidation and may be beneficial in combination with nitrates. NO promotes endothelial and vascular homeostasis and appropriate myocardial remodeling and contractility, which improves EF by up to 5%. Hydralazine–nitrate combination may also directly reduce pulmonary vascular resistance in patients with left HF-associated pulmonary hypertension. Black patients frequently have a defective genetic variant of NO synthase that may explain the particular benefit; many white patients also have this variant, and thus pharmacogenomics rather than race may guide hydralazine–nitrate therapy in the future.

Dosing: start hydralazine 12.5 mg TID, increase it to 25 mg TID in 2–4 days, then increase by 25 mg/dose Qweek. Target is ~50–75 mg TID (may reduce to BID in renal failure). Start isosorbide dinitrate 20 mg TID and titrate up to 40 mg TID.

BiDil is a pill that combines 37.5 mg of hydralazine and 20 mg of isosorbide dinitrate (ISDN); BiDil may be started at one-half pill TID and titrated up to two pills TID. Beware, hydralazine causes a lupus syndrome of rash and arthralgia in 5-10% of patients.

VI. Drugs that improve symptoms and morbidity

A. Diuretics

  • Overview and effect of disease states. Thiazide diuretics are weak diuretics that block the distal tubule’s Na-Cl transporter, which reabsorbs ~3–5% of sodium, and are mainly used in chronic HTN rather than HF. Loop diuretics (furosemide, bumetanide, torsemide) block the ascending loop of Henle (Na-K-2Cl transporter), which reabsorbs ~25% of sodium, and are often the mainstay diuretics in HF. Both loop and thiazide diuretics are secreted by the proximal tubule into the lumen and work from the luminal side of the nephron.

    In HF and renal failure with reduced renal blood flow, less diuretic reaches the kidneys and gets secreted into the lumen, implying the need for higher doses. Moreover, the reduced amount of renally filtered sodium reduces the total amount of sodium that can be elimi- nated with every single diuretic dose, implying the need for more frequent administration.134

    All diuretics are highly bound to albumin, which traps them and delivers them to the secretory site of the nephron. In very low albumin states (<2 g/dl), some of the diuretic is lost in the extravascular space, and in nephrotic syndrome, even the diuretic that gets to the nephron’s lumen is bound to the urinary albumin, preventing it from acting on the sodium pump.

  • Renal failure and diuretics (especially thiazide). In patients with GFR < 30 ml/min, less sodium is filtered overall. Thus, less sodium can be eliminated by any diuretic (especially a thiazide diuretic), particularly because each of the remaining nephrons is already maximizing its sodium elimination. Hence, thiazide diuretics are less effective in patients with advanced renal failure, but, contrary to common belief, remain effective after cumulative dosing or in combination with a loop diuretic that increases the salt delivery to the distal tubule.105,106 Metolazone, a thiazide-like diuretic that also works on the proximal tubule, maintains its efficacy in advanced renal failure; it also has a prolonged duration of action (~1 week). All thiazide diuretics are renally eliminated and have a long half-life, and thus cumulate in renal failure (not just metolazone), which further potentiates their effect after multiple dosing. Some data suggest that thiazide monotherapy is effective in hypertensive patients with advanced CKD after multiple dosing.135
  • Diuretic threshold. A diuretic dose is only effective if it exceeds the required threshold at the tubular level. This threshold varies between individuals and diseases. Once the threshold is exceeded, there is a point at which maximal natriuresis is achieved, and beyond which no further natriuresis is gained with higher doses.134
  • Pharmacokinetics of loop diuretics.134 Approximately 50% of oral furosemide is absorbed, with a high intra- and inter-individual variation (10–90%). Thus, in general, the effective oral dose of furosemide is numerically twice the intravenous dose. Bumetanide and torsemide are more consistently and thoroughly absorbed (80–100%). With all these agents, the absorption is slowed by intestinal edema and poor intestinal flow; this flattened absorption curve prevents peaking of the diuretic concentration above the diuretic threshold and requires higher doses or an intravenous course of therapy before oral dosing is effective again.

