Portal de Periódicos da CAPES

Doppler Echocardiography Yields Dubious Estimates of Left Ventricular Diastolic Pressures

2009; Lippincott Williams & Wilkins; Volume: 120; Issue: 9 Linguagem: Inglês

10.1161/circulationaha.109.869628

1524-4539

Carsten Tschöpe, Walter J. Paulus,

Cardiovascular Health and Disease Prevention

HomeCirculationVol. 120, No. 9Doppler Echocardiography Yields Dubious Estimates of Left Ventricular Diastolic Pressures Free AccessArticle CommentaryPDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessArticle CommentaryPDF/EPUBDoppler Echocardiography Yields Dubious Estimates of Left Ventricular Diastolic Pressures Carsten Tschöpe, MD and Walter J. Paulus, MD, PhD Carsten TschöpeCarsten Tschöpe From the Department of Cardiology and Pneumology, Benjamin Franklin Campus, Charité University Hospital, Berlin, Germany (C.T.), and Institute for Cardiovascular Research VU, VU University Medical Center Amsterdam, Amsterdam, the Netherlands (W.J.P.). and Walter J. PaulusWalter J. Paulus From the Department of Cardiology and Pneumology, Benjamin Franklin Campus, Charité University Hospital, Berlin, Germany (C.T.), and Institute for Cardiovascular Research VU, VU University Medical Center Amsterdam, Amsterdam, the Netherlands (W.J.P.). Originally published1 Sep 2009https://doi.org/10.1161/CIRCULATIONAHA.109.869628Circulation. 2009;120:810–820Quantification of left ventricular (LV) diastolic function is necessary to diagnose heart failure (HF) when LV systolic function is normal.1–4 Furthermore, repetitive assessment of LV filling pressures is an important guide for titration of diuretic treatment and can predict survival of HF patients.5 Because of patient discomfort and the risks involved in invasive procedures, a noninvasive estimate of diastolic LV function and pressures is highly desirable. In current cardiological practice, noninvasive evaluation of diastolic LV function is based on Doppler echocardiographic visualization of LV inflow and/or LV tissue reextension. LV inflow and LV tissue reextension, however, are only indirectly related to LV filling pressures through laws of physics such as the Bernoulli principle and Laplace law. Noninvasive estimates of LV filling pressures can therefore be offset not only by limitations of the imaging technique but also by shortcomings inherent to derivation of pressures from inflow or reextension signals. As a result of these problems with noninvasive estimates of LV diastolic function and pressures, the cardiological community has witnessed over the past 20 years repetitive cycles in which a Doppler echocardiographic index was first proposed as robust and shortly thereafter discredited by contradictory evidence. The latest of such cycles involved the ratio of early transmitral velocity to tissue Doppler mitral annular early diastolic velocity (E/E′). The value of the E/E′ ratio as a reliable estimate of LV filling pressures was demonstrated in a variety of cardiac diseases6–10 and endorsed by European and American consensus statements on diastolic HF4 and diastolic LV dysfunction11 before being seriously questioned both in hypertrophic12 and dilated cardiomyopathy.13 This continuing uncertainty14 surrounding the value of noninvasive estimates of LV filling pressures and diastolic LV dysfunction asks for a reappraisal of physiological assumptions linking LV filling pressures to myocardial reextension kinetics, of pitfalls of diastolic LV dysfunction indices, and of limitations of current Doppler echocardiographic imaging techniques.Response by Little and Oh on p 820From Myocardial Lengthening to Ventricular Filling PressureThe original experimental15,16 and clinical17–19 studies establishing the E/E′ ratio as a reliable estimate of LA pressure emphasized the importance of LV relaxation kinetics for E′. These studies suggested E′ to be strongly associated with LV relaxation kinetics and minimally affected by LA pressure. Hence, because E′ corresponded with LV relaxation kinetics and E depended on both LV relaxation kinetics and LA pressure, the E/E′ ratio could serve as a reliable measure of LA pressure. In view of the recent controversy surrounding E/E′ as an estimate of LA pressure, the relative importance for early diastolic myocardial lengthening of residual LV relaxation pressure, myocardial restoring forces, and lengthening loads needs to be reassessed.Residual LV relaxation pressure after mitral valve opening has been extrapolated from an exponential curve fit to isovolumic LV pressure decay used to determine τ. By subtracting this residual "active" LV relaxation pressure from measured early diastolic LV pressure, a "passive" early diastolic LV pressure could be calculated.20 In ischemic heart disease21–23 and in diastolic HF,24 these calculations revealed substantial residual LV relaxation pressure after mitral valve opening. In isolated papillary muscle experiments with physiological sequence relaxation, residual LV relaxation forces also were observed even at a time