Cardiac Imaging to Evaluate Left Ventricular Diastolic ... · pressure and therefore a reduction in the early dia-stolic transmitral pressure gradient and the early transmitralvelocity(E)andacompensatoryincreasein

J A C C : C A R D I O V A S C U L A R I M A G I N G V O L . 8 , N O . 9 , 2 0 1 5

ª 2 0 1 5 B Y T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y F O U N DA T I O N I S S N 1 9 3 6 - 8 7 8 X / $ 3 6 . 0 0

P U B L I S H E D B Y E L S E V I E R I N C . h t t p : / / d x . d o i . o r g / 1 0 . 1 0 1 6 / j . j c m g . 2 0 1 5 . 0 7 . 0 0 4

STATE-OF-THE-ART PAPERS

Cardiac Imaging to EvaluateLeft Ventricular Diastolic Function
Frank A. Flachskampf, MD, PHD,* Tor Biering-Sørensen, MD, PHD,y Scott D. Solomon, MD,yOlov Duvernoy, MD, PHD,z Tomas Bjerner, MD, PHD,z Otto A. Smiseth, MD, PHDx
JACC: CARDIOVASCULAR IMAGING CME

CME Editor: Ragavendra R. Baliga, MD

This article has been selected as this issue’s CME activity, available online

at http://www.acc.org/jacc-journals-cme by selecting the CME tab on the

top navigation bar.

Accreditation and Designation Statement

The American College of Cardiology Foundation (ACCF) is accredited by

the Accreditation Council for Continuing Medical Education (ACCME) to

provide continuing medical education for physicians.

The ACCF designates this Journal-based CME activity for a maximum of 1

AMA PRA Category 1 Credit(s) �. Physicians should only claim credit

commensurate with the extent of their participation in the activity.

Method of Participation and Receipt of CME Certificate

To obtain credit for this CME activity, you must:

1. Be an ACC member or JACC: Cardiovascular Imaging subscriber.

2. Carefully read the CME-designated article available online and in this

issue of the journal.

3. Answer the post-test questions. At least 2 out of the 3 questions

provided must be answered correctly to obtain CME credit.

4. Complete a brief evaluation.

5. Claim your CME credit and receive your certificate electronically by

following the instructions given at the conclusion of the activity.

From the *Institutionen för Medicinska Vetenskaper, Uppsala Universitet, U

and Women’s Hospital, Harvard Medical School, Boston, Massachusetts; zInRadiologi, Uppsala Universitet, Uppsala, Sweden; and the xDepartment

Rikshospitalet, Oslo University Hospital, Center for Cardiological Innovatio

Heart Failure Research, University of Oslo, Oslo, Norway. Dr. Bjerner has se

other authors have reported that they have no relationships relevant to the

Manuscript received April 20, 2015; revised manuscript received July 2, 2015

CME Objective for This Article: At the end of this activity the reader should

be able to: 1) review the physiology and pathophysiology of left ven-

tricular diastolic function, including the relation to imaging parameters;

2) understand how different cardiac imaging techniques assess left

ventricular diastolic function, and which limitations exist for these

approaches; and 3) summarize the published experience with cardiac

imaging from clinical trials in which left ventricular diastolic function was

important for selection of patients or as an outcome variable.

CME Editor Disclosure: JACC: Cardiovascular Imaging CME Editor

Ragavendra R. Baliga, MD, has reported that he has no relationships to

disclose.

Author Disclosures: Dr. Bjerner has served on advisory boards for

Carestream Health. All other authors have reported that they have no

relationships relevant to the content of this paper to disclose.

Medium of Participation: Print (article only); online (article and quiz).

CME Term of Approval

Issue Date: September 2015

Expiration Date: August 31, 2016

ppsala, Sweden; yCardiovascular Division, Brighamstitutionen för Kirurgiska Vetenskaper, Enheten för

of Cardiology and Institute for Surgical Research,

n, K.G. Jebsen Cardiac Research Centre, Centre for

rved on advisory boards for Carestream Health. All

content of this paper to disclose.

, accepted July 15, 2015.

http://www.acc.org/jacc-journals-cme

http://crossmark.crossref.org/dialog/?doi=10.1016/j.jcmg.2015.07.004&domain=pdf

http://dx.doi.org/10.1016/j.jcmg.2015.07.004

Flachskampf et al. J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 8 , N O . 9 , 2 0 1 5

Cardiac Imaging for LV Diastolic Function S E P T E M B E R 2 0 1 5 : 1 0 7 1 – 9 3

1072

Cardiac Imaging to Evaluat
eLeft Ventricular Diastolic Function
ABSTRACT

Left ventricular diastolic dysfunction in clinical practice is generally diagnosed by imaging. Recognition of heart

failure with preserved ejection fraction has increased interest in the detection and evaluation of this condition

and prompted an improved understanding of the strengths and weaknesses of different imaging modalities for

evaluating diastolic dysfunction. This review briefly provides the pathophysiological background for current

clinical and experimental imaging parameters of diastolic dysfunction, discusses the merits of echocardiography

relative to other imaging modalities in diagnosing and grading diastolic dysfunction, summarizes lessons from

clinical trials that used parameters of diastolic function as an inclusion criterion or endpoint, and indicates

current areas of research. (J Am Coll Cardiol Img 2015;8:1071–93) © 2015 by the American College of Cardiology

Foundation.