    The duration of action is short (~4–6 h) and the nephron may avidly retain sodium between doses, which shows the importance of splitting the dose when a high total daily dose is required. Torsemide is twice longer acting (~12 hours) than the other two agents, has better oral bioavailability than furosemide, has a hepatic metabolism, and is given once daily; 2 small randomized trials showed that outpatient torsemide is superior to furosemide (less HF hospitalizations, reversal of myocardial fibrosis).136,137 It may be considered in patients with erratic diuretic effect or those with refractory fluid retention. While torsemide and bumetanide are metabolized by the liver, furosemide is renally excreted; thus, in renal failure, the elimination of furosemide is delayed, which prolongs its action and narrows its dose equivalence to the other two agents.

    Bumetanide dose equivalent: 1 mg = 40 mg of furosemide (or 20 mg in advanced renal failure); torsemide dose equivalent: 20 mg = 40 mg furosemide.

    Schematic illustration of diuretic response in a normal individual and in HF.

    Figure 4.6 Diuretic response in a normal individual and in HF. Note that no diuretic response occurs before achieving the threshold dose. In HF, even more so decompensated HF: (i) the threshold dose is higher, (ii) the diuresis achieved with the threshold dose is lower, and (iii) the maximal effect is lower. This suggests the value of frequent dosing of the threshold dose, rather than higher dosing, to increase the overall efficacy. Figure inspired by reference 134.

  • Practical use of oral loop diuretics:

    • Use the lowest chronic oral dose to maintain euvolemia: furosemide 20 mg once daily, up to 40–80 mg 2–3 times daily. If the patient has a diuretic response to 40 mg of furosemide and is requiring a total of 80 mg per day, it is better to give 40 mg BID than 80 mg Qday to provide a more sustained diuretic effect and less metabolic perturbations.
    • The patient needs to take extra dose(s) if weight increases > 3 lb in < 5 days. He needs to continue taking an extra dose of diuretic daily until the weight returns to baseline. If in 1–2 days, the weight continues to rise, a clinic visit is immediately warranted and metolazone may be added for 1-2 days.
    • Euvolemic outpatients who are minimally symptomatic on a small daily dose of furosemide ≤ 40 mg may consider furosemide withdrawal (ReBIC trial).138 On the other hand, patients with severe HF requiring high chronic doses of furosemide (>200 mg/d) may be supplemented with a biweekly or triweekly dose of chlorthalidone or metolazone.
    • Oral potassium (~20–40 mEq per day) may be added, particularly if the patient is not receiving an aldosterone antagonist. K goal is ≥ 4 mEq/l. Patients requiring high doses of potassium are often magnesium-deficient and should receive magnesium supplementation.

B. Digoxin

Digoxin has the unique property of combining a mildly positive inotropic action with a negative chronotropic action, making it the only positive inotrope that does not increase mortality. Digoxin inhibits the sarcolemmal Na/K-ATPase, which increases intracellular Na and subsequently increases the intracellular Ca through the Ca/Na sarcoplasmic passive exchange, leading to an inotropic effect. Digoxin also increases the vagal tone, which counteracts some of the catecholaminergic hyperactivity in HF, reduces HR, increases HR variability, and achieves some vasodilatation. This neurohormonal modulation is achieved at low levels of digoxin and actually explains most of the digoxin’s benefits in HF and the rate control in AF.141,142 EF improves with digoxin, including low-dose digoxin (up to 5%).142

The inotropic action of digoxin becomes stronger and begins to offset the therapeutic benefits provided by neurohormonal modula- tion at moderate and high serum digoxin levels, possibly explaining the increased mortality with higher levels.141 Higher doses may further improve EF and contractility parameters without further symptomatic or rate improvement.143 In addition, some believe that high digoxin levels have fewer neurohormonal effects than low levels.

Digoxin is indicated in HF that is still symptomatic despite all the above medications (class IIa indication). In the DIG trial of patients with HF and sinus rhythm,144 the addition of digoxin to standard therapy reduced pump failure death but was offset by an increase in arrhythmic death, making it neutral on mortality. Digoxin was overall beneficial and reduced HF hospitalizations by ~30%, particularly in the more severe cases (EF < 25%, class III or IV). Women had more hazard with digoxin, probably because of higher serum digoxin levels.

In a patient with HF and AF, digoxin is added as a second-line agent when β-blockers do not achieve appropriate rate control.