when muscle lengthening was completed.25,26 Furthermore, in occasional patients with LV hypertrophy caused by hypertrophic cardiomyopathy or aortic stenosis, measured early diastolic LV pressure almost resembled residual LV relaxation pressure, with a continuous decline throughout early diastole and a minimum diastolic LV pressure observed just before atrial contraction (Figure 1).27 Involvement of cardiomyocyte calcium handling in this early diastolic LV pressure decline was supported by its appearance after postextrasystolic potentiation and by its disappearance after calcium channel blockers.27,28 Throughout this early diastolic LV pressure decline, the mitral valve was open with minimal LV filling (Figure 1). This suggested a near equilibrium in these patients within the LV wall between residual LV relaxation force, which opposes LV filling, and forces that promote LV filling. The latter consist of restoring forces within the cardiomyocytes resulting from end-systolic compression and of lengthening loads imposed by left atrial (LA) pressure. Taken together, these observations support residual LV relaxation pressure after mitral valve opening as important for early diastolic myocardial lengthening or for E′ but do not suggest that restoring forces and lengthening loads can be overlooked. Download figureDownload PowerPointFigure 1. Left, Continuous diastolic LV pressure decline in an occasional patient with LV hypertrophy related to aortic stenosis or hypertrophic cardiomyopathy. This continuous diastolic LV pressure decline, recorded with a catheter-tip micromanometer, resulted from substantial residual LV relaxation pressures in early and mid diastole (arrow). Right, Simultaneous LV pressure and LV cavity echocardiogram in the same patient. There was minimal LV filling during the diastolic LV pressure decline. This suggests a near equilibrium within the LV wall between residual LV relaxation force, which opposes LV filling, and lengthening load, which promotes LV filling.When peak diastolic lengthening velocity of an isolated papillary muscle strip was plotted against systolic shortening, a relation appeared between peak diastolic lengthening velocity and systolic shortening (Figure 2).29 A higher lengthening load shifted the relation upward, but a higher calcium concentration had no effect. These findings imply that during diastolic lengthening, normal cardiac muscle behaves like a spring: When the spring is more forcefully compressed during systole or when a heavier load is suspended on the spring, diastolic lengthening velocity is higher. The cardiomyocyte protein responsible for this spring-like behavior of cardiac muscle has meanwhile been identified as the giant cytoskeletal protein titin.30 Titin acts as a bidirectional spring affecting early diastolic muscle lengthening kinetics and late diastolic muscle extension. Its spring properties are altered by transcriptional (ie, isoform shifts) and posttranslational (ie, phosphorylation, oxidation) modifications.31 Higher expression of the compliant N2BA titin isoform is observed in patients with HF and reduced LV ejection fraction (HFREF)32–34 but not in patients with HF and normal LV ejection fraction (HFNEF).35 Furthermore, in both HFREF and HFNEF patients, reduced overall phosphorylation of titin36 and reduced phosphorylation of the noncompliant N2B titin isoform37 have just been reported. Lower phosphorylation of titin, especially of its noncompliant N2B isoform, stiffens its spring characteristics.38 These titin-related restoring forces within the cardiomyocyte affect LV filling kinetics, as evident from associations in HFREF patients between titin isoform shifts and E/A ratio, exercise tolerance, or symptomatic status.33 Restoring forces can therefore not be overlooked as determinants of myocardial reextension, and use of E′ as exclusively related to LV relaxation kinetics should therefore be questioned. Download figureDownload PowerPointFigure 2. Spring-like mechanics of normal cardiac muscle during diastolic lengthening. In isolated cat papillary muscle strips subjected to physiological sequence relaxation, peak diastolic lengthening velocity is related to systolic shortening. Higher lengthening loads shift the relation upward. Redrawn from Goethals et al.29Very recent in vivo experiments in instrumented anesthetized dogs39 reappraised the importance of lengthening loads arising from LA pressure after mitral valve opening. In this study, E′ measurements, which were recorded under a variety of conditions such as volume loading, dobutamine infusion, and coronary artery occlusion, showed a close relation with mitral valve opening pressure and only a modest relation with restoring forces and LV relaxation kinetics (τ). This finding again undermines the physiological assumptions underlying the E/E′ ratio as a reliable estimate of LV filling pressure.As nicely illustrated by the continuous early diastolic LV pressure decline in the occasional patient with LV hypertrophy (Figure 