Indications and patient selection


Heart failure (HF) is a clinical syndrome characterized by breathlessness and fatigue. Hemodynamically, HF is caused by the inability to generate adequate cardiac output to meet end organ and peripheral perfusion demands, or to do so only at the expense of elevated cardiac filling pressure.

The diagnosis of HF is made on the basis of a constellation of clinical signs and symptoms that is supported by evidence of inadequate perfusion and/or elevated filling pressure. History and physical examination have a modest sensitivity for detection of elevated filling pressure, but poor sensitivity to detect systolic dysfunction or chamber remodeling.

Echocardiography is a valuable tool that provides critical information regarding cardiovascular function, morphology, and hemodynamic status, which can both enhance the diagnosis of HF and also be useful to guide therapeutic decision making when used in the proper context. This chapter will focus on the use of echocardiography for the evaluation and management of HF.

Continue Reading

Echocardiography: principles and equipment

Ultrasound imaging uses sound waves created by vibrating piezoelectric crystals within the ultrasound transducer that both send and receive signals. Sound waves are generated at a frequency beyond the range of detection of the human ear, >20,000 cycles/second, or Hertz (Hz).

Image formation is based upon the return of sound waves back to the transducer following reflection off various tissues. The transducer transforms the mechanical sound wave into an electrical signal, which is then converted into a digital image by software onboard the ultrasound scanner.

Transthoracic echocardiography (TTE) represents the foundation of modern echocardiographic imaging and uses a hand-held ultrasound transducer positioned in standard transthoracic locations (windows) during imaging. Standard TTE transducers emit at a frequency range of 1 to 5 million Hz (MHz).

Imaging is performed predominantly in the left lateral decubitus position to bring the heart as close to the transducer as possible, although certain structures are best imaged when the patient is supine. A standard comprehensive TTE takes approximately 30 to 60 minutes to perform, depending on the findings and technical difficulty of the study.

Imaging modalities

Two-dimensional (2-D) echocardiography produces a cine imaging loop of a cardiac structure in a tomographic imaging plane (Figure 1). The standard acquisition ‘clip’ is 3 to 4 cardiac cycles; however, longer loops or still frame images can also be acquired.

The depth and width of the imaging sector can be adjusted based on the desired structure being imaged, but increasing the depth or width of a given sector comes at the expense of temporal resolution.

M (motion)-mode is one of the oldest modalities of echocardiography (Figure 2). A one-dimensional ultrasound beam records tissue movement as it passes through a cardiac structure. M-mode has the highest temporal resolution of all echocardiography modalities (>1,000 MHz), allowing for identification of subtle or fine movements of small or fast moving structures, such as endocardium or cardiac valves that are difficult to appreciate with the naked eye.

Doppler imaging forms the basis of the majority of cardiac hemodynamic assessment by echocardiography. Doppler echocardiography relies on the change of frequency of a returning ultrasound signal occurring when a moving reflector (such as red blood cells) is imaged (the Doppler effect).

The prime determinants of the recorded Doppler signal are the velocity and direction of the moving structure being imaged. The key principle of Doppler imaging is that the measured velocity of blood flow reflects the relative pressure gradient between the two chambers through which flow is traveling, so higher velocity is reflective of a higher pressure gradient across the region of interest.

The different types of Doppler imaging include: continuous wave (CW), pulsed wave (PW), color flow, and tissue Doppler imaging (TDI). CW Doppler requires two crystals, one continuously emitting and the other receiving. CW Doppler allows for the detection of high velocity signals, but has range ambiguity.

Pulsed wave (PW) Doppler involves fixed interval pulse firing from the transducer, with periods of signal reception in between pulses. PW Doppler permits sampling of velocities from specific regions of interest (sample volumes). PW Doppler can only detect flow velocities up to an upper limit (frequency shift), referred to as the Nyquist limit.

Color flow Doppler imaging is a type of PW imaging that assigns a color coding to flow, based on velocity and direction (orange/red for flow towards the transducer, blue for flow away, with a brighter shade indicating higher velocity). Because it uses pulsed signals, color flow imaging also has a Nyquist limit.

TDI is applied to tissue movement instead of blood flow. TDI filters out higher velocity signals so as not to be contaminated by blood flow adjacent to the tissue region being sampled.

Details of how the procedure is performed

Echocardiographic assessment of cardiac structure and function

Chamber and vascular dimensions

Evaluation of chamber size is an important function of echocardiography. Short and long axis measurements of the left ventricle (LV) are performed, indexed to body surface area and compared to standardized population based measurements. Chamber size has important clues to etiology.

HF with reduced ejection fraction (HFrEF) is typically characterized by eccentric LV remodeling, meaning that the chamber becomes dilated. The extent of LV dilatation is a robust marker of disease severity, and therapies that reduce LV size in HFrEF (a process termed reverse remodeling) improve morbidity and mortality.

Thus serial evaluation of LV dimension provides important prognostic information regarding response to treatment in HFrEF. In contrast, patients with HF and preserved EF (HFpEF) typically display normal ventricular dimension.

Measurements of LV size are routinely performed using 2-D images of the long and short axis at end-systole and end-diastole. M-mode measurements of LV size are also typically performed and are regarded as more accurate due to the superior temporal resolution of M-mode (Figure 2).

However, proper orientation depends on the ultrasound beam being perpendicular to the imaging axis of the chamber, which is frequently guided by the corresponding 2-D image. LV volume is also frequently measured, especially when the LV is dilated and/or the systolic function is reduced.

Endocardial border tracings are performed in two orthogonal apical long axis 2-D images at end-systole and end-diastole to provide a ‘biplane’ measurement of LV volume. Volume is calculated through the summation of sequentially stacked elliptical disks, a technique referred to as the Simpson method.

Measurements of right ventricular (RV) chamber size have also been proposed, though the complex geometry of the RV makes precise quantification more difficult. Measurements are performed in the long axis view from the apical imaging window, with the imaging sector centered on the RV.

Two-dimensional RV volume measurements should not be performed due to the asymmetric shape of the RV. RV enlargement occurs with pulmonary hypertension in HFrEF and HFpEF and is also a marker for more advanced disease.