Once a patient is on digoxin, the withdrawal of digoxin worsens HF and increases HF hospitalizations. Therefore, once digoxin is started, it should not be withdrawn unless toxicity occurs.145

Dosing: 0.125 mg daily in most patients. In small patients, women, old patients > 70, and patients with renal failure, every-other-day dosing is considered. A dose of 0.25 mg daily may be used in young, large men with normal renal function. Loading doses are not necessary; loading is used for supraventricular arrhythmias rather than HF.

Digoxin level monitoring is encouraged in all patients, once steady state is achieved (~2 weeks after digoxin initiation). This is even more important in cases of renal failure, female sex, drug interactions, or suspicion of toxicity. The goal trough level is ≤ 0.8 ng/ml (serum level is checked at least 6 hours after the last dose). Trough levels of 0.5–0.8 ng/ml were associated with a lower mortality than placebo in the DIG trial; trough levels > 1.1 ng/ml were associated with a higher mortality than placebo.141 In addition, a dosing study has shown that all of the hemodynamic, EF, and autonomic improvement with digoxin is seen at the lower dose (0.125 mg/day) and lower concentrations, without further improvement at higher doses.142 This reflects the importance of neurohormonal modulation over further inotropic activity.

Signs of toxicity:

  • Bradyarrhythmias, tachyarrhythmias (atrial tachycardia with block, non-paroxysmal junctional tachycardia, and VT, including the charac- teristic bidirectional VT)
  • Nausea and vomiting, visual scotomas or halos, neurologic changes

Digoxin toxicity is usually seen with levels > 1.5–1.8 ng/ml, but may be seen at lower levels, particularly over the long term or when hypokalemia coexists. Digoxin half-life is ~36 hours, so digoxin effects are expected to last ~4 days after discontinuation. Digoxin Fab antibodies (DigiFab) are indicated for bradyarrhythmias or tachyarrhythmias associated with hemodynamic instability.*

C. Ivabradine

Ivabradine is a pure heart rate reducer (sinus node inhibitor). In the SHIFT trial of patients with class II–IV systolic HF (EF ≤ 35%), prior HF hospitalization in the last year, and sinus rate ≥ 70 bpm despite a maximally tolerated β-blocker dose, ivabradine reduced HF hospitalizations and HF death by 26%.117 This benefit is consistent among patients with severe class IV HF. However, being a sinus node blocker, it has no role in patients with AF; also, it slightly increases the risk of AF occurrence in patients with sinus rhythm. After proper β-blocker titration, only 14% of HF patients would qualify for ivabradine.146 See discussion under Section 1.V.B, last paragraph.

D. Glifozins= Sodium-glucose cotransporter-2 (SGLT-2) inhibitors (class I recommendation in HFrEF per ESC)

SGLT-2 is the main glucose transporter in the proximal renal tubule and is responsible for reabsorption of the filtered glucose in diabetes but also in normal individuals; Na is co-absorbed with glucose. SGLT-2 inhibitors cause glycosuria, and along with it, osmotic diuresis and natriuresis.

Trials of diabetic patients have consistently shown that glifozins reduce HF hospitalizations (by ~35%) regardless of baseline HF. In a paradigm shift, DAPA-HF trial specifically randomized HF patients with EF <40%, with or without (58%) diabetes, with NYHA II-IV, to dapaglifozin (10 mg/day) vs placebo; a similarly striking reduction of HF hospitalizations was seen in diabetic and non-diabetic patients, along with symptomatic improvement and significant mortality reduction (17%), even though patients were already receiving standard diuretic and HF therapies and mostly had NYHA class II.147 A similar reduction of HF hospitalizations was demonstrated with empaglifozin 10 mg/d in HFrEF with or without diabetes (Emperor-reduced trial). Furosemide dose was unaltered in most patients. This benefit is related to the diuretic effect but also to direct cardiac metabolic effects (↑ blood ketones, an efficient heart fuel, and ↑ myocardial glucose oxidation >fat); and renal effects: GFR declined at a much slower slope with empaglifozin (Emperor-reduced) or dapaglifozin (DAPA-CKD) vs. placebo. Glifozins are avoided if GFR<30.