1), E′ is determined by a balance between forces opposing LV filling such as residual LV relaxation pressure and forces promoting LV filling such as restoring forces and external lengthening loads (Figure 3). In HF, this balance between forces shifts. Slower LV isovolumic relaxation raises residual LV relaxation pressures in early diastole; this occurs regardless of HF phenotype.24,35,40–42 Because of a larger reduction in myocardial systolic shortening in HFREF than in HFNEF,43 restoring forces will fall especially in HFREF.44 LA pressure at mitral valve opening rises, but the resultant elevation of early diastolic lengthening load will be smaller in a concentrically remodeled LV of HFNEF than in an eccentrically remodeled LV of HFREF. Moreover, a normal LVEF has recently been shown to protect cardiomyocytes against the excessive late systolic load imposed by arterial wave reflection.45 Similarly, a normal LVEF also could protect cardiomyocytes against excessive isovolumic relaxation or early diastolic loads. Download figureDownload PowerPointFigure 3. Forces acting on the cardiomyocyte during early diastolic lengthening. The kinetics of early diastolic cardiomyocyte lengthening result from a balance between residual crossbridge interaction (active force; AF), titin-mediated recoil (recoil force; RF), and lengthening load (external force; EF). AF opposes lengthening (−), and both RF and EF promote lengthening (+). In HF, this balance between forces shifts. In both HFNEF and HFREF, there is more residual crossbridge interaction. Titin-mediated recoil is better preserved in HFNEF than in HFREF because of smaller LV end-systolic volume. Lengthening loads are higher in HFREF than in HFNEF because of eccentric LV remodeling in HFREF and concentric LV remodeling in HFNEF.Because HFNEF and HFREF have unequal effects on early diastolic forces that control cardiomyocyte lengthening, it is not surprising that the reliability of E/E′ as an estimate of diastolic LV dysfunction or LV filling pressures also differs in both conditions. In 86% of patients with HFNEF, E/E′ detected diastolic LV dysfunction evident from conductance catheter pressure-volume loop analysis,46 whereas in only 53% of patients with HFREF, E/E′ correctly identified a pulmonary capillary wedge pressure (PCWP) >18 mm Hg.13 This divergence is explained by lengthening load, which is more important for early diastolic myocardial extension and E′ in HFREF. In HFREF, dividing E by E′ no longer simply corrects for residual LV relaxation pressure, and the E/E′ ratio becomes unreliable as an estimate of LV filling pressure. In HFNEF, residual LV relaxation pressure probably remains the dominant determinant of early diastolic myocardial extension and E′. Hence, in HFNEF, dividing E by E′ continues to correct for the effect of residual LV relaxation pressure on mitral E velocity, and the E/E′ ratio still yields a reliable estimate of LV filling pressures. The prominent control of early diastolic LV filling by LA pressure also explains why a simple mitral E velocity measurement also is reliable as an estimate of LV filling pressures in HFREF but correlates poorly with LV filling pressures in patients with LVEF >50%.47,48 Even in the recent critical study on the mitral E/E′ ratio in decompensated HFREF patients with advanced HF and resynchronization therapy, a significant but weak correlation was observed between mean PCWP and mitral E velocity.13From these pathophysiological insights, it becomes evident that current Doppler echocardiographic techniques for estimation of LV filling pressures or diastolic LV dysfunction cannot be used as a "1 size fits all" tool but that a "tailored" approach is needed. Such a tailored approach with different strategies in HFREF and HFNEF is indeed proposed in the recent recommendations for the echocardiographic evaluation of LV diastolic function issued by the American Society of Echocardiography and the European Association of Echocardiography.11 Their recommendations for the echocardiographic diagnosis of diastolic LV dysfunction in HFNEF also correspond to earlier guidelines published in a consensus document for the diagnosis of diastolic HF written by the Heart Failure and Echocardiography associations of the European Society of Cardiology.4 Finally, different noninvasive diagnostic strategies for diastolic LV dysfunction in HFNEF and HFREF provide support for HFNEF and HFREF as distinct HF phenotypes, with LV remodeling in both conditions driven by dissimilar gene programs.4,35 This unique course of LV remodeling in HFNEF and HFREF is further endorsed by recent large multicenter trials or registries in which treatment with angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, or β-blockers yielded positive outcome in HFREF49–51 but neutral outcome in HFNEF.49,51,52Pitfalls of Diastolic Dysfunction IndexesDiastolic LV function can be assessed in each of the 4 phases of diastole: isovolumic relaxation, rapid filling, slow filling, and atrial