Ventricular function has important effects on atrial size and function. Any chronic increase in LV filling pressure may result in left atrial (LA) enlargement. This may be seen in HFrEF, HFpEF, or valvular disease (particularly mitral valve disease). Similar to the LV, greater LA enlargement is a marker for the worst outcome in HF and non-HF populations.

It is also useful to assess LA size when contemplating the treatment of atrial fibrillation, with greater LA remodeling suggesting less of a likelihood to achieve durable restoration of sinus rhythm. Long and short axis measurements of the LA are performed at end-systole.

Measurement of LA volume is performed using the area-length method by tracing the endocardial border in two orthogonal apical long axis 2-D images at end-systole, and then indexed to body surface area. Right atrial (RA) size is less well studied but may increase with RV failure, particularly in patients with HF complicated by or caused by pulmonary hypertension.

Assessment of ventricular systolic function

Ejection fraction (EF), defined as the stroke volume divided by the LV end-diastolic volume, is the most commonly used measure of LV systolic function in practice (Table 1). LVEF has been proven in multiple clinical trials to be one of the most important predictors of prognosis in HF patients.

However, many fail to appreciate that EF is highly afterload sensitive. For example, a patient with HFrEF and an LVEF 35% who is given an acute dose of nitroprusside may enjoy an increase in LVEF to 45% with no change contractility.

Conversely, a patient with HF and severe mitral regurgitation (MR) may have an LVEF of 50% prior to valve surgery, that drops to 35% with the increase in afterload induced by restoring mitral valve competence. In HFrEF, the LVEF is typically low because the denominator of the EF equation (LV end-diastolic volume) is increased, rather than the numerator (stroke volume) being low, at least until the advanced stages of HF.

Calculation of LVEF can be performed by multiple methods. The percentage change in squared LV short axis dimensions from end-diastole to end-systole is a frequently used method. This can be performed using 2-D or M-mode based measurements.

Fractional shortening (FS) is a related parameter of systolic function based on the percentage change in the length of short axis dimensions during systole, and is less frequently used in adult echocardiography. FS can be related to measures of afterload (wall stress) to derive more robust, load-independent measures of LV contractility, but this is used predominantly in research rather than clinical practice. Other load-independent measures of LV contractility exist but are also seldom used in clinical practice.

The Simpson biplane method, although not without limitations (discussed below), is generally regarded as one of the more reliable echocardiographic techniques for calculation of LVEF. The stroke volume is determined by the difference between calculated end-diastolic and end-systolic volumes, which is then divided by the end-diastolic volume.

This method should be performed during the initial TTE for all patients with newly discovered LV systolic dysfunction. The subjective, qualitative interpretation of LVEF (‘eyeball’ method) is used by all echocardiographers as a reference guide for the accuracy of the calculated LVEF, and correlates closely with quantitative methods by experienced echocardiologists.

Because 2-D RV volume measurements are unreliable, RVEF is rarely used for quantifying systolic function. Other quantitative measurements of RV systolic function include fractional area change (FAC), which is calculated using end-diastolic and end-systolic areas measured from long axis images of the RV from an apical imaging window.

Multiple nonvolumetric measures of systolic function also exist. Examples include the rate of pressure rise in the RV (dP/dt), the index of myocardial performance (IMP, or Tei index), tricuspid annular plane systolic excursion (TAPSE), and TDI-based parameters, such as the peak systolic velocity of the tricuspid annulus and the rate of myocardial acceleration during isovolumic contraction.

In pulmonary arterial hypertension, RV function is the most potent predictor of outcome, and TAPSE has emerged as one of the best measures in this regard. RV function is increasingly being recognized for its important implications for outcome in advanced HF, and the assessment of RV function is an area of active research in HF. The degree of RV systolic function also has important implications for the likelihood of good outcome after left ventricular assist device implantation in HFrEF.

Assessment of LV diastolic function

The assessment of LV diastolic function is an integral component of the echocardiographic evaluation of patients with suspected or confirmed HF. This is particularly true of patients presenting with dyspnea or other symptoms of HF who have a normal LVEF. It is now recognized that patients with significant systolic LV dysfunction invariably have some degree of diastolic impairment.

Closely related to the assessment of diastolic function is the estimation of LV filling pressure. Early studies showed moderate correlations between echocardiographic estimates of filling pressure and gold standard catheterization data; but more recent studies have raised questions regarding the veracity of these estimates.

PW Doppler of the mitral valve inflow represents an important component of LV diastolic function assessment. Primary measurements include the peak early diastolic filling velocity (E-wave), the peak late diastolic filling velocity secondary to atrial contraction (A-wave), the E/A ratio, and the deceleration time (DT) of the E-wave.

The E/A ratio is a key parameter used for grading of diastolic dysfunction (Figure 3). With impaired LV relaxation, the E-wave duration becomes prolonged (reflected by a prolonged DT) and the peak velocity (E) becomes reduced. This describes grade I diastolic dysfunction, when LV filling pressure is believed to remain at near normal levels and atrial contraction becomes responsible for a greater amount of diastolic filling.

With worsening diastolic function, LV filling pressure and corresponding LA pressure begin to rise, increasing the E-wave velocity and E/A ratio while simultaneously reducing the DT. This leads to the ‘pseudo-normalized’ mitral inflow pattern characteristic of grade II diastolic dysfunction.

This pattern is also characteristic of normal diastolic function, and differentiation between the two relies on additional diastolic function parameters (including evaluation of mitral inflow during Valsalva maneuver) and ancillary echocardiographic findings (particularly TDI measures; see later in this section). A further increase in filling pressure leads to an even higher E-wave velocity and E/A ratio, and a shorter DT, leading to the so-called ‘restrictive physiology’ of grade III (reversible) and grade IV (irreversible) diastolic dysfunction.

The other key parameter involved in the assessment of LV diastolic function is the peak early diastolic filling velocity of the mitral annulus (e’), which is measured with TDI (Figure 3). The e’ can be measured at the septal mitral annulus, the lateral mitral annulus, and/or both to produce an average of the two.

Prior studies have shown that e’ correlates modestly with invasive measures of LV relaxation, at least in patients with diastolic dysfunction. The e’ velocity falls with worsening LV relaxation impairment. Therefore, an abnormal e’ velocity suggests diastolic dysfunction.