VII. Devices


Fifty percent of patients with HF die of VT or VF. The remaining 50% of patients die of end-stage, low-output HF. Late MI contributes to 30–50% of both types of mortality in ischemic HF.148Sudden death is less common in absolute value, but more common in relative value, in mild class II HF than severe class IV HF.

An ICD is indicated as a secondary prevention measure if the patient has a history of sustained VT. It is indicated as a primary preven- tion measure in ischemic or non-ischemic cardiomyopathy with a low EF (≤35%) that is persistent over time despite optimal medical ther- apy, a functional class of II or III, and an expected survival of over 1 year. Patients with functional class I qualify for ICD in ischemic cardiomyopathy with EF < 30%:8,149

  • Wait > 40 days after an MI.
  • Wait > 3 months after revascularization for chronic ischemia.
  • Wait 3–6 months in non-ischemic cardiomyopathy, to ensure that it is not reversible and that it persists after a few months of medical therapy and BP control. Three to six months of waiting are reasonable but are not necessary according to the ACC guidelines.

Patients with functional class IV do not have an indication for ICD because of their high mortality. While they have a higher risk of VT/ VF than other functional classes, they also have a much higher risk of death from pump failure; ICD simply converts their mode of death from VT to pump failure. If class IV patients are ambulatory and have an indication for BiV pacemaker, the combo BiV/ICD is indicated and improves their mortality as well as their quality of life.150 Similarly to class IV patients, patients with advanced CKD (≥ stage 4 in MADIT 2 trial) or those with many comorbidities (eg, age>70 with multiple HF hospitalizations) do not clearly benefit from ICD; ICD simply converts their death modality.151

B. Biventricular (BiV) pacemaker = cardiac resynchronization therapy (CRT)

Approximately 20–30% of HF patients have QRS > 120 ms, mostly LBBB (80%). This leads to dyssynchronous contraction of the LV, which means that different LV segments contract at different times (intraventricular dyssynchrony), and the LV and RV contract at different times (interventricular dyssynchrony), leading to an ineffective ventricular effort and sequential rather than simultaneous mitral leaflets movement (→ MR). A prolonged QRS independently correlates with increased mortality and symptoms. CRT paces LV and RV at approximately the same time and restores the delayed lateral LV contraction.

CRT is indicated in patients with EF ≤ 35% + QRS ≥ 150 ms (mainly LBBB morphology or RV-paced rhythm) + functional class II, III, or ambulatory IV despite an adequate medical regimen, whether they are in sinus rhythm or AF. These same patients have an indication for a CRT/ICD combined device. Patients with LBBB but a QRS 130–150 ms, or RBBB with a QRS > 150 ms, have a class IIa indication for CRT.8,51

Thirty percent of patients do not respond to CRT, i.e., do not achieve reverse remodeling on echo, for the following reasons:

  • Wide QRS does not always mean dyssynchrony. The wider the QRS (i.e., QRS > 150 ms), the more likely a response is seen. Echocardiographic dyssynchrony features help select patients whose QRS is 130–150 ms, according to the CARE HF trial.152
  • Ischemic cardiomyopathy with large infarcts is poorly responsive to CRT. Irreversible scars respond less well to CRT than non-ischemic cardiomyopathy.
  • LV lead placement may not be optimal (apex rather than the basal lateral wall).
  • HF with a pure RBBB is less likely to respond to CRT, especially if the RV is the main side that is dyssynchronous and delayed. RBBB, per se, is associated with at least the same negative prognosis as LBBB but is less responsive to CRT;153 this partly explains why, in CRT registries, RBBB is associated with a higher mortality than LBBB. Patients who, in addition to RBBB, have LV dyssynchrony and slow left bundle conduction may respond to CRT (RBBB would be > 150 ms or echocardiographic features of LV dyssynchrony would be present).
  • Echocardiographic evaluation of dyssynchrony is likely helpful in selecting patients with QRS 130–150 ms and those with RBBB. Conversely, it is not helpful for QRS <130 ms. Approximately 30% of HF patients with QRS <130 ms have dyssynchrony on echo; however, CRT is not beneficial in these patients.154
Nov 27, 2022 | Posted by in CARDIOLOGY | Comments Off on Heart Failure

Full access? Get Clinical Tree

Get Clinical Tree app for offline access