contraction. Each of these 4 phases uniquely reflects cardiomyocyte, myocardial, or LV physiology and is variably accessible to invasive or noninvasive evaluation. Indexes of diastolic LV dysfunction derived from each of these 4 phases therefore have different clinical implications, have to meet specific clinical or scientific needs, and often have specific pitfalls frequently overlooked when noninvasive surrogate measures are proposed.Isovolumic relaxation time, which corresponds to the time interval from aortic valve closure to mitral valve opening, is difficult to appreciate from simultaneous LV pressure, aortic pressure, and PCWP recordings but is easily measured by continuous-wave Doppler from the simultaneous display of the end of aortic ejection and the onset of mitral inflow. Its clinical value as an index of diastolic LV function is limited because it depends on arterial or mitral valve opening pressures and therefore is not uniquely related to LV isovolumic relaxation rate. An invasively determined time constant of an exponential curve fit to LV pressure fall (τ) is a better approximation of LV isovolumic relaxation rate. Three potential pitfalls should be addressed when calculating τ. First, does a monoexponential curve fit adequately describe LV pressure decay? Second, which start and end points have to be used for the curve fitting procedure? Third, which value is assigned to the asymptote pressure of the fit (Pinf)? Although biexponential, polynomial, and logistic models have all been proposed, a single monoexponential curve fit usually describes LV pressure decay adequately and yields a satisfactory correlation coefficient (ie, r>0.99). Exceptions are patients with hypertrophic cardiomyopathy and aortic stenosis. Because of some deviation of LV pressure decay from an exponential decline, a higher starting point or a higher end point will erroneously prolong τ.53 This usually has no implications except when τ values are compared under widely varying LV loading conditions (eg, between control subjects and hypertensive patients). Pinf is the final pressure to which LV pressure would decay in the absence of LV filling. The use of an exponential curve fit that allows Pinf to vary from 0 is mathematically the more correct analysis of LV relaxation kinetics. However, it yields values of τ that can be vastly different from the values of τ from an exponential curve fit with Pinf=0 (eg, 40 versus 35 ms in control subjects, 62 versus 43 ms in patients with aortic stenosis, and 74 versus 47 ms in those with hypertrophic cardiomyopathy).28 Noninvasive attempts54 to quantify τ all failed to account for these potential errors intrinsic to the τ concept.Most echocardiographic efforts to noninvasively measure LV filling pressures looked at the rapid LV filling phase of diastole and used mitral flow velocity Doppler (E wave), pulmonary venous flow velocity Doppler (D wave), color M-mode flow propagation velocity (Vp), and tissue Doppler mitral annular velocity (E′ wave). The peak values of these measurements are recorded at a single time point of the rapid LV filling phase. The LV diastolic pressure derived from these measurements estimates the LV pressure value at the corresponding time point, which occurs in the interval between Y descent of the LA pressure wave and the second diastolic LV-LA pressure crossover (Figure 4). This diastolic LV pressure value, however, is variably related to LV end-diastolic pressure (EDP) or mean LA pressure. In the presence of small V waves, the pressure value at the trough of the Y descent, which almost coincides with the second diastolic LV-LA pressure crossover, is significantly lower than mean LA pressure or LVEDP, which are of comparable magnitude.55 In the presence of large V waves, the pressure value at the trough of the Y descent is again significantly lower than mean LA pressure or LVEDP, but mean LA pressure also is 30% higher than LVEDP. Hence, apart from conceptual problems relating lengthening or inflow to pressures, it seems difficult for an echocardiographic index derived from a single snapshot measurement during the rapid LV filling phase to provide a reliable estimate of mean LA pressure or PCWP, which averages LA pressure or PCWP over the entire cardiac cycle. Download figureDownload PowerPointFigure 4. Simultaneous recordings of mean PCWP, phasic PCWP, and diastolic LV pressure (LVP) in a patient with large V waves. Pressure values corresponding with LV minimum diastolic pressure (LVMDP), Y trough, LVEDP, and X trough differ from mean PCWP. Approximate positions of mitral flow velocity E wave and of tissue Doppler mitral annular E′ wave are also indicated at peak PCWP-LV pressure gradient and just before the second PCWP-LV pressure crossover, respectively. The phasic PCWP values corresponding to E and E′ also differ from mean PCWP and from LVEDP.During the slow LV filling phase, residual effects of LV relaxation and "dynamic" effects of fast LV inflow have dissipated. This phase is used to