Because the mitral Doppler E velocity varies directly with preload and TDI e’ estimates relaxation with less load-dependence, the E/e’ ratio has been proposed and demonstrated to represent a reasonable marker of LV filling pressures. A septal E/e’ ≥15, a lateral E/e’ ≥12, or an averaged E/e’ ≥13 have been used by some authors to indicate increased filling pressure.

Elevated E/e’ (>15) has been proposed by the European Society of Cardiology as being sufficient stand-alone evidence to make the diagnosis of HFpEF. However, recent studies have questioned the utility of E/e’, showing that it may not track with filling pressure, and many patients with HFpEF may lack elevated E/e’ while patients without HFpEF may have elevated E/e’.

A recent study found that while E/e’ correlated with LV filling pressure, the addition of E/e’ data to careful physical examination did not enhance assessment of left heart filling pressure. Like most diagnostic tests, E/e’ is probably most useful when it is very high or very low, and intermediate values are less helpful in establishing volume status.

We would not recommend relying on elevated E/e’ as stand-alone evidence of HFpEF, but it is a useful marker and when other confirmatory data are present, it increases the probability of disease.

There are a number of additional parameters used for the echocardiographic evaluation of LV diastolic function and filling pressure. These include: the pulmonary vein PW Doppler signal (including the ratio of systolic to diastolic pulmonary venous flow velocity, and the difference between pulmonary vein atrial flow reversal duration and mitral inflow A-wave duration), isovolumic relaxation time, mitral inflow propagation time, and estimated pulmonary artery pressure, among others.

These parameters are used predominantly as supportive data when the principle parameters present indeterminate or contradictory information. Table 2 describes the various echocardiographic indicators that are often used to estimate volume status.

Assessment of other important echocardiographic findings in HF

Cardiac output can be measured by echocardiography. This is because the flow volume passing through a given region can be calculated as the velocity time integral (VTI) of the Doppler signal, multiplied by the cross-sectional area of the region.

Placement of the PW Doppler sample volume in the LV outflow tract (LVOT) and measurement of the LVOT diameter provides the data required for calculation of LV stroke volume, from which cardiac output and cardiac index (normal range 2.5-4.0 L/min/m2) can be determined (Figure 4).

LV mass can be determined from echocardiography, based on an empirically derived formula from Devereaux and colleagues. LV mass varies directly with body size and changes with composition, and is therefore typically scaled to body surface area (gm/m2) or to allometric powers of height (gm/ht2.7).

LV hypertrophy (LVH) is quantified if the LV mass exceeds threshold values determined in population-based samples. LV hypertrophy may be eccentric (associated with chamber dilatation) or concentric (with normal chamber size). This is usually determined by measuring the relative wall thickness (defined as posterior wall thickness*2/LV end-diastolic dimension). If the relative wall thickness is increased but there is not LVH, this is termed LV concentric remodeling.

Patients with HFrEF typically (though not invariably) display eccentric LVH, and 50% to 60% of patients with HFpEF display either concentric LVH or concentric remodeling. The presence and extent of LVH has recently been shown to predict outcome in HFpEF. Regression of LVH is an important goal of antihypertensive treatment and is associated with improved morbidity and mortality, and lower risk of HF.

RA pressure can be estimated by imaging of the inferior vena cava (IVC). The IVC is typically imaged from the subcostal imaging window, where its junction with the RA can be visualized. Under normal loading conditions, the IVC should collapse by >50% with forced inspiration.

Dilation of the IVC and reduced inspiratory collapse are signs of elevated RA pressure, and IVC imaging provides useful confirmatory evidence (along with jugular distention on examination) of the presence or absence of systemic venous congestion. There is data suggesting that greater IVC distention is associated with a worse outcome in HF and non-HF populations, and reversal of IVC dilatation with therapy (e.g., diuretics) may enhance confidence that a euvolemic state has been achieved.

Visualization of the hepatic vein is typically performed during subcostal imaging as well. Hepatic vein dilation often accompanies IVC dilation. In addition, PW Doppler imaging of the hepatic vein can demonstrate blunting of systolic forward flow, also suggesting elevated RV filling pressure.

Systolic flow reversals in the hepatic vein are often seen with severe tricuspid valve regurgitation (in which case the Doppler profile is typically late peaking), or with severely elevated RV filling pressure due to RV failure (which causes an early peaking Doppler profile).

Pulmonary artery pressurecan be estimated using the peak end-expiratory CW Doppler velocity of tricuspid valve regurgitation. Using the modified Bernoulli equation (pressure = 4*velocity2), this velocity can be converted into a pressure gradient that when added to the estimated RA pressure provides an approximation of right ventricular systolic pressure (RVSP, Figure 5).

The overall correlation between echo-estimated RVSP and catheterization measurements is reasonable (r~0.7) but the disagreement may be substantial in many cases. Pulmonary hypertension (PH) is a catheterization-based diagnosis, but its presence and severity can be suggested by echo-estimated RVSP (Table 3).

The development of PH in HF patients is an important milestone in disease progression, typically indicating chronically elevated LV filling pressure, often with secondary elevation in pulmonary vascular resistance. The extent of RVSP elevation may serve as another indirect surrogate for left heart filling pressure in HF, because the pulmonary capillary wedge pressure adds to the RVSP in series rather than in parallel, as part of the pulmonary vascular circuit.

The presence of PH is a potent predictor of a worse outcome in both HFrEF and HFpEF, and recent studies have evaluated the role of treating PH in HF patients, though large scale trials are not yet available.

Patients with HF often have concurrent valvular heart disease, which can be either a consequence of HF, a cause of HF, or both. For example, patients with significant biventricular systolic dysfunction often have some degree of both mitral (Figure 6) and tricuspid valve regurgitation (Figure 5) that is typically functional in nature (usually due to chamber enlargement and hypokinesis rather than organic valve disease). Mitral regurgitation in HFrEF is an important determinant of PH.

Echocardiography is the imaging standard for the quantification of valvular lesion severity, and for determining the etiology of the lesion. In addition to annular dilatation, chamber enlargement can result in leaflet tethering causing atrioventricular valve regurgitation.

The degree of functional mitral or tricuspid insufficiency varies as a function of volume status and can be used as other indirect evidence of decompensation. Indeed, diuretics and vasodilators are well known to reduce the severity of mitral regurgitation in HFrEF. Other common causes of valvular dysfunction in HF include tricuspid valve regurgitation secondary to leaflet impingement by device leads and mitral valve regurgitation secondary to ischemic mitral valve apparatus.

Pericardial disease may present with signs and symptoms of HF, particularly with evidence of right heart failure (jugular distention, edema, and ascites). Pericardial effusion can develop as a consequence of volume overload (Figure 7) and is often seen in right heart failure due to pulmonary hypertension, where its presence carries adverse prognostic impact. Acute inflammatory causes of HF such as myocarditis can often have pericardial involvement, leading to pericarditis, effusion, and/or a constrictive hemodynamic pattern.

Etiology specific echocardiographic findings in HF

Causes of HF with reduced EF

1. Ischemic cardiomyopathy – Coronary disease is the leading cause of HFrEF in the United States, accounting for two thirds of cases. The hallmark of ischemic cardiomyopathy is the presence of regional wall motion abnormalities (RWMAs) in the distribution of one or more major epicardial coronary artery perfusion territories on 2-D imaging.

However, patients with severe double- or triple-vessel disease may have global systolic dysfunction, with or without regional variation, and many patients with nonischemic cardiomyopathy may also have RWMAs. Prior transmural myocardial infarction (MI) often has a distinctive appearance, with the infarcted region appearing thinned and echodense (‘bright’ or ‘scarred’). In addition, over time, aneurysms (or even pseudoaneurysms) can form in infarcted territory, with a commonly involved region being the LV apex following a left anterior descending coronary artery territory MI.

It is important to note that echocardiography cannot distinguish between viable and nonviable myocardium in akinetic regions. Such a distinction is important for patients for whom revascularization therapy is being considered, and would require performance of a dobutamine stress echocardiogram (see Advanced echocardiographic modalities of HF evaluation section.

Another important finding of ischemic cardiomyopathy is mitral valve regurgitation occurring due to ischemia or infarction of the papillary muscle(s) or surrounding myocardium. The posteromedial papillary muscle is more commonly affected, owing to its single source of perfusion (typically from the right coronary artery), whereas the anterolateral papillary muscle is less vulnerable due to its dual vessel blood supply.

2. Dilated cardiomyopathy – The combination of LV dilatation and global hypokinesis suggests a dilated cardiomyopathy (Figure 1). Often there is some degree of regional variability of LV contractile function; however, all myocardial segments appear hypokinetic.

RV involvement is variable, ranging from normal to a severely reduced systolic function. Valvular lesions are also variable; however, some degree of functional mitral valve regurgitation is common, often in the absence of a discernible murmur.

3. Myocarditis and inflammatory cardiomyopathies – Depending on the extent of inflammation, acute myocarditis can present as global systolic dysfunction, or with RWMAs. When the latter occurs, they are typically not located in the distribution of a coronary artery perfusion territory.

Acute myocarditis, regardless of the cause, typically presents with normal LV size, as the ventricle has not had time to remodel during the inflammatory insult. Edema and necrosis in myocarditis may cause an increased wall thickness on 2-D imaging. If the myocardial injury is significant and sustained, eccentric ventricular remodeling will occur with time.

Certain types of connective tissue diseases may lead to valvular involvement, such as the immune complex (nonbacterial) verrucous vegetations associated with systemic lupus erythematosus (‘Libman-Sacks vegetations’). Such valvular involvement has a tendency to cause regurgitant rather than stenotic lesions, and may progress to require surgical repair or replacement.

Pericardial involvement can also occur. Certain collagen vascular disorders and vasculitides, such as Kawasaki’s disease and Takayasu’s arteritis, are also associated with vascular dilatation, including the aorta and/or coronary arteries. Imaging of the ascending aorta, aortic arch, and upper abdominal aorta should be performed in all patients when such diagnoses are suspected.

4. Toxic cardiomyopathies – Most toxic cardiomyopathies have a similar appearance to dilated cardiomyopathy. A globally hypokinetic LV with varying degrees of dilatation is the typical 2-D imaging appearance. Examples include alcohol cardiomyopathy, chemotherapy-induced cardiomyopathy (such as occurring with anthracycline, trastuzumab, and cyclophosphamide therapy), certain trace elements (such as cobalt and arsenic), and others.

5. LV noncompaction (LVNC) – Failure or arrest of myocardial compaction during fetal development will result in LVNC. The myocardium appears echocardiographically as two distinct layers, an outer ‘compacted’ layer (normal in appearance), and an inner ‘noncompacted’ layer. The noncompacted layer appears spongy and trabeculated, and deep intratrabecular recesses can be appreciated with cavitary blood flowing within the recesses.

The most commonly used echocardiographic criteria for the diagnosis of LVNC is an end-systolic ratio of noncompacted to compacted myocardium of 2:1 in a short axis image. The distribution of noncompaction is most often the LV apical, and inferior and lateral walls; however, other distributions have been reported.

Global systolic dysfunction and chamber dilatation are not uncommon. Many cardiomyopathies are also associated with increased trabeculations, and it is not uncommon for the diagnosis of LVNC to be erroneously made. LVNC is typically genetic, and first-degree relatives should undergo screening echocardiography.

6. Others – There are multiple other types of cardiomyopathy that present with LV dilatation and global systolic dysfunction, and these are indistinguishable from each other by echocardiography. Examples include peripartum cardiomyopathy, hereditary hemochromatosis cardiomyopathy, tachycardia-induced cardiomyopathy, and certain inherited conditions such as some muscular dystrophies and myotonic dystrophy.

Stress-induced cardiomyopathy (also called “apical ballooning syndrome” or “Takotsubo’s cardiomyopathy”) has a distinct appearance. This condition, typically caused by acute physiologic stressors, such as emotional stress, surgery, trauma, or infection, causes LV dilatation and dysfunction of the anterior wall around the apex and the mid to distal inferior wall, with preserved or hyperdynamic basal contractile function.

The echocardiographic appearance resembles that of an apical MI, but is usually more dramatic than the area subtended by a single coronary artery. Patients with severe coronary disease have been erroneously diagnosed as having apical ballooning syndrome, and exclusion of obstructive coronary artery disease is required. Rare cases of “reverse” or atypical stress-induced cardiomyopathy have been reported, where basal dilatation and dysfunction occur and apical function is preserved.