construct diastolic LV pressure-volume relations from a single cardiac cycle and allows LV stiffness, the slope of the diastolic LV pressure-volume relation, to be derived under so-called "static" conditions. A major drawback is the limited range of diastolic LV pressures of these single-cycle LV pressure-volume relations. An overlapping range of diastolic LV pressures is therefore frequently missing when diastolic LV pressure-volume relations are being compared before and after intervention or from different populations. Similar diastolic LV pressure levels, however, are essential to compare LV stiffness moduli. Three approaches have been used to overcome this problem: construction of the diastolic LV pressure-volume relation from multiple LVEDP points obtained during transient caval occlusion,46 inclusion of the fast LV filling phase through calculation of a "passive" early diastolic LV pressure that accounts for residual "active" LV relaxation,24 and extrapolation along a curve fit to the diastolic LV pressure-volume relation to a common diastolic LV pressure level. The first approach is the gold standard for LV stiffness measurements and can be obtained only at cardiac catheterization with conductance catheters and balloon caval occlusions. The second approach is open to critique because it presumes that residual LV relaxation after mitral valve opening has an exponential decay similar to that observed during isovolumic relaxation. Experimental and clinical studies using a mitral valve occluder or a balloon mitral valvuloplasty catheter, however, showed measured diastolic LV pressure during obstructed mitral inflow to deviate significantly from diastolic LV pressure predicted from isovolumic LV pressure decay.56,57 The third approach was already being heavily criticized >30 years ago because it requires long extrapolations to achieve a common diastolic LV pressure level.58 Despite these critiques, a variant of this approach has recently been introduced to noninvasively compare diastolic LV stiffness in patient populations with arterial hypertension and HFNEF.59 This variant uses not only a single-beat but also a single-measurement approach because it noninvasively assesses a single value of LVEDP and LV end-diastolic volume. It subsequently derives the entire LVEDP-end-diastolic volume relation from this single value of LVEDP and LV end-diastolic volume under the assumption that volume-normalized LVEDP-LV end-diastolic volume relations always have a similar shape regardless of animal species or heart size.60 These assumptions, however, have been tested only ex vivo in explanted HFREF hearts and in vivo in control subjects and HFNEF patients. The current use of this approach to clinically assess LV stiffness therefore seems premature, and its noninvasive application has to await further invasive in vivo validation against conductance catheter-derived LVEDP-LV end-diastolic volume relations.Limitations of Current Doppler Echocardiographic TechniquesMitral Inflow PatternsMitral inflow patterns have a U-shaped relation with diastolic LV dysfunction. Thus, patterns in normal subjects and HF patients at an intermediate stage of decompensation are similar. This made the use of mitral inflow patterns cumbersome and inspired many investigators to find other methods to measure LV filling pressures noninvasively. Especially in HFNEF patients,47,48 the use of mitral inflow as an estimate of diastolic LV dysfunction is limited. In these patients, who frequently have LV hypertrophy related to arterial hypertension, diabetes, and obesity, residual LV relaxation pressures are prominent, and myocardial lengthening load increased only slightly despite elevated mitral valve opening pressure because of a favorable Laplace relationship with small end-systolic LV cavity size and thick LV walls. In HFNEF patients, elevated mean PCWP or LA pressure will therefore not change the predominant control of early diastolic LV filling by residual LV relaxation pressure, so a slow LV relaxation pattern can coexist with elevated mean PCWP or LA pressure. Because of this limited reliability of mitral inflow patterns, mitral flow velocity Doppler is no longer withheld as an adequate diagnostic method for diastolic LV dysfunction in HFNEF, as evident from 2 recent consensus statements,4,11 in contrast to earlier guidelines in which it was still proposed to be of value.1,3 In patients with HFREF, myocardial lengthening load is the predominant control mechanism of early diastolic filling and easily overrides the influence of residual relaxation pressure. In these patients, mitral inflow pattern will track mean PCWP or LA pressure; therefore, a high mitral E/A ratio was proposed as first-line diagnostic evidence for diastolic dysfunction in HFREF patients.11 A major drawback for the use of mitral inflow patterns in HFREF patients is the elevation of the E wave by mitral regurgitation induced by mitral annular dilatation and eccentric LV remodeling. This E-wave elevation