Causes of HF with normal EF

1. HF with preserved ejection fraction (HFpEF) – Half of all HF patients have preserved EF, and this form of HF is growing relative to HFrEF at a rate of 1% per year. Use of the term HFpEF is typically equated with what was formerly described as “diastolic HF”—a form of HF seen with arterial hypertension, aging, obesity, and female sex.

Patients presenting with dyspnea and fatigue that are found to have low LVEF are easily diagnosed as having HFrEF. In contrast, patients with dyspnea and normal EF may be difficult to diagnose, as the differential diagnosis is broad and includes pulmonary disease, obesity, deconditioning, anemia, and other entities in addition to HFpEF.

Echocardiography can be extremely useful in these circumstances. HFpEF is typically associated with concentric LVH or concentric remodeling and LA enlargement, though neither is in itself necessary for a diagnosis.

Doppler findings of diastolic dysfunction increase the probability of HFpEF, particularly when they are very abnormal (i.e., very high E/A ration, short deceleration time, very high E/e’ ratio). Mild diastolic dysfunction (e.g., grade I) is very common with normal aging and is not helpful to rule in HFpEF as the diagnosis.

The presence of elevated RVSP provides an additional positive finding that may suggest elevation of LV filling pressure, particularly when observed in an older aged patient. Table 4 shows typical echocardiographic findings seen in HFpEF; the more that are present, the more likely the diagnosis of HFpEF is correct.

There are a number of other specific causes of HF that present with normal LVEF that are distinct from what we define as HFpEF above, and these are discussed below. Most of these entities have their own individualized treatment, so they are important to distinguish from HFpEF (which has no proven disease-specific treatment); echocardiography is incredibly useful in this regard.

2. Hypertrophic cardiomyopathy (HCM) – HCM is defined as the presence of LV hypertrophy (>15 mm wall thickness) in the absence of an identifiable cause, such as chronic hypertension or aortic valve stenosis, and is most often diagnosed by echocardiography. The LV cavity is normal size or smaller.

Different variations of LV wall hypertrophy can occur, with common patterns being asymmetric septal, midcavity, or apical hypertrophy. RV involvement is rare. An important feature of some patients with HCM is a dynamic outflow obstruction. Careful interrogation for the presence of intracavitary obstruction in a patient with recognized HCM is of vital importance.

The LVOT is the most common location for obstruction, resulting in a characteristic dynamic “reverse-dagger” shaped CW Doppler signal. LVOT obstruction is accompanied by systolic anterior motion (SAM) of the mitral valve, involving either the anterior leaflet or both mitral valve leaflets, as well as the valvular apparatus to some extent.

The regurgitant orifice is angled posteriorly by the anterior forces of SAM, creating a posteriorly directed regurgitant jet whose severity correlates directly with the degree of obstruction. The TDI E/e’ ratio has been found to correlate rather poorly with filling pressure in patients with HCM, and it should probably not be relied upon to estimate volume status in these patients.

Systolic gradients may be present at rest or require dynamic maneuvers to elicit them. Techniques to provoke an obstructive gradient include preload reducing maneuvers such as Valsalva, or pharmacologic agents, such as the inhaled vasodilator amyl nitrate. The presence of a symptomatic obstructive gradient is an indication for treatment, initially with pharmacologic agents and subsequently with invasive measures, such as septal ablation or surgical myectomy.

3. Restrictive cardiomyopathy (RCM) – The classic 2-D echocardiographic appearance of idiopathic RCM is a normal or small-sized LV cavity, with or without increased wall thickness (depending upon the cause) with severely dilated atria (Figure 8). Unlike the other cardiomyopathies that derive their names from morphologic criteria, RCM is based upon the restrictive physiology present, characterized by profound diastolic dysfunction (usually ≥ grade 3 diastolic dysfunction).

4. Infiltrative cardiomyopathies – A subset of RCM includes the infiltrative cardiomyopathies, many of which have a similar 2-D echocardiographic appearance and Doppler hemodynamic profile. This includes thick LV walls, biatrial enlargement, and diastolic dysfunction.

For this reason, echocardiography is rarely used as the sole diagnostic investigation for infiltrative cardiomyopathies, and endomyocardial biopsy is frequently required for confirmation. Within this category, cardiac amyloidosis is the most common.

Patients with cardiac amyloidosis have characteristically thickened heart valves, RV wall thickening, and often a small pericardial effusion (Figure 9). Cardiac sarcoidosis has a more variable appearance echocardiographically, and may demonstrate significant systolic dysfunction in addition to diastolic abnormalities. LV aneurysms are not uncommon.

5. Radiation cardiomyopathy – Patients with radiation induced cardiac toxicity often have a variety of echocardiographic findings. This can include some degree of both systolic and diastolic dysfunction.

Perhaps one of the most characteristic findings is the presence of diffuse calcification of many endocardial and vascular structures, including valvular leaflets and apparatus, pericardium, myocardium, and the walls of the aorta. As a result, patients with radiation-induced cardiac toxicity are vulnerable to several types of heart disease in addition to HF.

6. Other – Other causes of HF occurring in the absence of a significantly reduced LVEF may include: hypertensive cardiomyopathy, cardiomyopathy of chronic kidney disease, diabetic cardiomyopathy, all of which may have a very similar echocardiographic appearance to HFpEF. Patients with eosinophil-mediated heart disease typically develop predominantly apical biventricular immune complex deposition leading to an endocardial myocarditis.

Over time this leads to apical thrombus formation and eventual fibrotic endomyocardial scarring that causes a restrictive hemodynamic physiology. Another characteristic feature includes fibrotic entrapment of the mitral valve posterior leaflet and apparatus leading to mitral valve regurgitation.

Causes of predominantly right heart failure

1. Arrhythmogenic right ventricular cardiomyopathy (ARVC) – The diagnosis of ARVC can be made by tissue confirmation, or satisfaction of International Task Force Criteria, which includes echocardiographic criteria.

Major echocardiographic criteria are the presence of regional RV akinesis, dyskinesia or aneurysm, and either RV outflow tract (RVOT) dilatation (long axis diameter ≥ 32 mm or short axis diameter ≥ 36 mm) or an FAC ≤ 33%. Minor criteria include regional RV akinesia or dyskinesia and either an RVOT diameter of 29 to 31 mm in the long axis or 32 to 35 mm in the short axis, or an FAC of 34% to 40%.

2. Pulmonary artery hypertension (PAH)- Patients with PAH have RV dilatation and hypokinesis, along with RA enlargement. Patients may initially develop RV wall hypertrophy; however, over time, the RV will dilate, often severely.

LV size is typically normal or reduced. While LV contractile function is typically normal, paradoxical septal motion secondary to RV pressure overload may reduce the calculated LVEF. The ventricular septum often appears flattened or “D-shaped.”

Noninvasive estimation of pulmonary artery pressure by TTE has become an important component of the serial evaluation of patients with PAH (Figure 5). However, tricuspid valve regurgitation velocity may significantly underestimate RVSP in the setting of severe RV hypokinesia.

PW Doppler measurement of the RVOT typically demonstrates a notched Doppler signal due to midsystolic closure of the pulmonary valve. Pulmonary valve regurgitation may be present, and the CW Doppler velocity at end-diastole provides an estimate of diastolic pulmonary artery pressure.

PAH (i.e., PH due to pulmonary arterial disease alone) may be difficult to distinguish from HFpEF with PH; but older patient age, greater LA enlargement, LVH, and more severe diastolic dysfunction favor HFpEF, whereas RA enlargement and RVH favor PAH.

High output HF

High output HF occurs in the setting of an increased cardiac output that is typically caused by excessively low systemic vascular resistance. Examples include severe anemia, hyperthyroidism, arteriovenous fistula, Paget’s disease, and chronic liver disease. TTE will demonstrate a normal or hyperdynamic LVEF with an elevated cardiac output based upon Doppler evaluation.

It is not uncommon to identify patients who have previously been diagnosed as having “HFpEF” or “RCM” who are found to have an inordinately high cardiac output on echocardiography. An aggressive search should be undertaken looking for the reversible causes listed above; as this can effectively “cure” the HF (even if the LVEF has already begun to drop).

Correlation of physical examination findings with echocardiographic findings

Auscultation – One of the most common indications for performance of a TTE is the presence of a new murmur on physical examination. Echocardiography is the imaging standard for the evaluation of valvular heart disease, and is very sensitive to the detection of even trivial or mild valvular lesions that are inaudible on physical examination.

Echocardiography can also correlate well with common cardiac heart sounds heard in HF. For example, an S4 gallop rhythm correlates with an increased A-wave velocity as seen in grade I-II diastolic dysfunction, while an S3 correlates with an increased E-wave velocity as typically seen in patients with grade II or higher diastolic dysfunction (Figure 3).

Precordial pulsation- Patients with LV chamber enlargement will demonstrate an inferiorly/laterally displaced and diffuse apical impulse. A focal, sustained impulse typically corresponds with LV hypertrophy on echocardiogram, while a hyperdynamic impulse can be appreciated in patients with an increased stroke volume.

Echocardiography can confirm the presence of HOCM in patients suspected of this diagnosis on the basis of a double or triple LV apical impulse. Similarly, echocardiography can confirm the presence of RV hypertrophy in patients with a sustained precordial lift, while a hyperdynamic precordium may correspond with RV enlargement.

Patients with a murmur thrill (grade IV or higher) secondary to a significant valvular lesion should undergo echocardiography to evaluate the etiology and quantify the severity of the lesion. Lastly, the palpable P2 felt in the left second intercostal space in patients with pulmonary artery enlargement due to PAH will correspond to findings of significant PAH on echocardiography.

Other physical findings in HF – Patients with severe tricuspid valve regurgitation typically have a prominent ‘cv’ wave upon inspection of the internal jugular vein, a finding that correlates with systolic hepatic vein flow reversals on PW Doppler imaging. This is comparable to the systolic blunting or even reversal of pulmonary vein Doppler flow in patients with severe mitral valve regurgitation (although no direct corresponding physical examination finding for this exists).

Advanced echocardiographic modalities for HF evaluation

Stress echocardiography – The most common indication for stress echocardiography is the identification of stress-induced RWMAs, signifying the presence of obstructive coronary artery disease. In addition to its value for identifying an ischemic etiology in patients with newly discovered LV systolic dysfunction, stress echocardiography has additional value for the evaluation of HF.

In patients known to have an ischemic cardiomyopathy, the use of the dobutamine can aid in identifying regions of myocardium that are still viable. The so-called “biphasic” response describes how the contractile function of segments that are akinetic or severely hypokinetic at rest improve with low dose (2.5 to 5 mcg/kg/minute) dobutamine, but then returns to baseline or even worsens with higher doses (secondary to ischemia). This finding is very sensitive for the presence of myocardial viability and is highly accurate at predicting improvement in LV function following revascularization.

Exercise echocardiography is increasingly used to determine whether exertional dyspnea is due to cardiac sources, but data validating the efficacy of this approach is currently lacking and this remains an area of active study. Based upon the available evidence, we do not advocate using exercise Doppler measures to rule in or exclude HF as a source of dyspnea.

Contrast echocardiography – Ultrasound contrast uses a perflutren (a perfluoropropane gas) solution of microbubbles that resonate at harmonic frequencies when stimulated by ultrasound waves. The solution is intravenously injected and the microbubbles are small and durable enough to traverse the pulmonary circulation and opacify the LV cavity, allowing for improved endocardial border visualization.

In addition to use in patients with suboptimal endocardial border definition by 2-D imaging, contrast echocardiography has other advantages for the evaluation of patients with HF. Contrast can help identify the presence of intracardiac thrombus, particularly LV apical thrombus (a finding that can be difficult to visualize clearly on 2-D imaging alone in many patients) (Figure 10).

Contrast can also help visualize the presence of an LV aneurysm or pseudoaneurysm. And contrast echocardiography is of particular value in patients with suspected LVNC, as it can identify intracavitary blood flow between the deep intratrabecular recesses that characterize this disorder, in addition to aiding visualization of possible thrombus within those recesses.

Transesophageal echocardiography (TEE) – TEE usually plays a limited role in the evaluation of HF patients. TEE can be used for improved visualization of myocardial structures in patients with suboptimal TTE image quality, and this is particularly true of critically ill patients who are mechanically ventilated.