can erroneously mimic a restrictive LV filling pattern; thus, some investigators have suggested that in the presence of mitral regurgitation, evidence of diastolic LV dysfunction should be inferred only from end-diastolic indexes such as the difference in duration between the pulmonary venous atrial reversal velocity (Ar) and the mitral A wave (Ar-A).61 Furthermore, a large V wave of mitral regurgitation causes earlier opening of the mitral valve and earlier peaking of the diastolic LV-LA pressure gradient and of the peak mitral E-wave velocity. Peak mitral E-wave velocity thereby coincides with a higher LA pressure on the downslope of the V wave and becomes a poorer estimate of LVEDP.Pulmonary Venous Flow PatternsPulmonary venous diastolic (D) velocity changes in parallel with mitral E velocity and therefore has similar shortcomings as a tool to estimate LV filling pressures or to grade diastolic LV dysfunction.62 Ar-A duration is more useful because it relates to the A-wave-induced LV pressure increase and end-diastolic LV stiffness.63 Moreover, it is the only noninvasive estimate of a diastolic LV compliance reduction, which has not yet sufficiently evolved to raise mean LA pressure. Its widespread application is hindered, however, by difficult procurement of high-quality pulmonary venous flow velocity recordings suitable for analysis.Pulmonary Artery Systolic PressureIn HF patients, high mean LA pressure or PCWP can be inferred from elevated pulmonary artery pressures. Pulmonary artery systolic pressure is derived from the tricuspid regurgitant jet by continuous-wave Doppler and the right atrial pressure.64 A correct estimation of right atrial pressure and the high variability of the relation between pulmonary artery pressure and PCWP are obvious drawbacks of this method. A recent report on patients with HFNEF showed their pulmonary hypertension to result both from a component reactive to elevated mean LA pressure and from a component of precapillary pulmonary hypertension65 possibly induced by insensitivity to nitric oxide, endothelin, or prostaglandin vasodilator signaling pathways.66 These observations are an important warning against indiscriminate use of pulmonary artery systolic pressures as estimates of PCWP in HFNEF patients.Tissue Doppler Annular VelocitiesDiastolic tissue velocities measured at the mitral annulus show low-velocity deflections during early filling (E′) and with atrial contraction (A′). E′ is presumed to correlate closely with LV relaxation indexes such as τ and to be relatively preload insensitive.67 These initial presumptions about E′, however, have recently been refuted by E′ measurements in instrumented dogs, which revealed close correlations of E′ with mitral valve opening pressure but weak correlations with LV relaxation rate under a variety of experimental conditions.39 Moreover, similar to mitral E flow, E′ appears to be age dependent.68–70 This age dependence also could detract from its prognostic value.71–73 The E/E′ ratio has been proposed as a reliable estimate of LV filling pressure.15–19 Because E depends on LA pressure, residual LV relaxation pressure, and age and because E′ is presumed to depend only on LV relaxation pressure, dividing E by E′ eliminates LV relaxation pressure and age, so the E/E′ ratio becomes a noninvasive estimate of LA pressure. Similar to E′, the value of the E/E′ ratio as a reliable estimate of LA pressure has recently been questioned both in patients with hypertrophic cardiomyopathy and in decompensated patients with resynchronization therapy.12,13Apart from conceptual problems involving E′, some practical issues detract from its usefulness. Septal and lateral mitral annular E′ velocities differ. Recent guidelines for the diagnosis of diastolic HF therefore recommend use of an E/E′ value that is the average of septal and lateral mitral annular E′.4 Furthermore, a value of E/E′ >15 is usually proposed as evidence for elevated LV filling pressure and a value of E/E′ <8 as evidence for normal LV filling pressure. As a consequence, there is a wide range of E/E′ values (8 15) for which additional investigations are required to obtain a LV filling pressure estimate. Further technical limitations include angle dependency, signal noise, signal drifting, spatial resolution, sample volume, and tethering artifacts. E′ also can be reduced erroneously by mitral annular calcification, surgical rings, or prosthetic valves.Deformation AnalysisMyocardial deformation (strain) is an important consequence of LV contraction. Myocardial strain and strain rate used to be measured with magnetic resonance imaging but recently have also been determined by speckle-tracking echocardiography, in which patterns of echocardiographic pixel intensity are identified and tracked throughout the cardiac cycle. Assessment of myocardial diastolic strain and diastolic strain rate avoids tissue Doppler-associated angulation errors and tethering artifacts. It has recently been used experimentally to evalua