More commonly, TEE is used to rule out the presence of LA appendage thrombus in patients with atrial fibrillation who are being evaluated for possible cardioversion to sinus rhythm (Figure 11). TEE also provides excellent visualization of cardiac valves and valvular apparatus, and is often used to determine the etiology and severity of valvular lesions that are poorly defined by TTE.

Three-dimensional (3-D) imaging – The use of 3-D imaging in HF is predominantly for the measurement of LV volume and EF. The ability to acquire a full volume sample in a single acquisition addresses some of the limitations of 2-D biplane volumetric assessment. Although 3-D imaging has been evaluated for predicting and measuring response to cardiac resynchronization therapy (CRT) in HF patients, this application requires further research before routine clinical use can be endorsed.

Strain imaging – Strain imaging measures the degree of relative myocardial deformation occurring throughout the cardiac cycle and can be measured with either TDI or a technique called “speckle-tracking.” The main advantage of strain is that it addresses the limitations of tethering and translational motion, which may confound the visual assessment of global and regional myocardial systolic function with TDI.

Strain imaging can also be used to evaluate diastolic function but is not as well vetted as TDI. Strain imaging has been applied to detect low-grade systolic dysfunction in several forms of HF, including amyloid cardiomyopathy and HCM. However, at this time, strain imaging is mainly used as a research tool.

HF due to pericardial disease

Pericardial diseases are readily identified by echocardiography. A significant pericardial effusion (Figure 7) can mimic the symptoms of heart failure, initially causing dyspnea, and subsequently leading to tachycardia, hypotension, shock, and death if progression to cardiac tamponade occurs.

Echocardiographic signs of cardiac tamponade include diastolic RA and RV collapse, and evidence of respiratory ventricular interdependence, such as a significant drop in mitral E-wave inflow velocity during inspiration and ventricular septal shift with inspiration.

Constrictive pericarditis is also a condition commonly confused with HF (especially right heart failure) because the signs and symptoms are identical. Echocardiography can suggest or confirm the presence of constrictive pericarditis, typically demonstrating respiratory ventricular interdependence.

Other findings include reversal of mitral annular diastolic e’ velocities by TDI, the so-called “annulus reversus” sign. Under normal circumstances, the lateral mitral annulus e’ velocity is greater (normal >10 cm/sec) than the septal annulus e’ velocity (normal >8 cm/sec).

This relationship becomes reversed in many patients with constrictive pericarditis. An additional finding is the presence of expiratory diastolic hepatic vein flow reversals on PW Doppler imaging (such a finding requires the use of a respirometer to display the phase of the respiratory cycle during imaging).

Rarely, cardiac masses large enough to cause symptomatic obstruction of flow may be detected by echocardiography .

Correlation of other investigations with echocardiographic findings

Noninvasive imaging

1. Cardiac magnetic resonance imaging (CMR) – CMR has superior spatial resolution when compared to echocardiography, and therefore has superior endocardial border definition. For this reason, volumetric evaluation of both LV and RV EF by CMR is generally more accurate than TTE.

In addition to this advantage, CMR has the added capability of delayed enhancement imaging through the use of paramagnetic contrast agents. This allows for accurate identification and volume quantification of myocardial scar and fibrosis, which is valuable for both diagnostic and prognostic purposes.

Pharmacologic stress imaging and perfusion imaging can also be performed. Disadvantages of CMR are its cost and availability in some centers, and (importantly for HF patients) CMR is contraindicated in patients with implantable cardioverter-defibrillator (ICD) due to magnetic interference with device function. Also, patients with significant renal dysfunction cannot receive MRI contrast agents due to the risk of nephrogenic systemic fibrosis.

2. Nuclear imaging – Cardiac single photon emission computed tomography (SPECT) and positron emission tomography (PET) are well validated techniques for demonstrating the presence of inducible ischemia and/or myocardial viability in patients with ischemic cardiomyopathy, and for quantification of LVEF. A disadvantage is patient exposure to ionizing radiation.

3. Cardiac computed tomography(CT) – Cardiac CT is the current noninvasive imaging standard for evaluating the patency of native coronary arteries and bypass grafts. Despite inferior temporal resolution to echocardiography and CMR, cardiac CT is rapidly developing the ability to provide functional assessment. Patient exposure to ionizing radiation is a recognized disadvantage.

Invasive evaluation

1. Cardiac catheterization – Right and left heart catheterization remains the gold standard for evaluation of the hemodynamic derangements occurring in HF, including measurement of LV and RV filling pressure, cardiac output, pulmonary artery pressure, and pulmonary and systemic vascular resistance.

Invasive study is frequently required for patients with advanced or end-stage HF who are being considered for invasive treatment options such as heart transplantation of ventricular assist device implantation. Catheterization may also be necessary when noninvasive studies are either nondiagnostic or provide discordant information.

Cardiac catheterization can be performed during bicycle exercise to identify whether exertional dyspnea is due to cardiac or noncardiac etiologies. Hybrid invasive and echocardiographic assessments may allow for detection of elevation of filling pressure and evaluation of dynamic changes in valvular disease or change in cardiac volumes during stress.

Limitations of echocardiography

Although echocardiography is an important tool for evaluation of the HF patient, it has a number of limitations that must be recognized and understood to facilitate its optimal use for patient diagnosis and management.

Image quality – Perhaps the most significant limitation of echocardiography is the potential for poor image quality to hamper optimal visualization of cardiac structures. Patients with a large body habitus or a barrel-shaped chest, and particularly those with significant pulmonary disease, are most susceptible to this problem.

While the use of contrast echocardiography can be very helpful to help overcome poor endocardial border definition, it only aids the assessment of ventricular chamber size and global and regional systolic function, and does not assist with visualization of finer structures such as valves. In patients with poor quality images, it is the responsibility of the interpreting echocardiographer to communicate in the report which data obtained during the examination may be less reliable or indeterminate, and to recommend additional investigations that may be more suitable (such as TEE, CMR, or cardiac CT).

Evaluation of LV systolic function- Although determination of LVEF is one of the primary indications for performing echocardiography, the reliability of this calculation may be compromised in certain situations. As described above, LVEF is highly load sensitive and not an ideal measure of contractile function.

The Simpson biplane method, generally regarded as the most reliable 2-D TTE method for calculating LVEF, relies on the geometric assumption of symmetrical elliptical disks stacked one on top of the other. This assumption may not reflect the true geometry of the ventricle, especially for patients with an ischemic cardiomyopathy and significant RWMAs.

Another potential problem occurs when the ultrasound transducer is not correctly positioned during imaging. The most common example is the foreshortened LV occurring when the transducer is not positioned at the true apex during apical imaging, causing the apex to appear truncated. In this circumstance, not only will calculated LVEF be incorrect, but important findings such as an apical aneurysm or a thrombus may be missed.

When calculation of LVEF is limited by such factors, echocardiographers often rely on their subjective interpretation of global LV function, or the “eye-ball” method. However, the accuracy of this is approach is highly dependent on echocardiographer experience, and significant interobserver variability may exist. The same can be said for assessment of both RMWAs and RV systolic function, both of which rely heavily on subjective interpretation.

Evaluation of LV diastolic function – Despite promising results from early validation studies, the accuracy and reproducibility of the echocardiographic evaluation of LV diastolic function and filling pressure has been questioned by more recent studies, particularly for patients with preserved LVEF. Though this topic remains controversial, clearly a careful evaluation of both diastolic function and LV filling pressure remains an important component of a comprehensive TTE examination.

In general we view diastolic indices as being useful when they are more severely deranged and when several independent indices (Table 2) show concordance. For example, a high E/A ratio with a short deceleration time, high E/e’ ratio, LA enlargement, and dilated IVC provides convincing evidence for advanced and/or decompensated HF. In contrast, an isolated E/e’ ratio of 13 without other corroborative evidence is less convincing and should not provide stand-alone evidence of HF unless other corroborating findings are present.

There are a number of circumstances where diastolic function assessment is limited by echocardiography, including prior valve surgery, significant mitral annular calcification, hypertrophic cardiomyopathy, atrial fibrillation and causes of abnormal septal motion such as left bundle branch block or ventricular pacing (on the septal mitral annulus e’ velocity).

Doppler imaging – All information obtained from Doppler imaging relies heavily on the angle of insonation, or the angle between the ultrasound imaging beam and the plane of the flow or structure being imaged. Ideally the two should be parallel; however, when the angle of insonation is increased, Doppler imaging will underestimate the true velocity.

This has important implications for many components of a comprehensive TTE examination, including the severity of valvular lesions and the estimation of cardiac output. When acquired data is unreliable for this reason, it should not be included in the TTE report.

What’s the evidence?

Ommen, SR, Nishimura, RA, Appleton, CP. “Clinical utility of Doppler echocardiography and tissue Doppler imaging in the estimation of left ventricular filling pressures: A comparative simultaneous Doppler-catheterization study”. Circulation. vol. 102. 2000. pp. 1788-94. (Details correlation of echocardiography derived measures of left ventricular relaxation and filling pressure with invasive catheter derived measurements in patients with a range of both normal and reduced left ventricular ejection fraction.)

Mullens, W, Borowski, AG, Curtin, RJ, Thomas, JD, Tang, WH. “Tissue Doppler imaging in the estimation of intracardiac filling pressure in decompensated patients with advanced systolic heart failure”. Circulation. vol. 119. 2009. pp. 62-70. (Critically evaluates the utility of Doppler echocardiography-derived measures of left ventricular filling pressure in patients with acutely decompensated heart failure and reduced ejection fraction.)

Lam, CS, Roger, VL, Rodeheffer, RJ. “Cardiac structure and ventricular-vascular function in persons with heart failure and preserved ejection fraction from Olmsted County, Minnesota”. Circulation. vol. 115. 2007. pp. 1982-90. (Reports the abnormalities in echocardiography-derived measures of ventricular function in a community cohort of patients with heart failure and preserved ejection fraction compared to healthy control subjects.)

Bhella, PS, Pacini, EL, Prasad, A. “Echocardiographic indices do not reliably track changes in left-sided filling pressure in healthy subjects or patients with heart failure with preserved ejection fraction”. Circ Cardiovasc Imaging. vol. 4. 2011. pp. 482-9. (Reports findings that question the utility of serial assessment of Doppler echocardiography-derived measures of left ventricular filling pressure compared with invasive measurements in patients with heart failure and preserved ejection fraction and healthy control subjects.)

Oh, JK, Hatle, L, Tajik, AJ, Little, WC. “Diastolic heart failure can be diagnosed by comprehensive two-dimensional and Doppler echocardiography”. J Am Coll Cardiol. vol. 47. 2006. pp. 500-6. (A review of two-dimensional and Doppler echocardiography-derived measures of left ventricular diastolic function in patients with preserved ejection fraction.)

Borlaug, BA, Nishimura, RA, Sorajja, P, Lam, CS, Redfield, MM. “Exercise hemodynamics enhance diagnosis of early heart failure with preserved ejection fraction”. Circulation: Heart Fail. vol. 3. 2010. pp. 588-95. (Describes how invasive measurement of cardiac hemodynamics during exercise can aid in the diagnosis of heart failure with preserved ejection fraction.)

Guazzi, M, Borlaug, BA. “Pulmonary hypertension due to left heart disease”. Circulation. vol. 126. 2012. pp. 975-90. (Detailed review describing the pathophysiology, clinical evaluation, management, and outcomes of pulmonary hypertension caused by left heart diseases.)

Fukuta, H, Little, WC. “The cardiac cycle and the physiologic basis of left ventricular contraction, ejection, relaxation, and filling”. Heart Fail Clin. vol. 4. 2008. pp. 1-11. (Describes the pathophysiologic basis for the development and progression of heart failure.)

Haddad, F, Hunt, SA, Rosenthal, DN, Murphy, DJ. “Right ventricular function in cardiovascular disease, part I: Anatomy, physiology, aging, and functional assessment of the right ventricle”. Circulation. vol. 117. 2008. pp. 1436-48. (Review of the anatomy, pathophysiology and clinical evaluation of the right ventricle in patients with right heart failure)

Ammash, NM, Seward, JB, Bailey, KR, Edwards, WD, Tajik, AJ. “Clinical profile and outcome of idiopathic restrictive cardiomyopathy”. Circulation. vol. 101. 2000. pp. 2490-6. (Describes the clinical and echocardiographic features and outcomes of patients with idiopathic restrictive cardiomyopathy.)