T he concept of left ventricular (LV) diastolicfunction as a characteristic separated fromsystolic function is not immediately com-

pelling. Classic functional diagrams of LV function,like the pressure-volume loop, show a smoothtransition between systolic and diastolic pressure-volume data. Furthermore, the term function, al-though appealing for the pump action of the LVin systole, seems less appropriate for the partiallypassive ventricular filling phase changes in volumeand pressure. Nevertheless, the term diastolic func-tion has become firmly rooted in cardiology anddenotes the ability of the LV to fill sufficientlyto produce the requested stroke volume withoutexceeding certain pressure limits during filling.LV pressure during diastole is nearly identical toleft atrial (LA) and pulmonary capillary pressurebecause the latter structures have an open commu-nication with the LV during diastole. IncreasedLA pressure implies pulmonary congestion, whichaccounts for dyspnea in patients with left heartfailure. Clinical interest has been boosted by therealization that about one-half of patients present-ing with heart failure symptoms have a “pre-served” LV ejection fraction (>50%), although atcloser inspection, systolic functional abnormalitiessuch as reduced longitudinal LV shortening areoften present. Hence, it has been assumed that alarge proportion of the “heart failure epidemic” isprimarily caused by diastolic LV dysfunction, andthis has been termed heart failure with preservedejection fraction (HFpEF). However, because symp-toms of heart failure, in particular dyspnea, are notalways cardiac in origin, direct evidence of suchdiastolic dysfunction should be sought before

settling on the diagnosis of diastolic dysfunctionto explain HFpEF. Alternative diagnoses includenoncardiac disorders such as pulmonary diseaseand obesity. Furthermore, valvular disease, suchas mitral stenosis, constrictive pericarditis, congen-ital heart disease, pulmonary arterial hypertension,and others, may cause HFpEF and should beruled out. Because exertional dyspnea is a commonsymptom of coronary artery disease, this diseasealso needs to be considered. Direct evidence ofdiastolic LV dysfunction can be provided by inva-sive measurements (LV pressure tracings or ideallypressure-volume data), natriuretic peptide levelsindicating myocardial stretch, or cardiac imaging—first and foremost, echocardiography (1). Imagingtechniques, however, lack the capability of directlymeasuring pressures, although they can measurevolumes and blood and tissue velocities; there-fore, by nature, they only provide indirectevidence of diastolic pressures or pressure-volumerelationships.

Like in any other diagnostic work-up, a clinicalquestion should always be the starting point. Reasonsfor referring patients for evaluation of diastolicfunction include: 1) symptoms or signs of heart failurein patients with preserved LV ejection fraction; 2) theneed for an estimate of LV filling pressure in patientswith known heart disease; and 3) assessment of car-diovascular risk.

In the following, we will review current knowledgeon how to assess diastolic LV function by cardiacimaging, how well validated such assessment is, andwhich prognostic implications these data have. Wewill also describe the most important current openquestions and unmet needs.

AB BR E V I A T I O N S

AND ACRONYM S

AV = atrioventricular

CMR = cardiac magnetic

resonance

CT = computed tomography

e0 = left ventricular regional

lengthening velocity, usually

the septal or lateral basal LV

tissue velocity, also called

mitral annular velocity

e0sr = early diastolic strain rate

HFpEF = heart failure with

preserved ejection fraction

IVRsr = strain rate during

isovolumic relaxation time

LA = left atrium/atrial

LV = left ventricle/ventricular

LVEDP = left ventricular end-

diastolic pressure

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 8 , N O . 9 , 2 0 1 5 Flachskampf et al.S E P T E M B E R 2 0 1 5 : 1 0 7 1 – 9 3 Cardiac Imaging for LV Diastolic Function

1073

DIASTOLIC LV FUNCTION: A BRIEF SUMMARY

OF PHYSIOLOGY AND ITS REFLECTION IN

ECHOCARDIOGRAPHIC PARAMETERS

Whereas LV systolic function is routinely quantified bymeasuring ejection fraction or deformation parame-ters such as global longitudinal strain, there is no singleclinical measure that quantifies LV diastolic function;instead, a large number of indexes based on cardiacimaging have been introduced.

The physiological hallmarks of LV diastolicdysfunction are impaired relaxation, loss of restoringforces, reduced diastolic compliance, and elevated LVfilling pressure. The latter is not a primary disturbanceof ventricular function but is a compensatory mecha-nism to maintain LV filling and stroke volume (2).In some patients, elevated filling pressure is observedonly during exercise; therefore, normal filling pressureat rest does not exclude clinically significant diastolicdysfunction or HFpEF (3,4). There is no single non-invasive index that provides a direct measure ofrelaxation, restoring forces, compliance, or LV fillingpressure. However, by using a combination of differentnoninvasive indexes, it is feasible in most patients todetermine if diastolic function is normal or impaired(Central Illustration).

In addition, structural changes (e.g., LV hypertro-phy) can indicate abnormalities of diastole and shouldbe considered.

INDEXES OF RELAXATION AND RESTORING FORCES.

Slowing of LV relaxation and loss of restoring forcesleads to reductions in velocities of LV lengthening anduntwisting and reduces the decay rate of LV pressureduring isovolumic relaxation. The latter response canbe measured invasively by micromanometer as pro-longation of the time constant tau of LV pressure fall.LV regional lengthening velocity (e0), however, can bemeasured noninvasively by tissue Doppler imagingand is a key parameter of diastolic function because itreflects relaxation as well as restoring forces (5,6)(Figure 1). It is inversely related to the relaxationconstant tau. Similar to most other parameters of LVfunction, e0 is modified by diastolic pressure, but inthe failing heart, e0 is less load dependent thantransmitral velocities (7). Furthermore, if LV relaxa-tion is slowed, e0 onset and peak, which normallyoccur close to onset and peak of the transmitral Ewave, are delayed, leading to less dependence of e0

on LA pressure (8). Table 1 shows normal referencevalues for e0.

LV early diastolic strain rate provides similar diag-nostic information as e0 and can technically be derivedfrom speckle tracking–based strain imaging, although

the comparatively low achievable frame ratesmake substantial information loss likely.Although it is independent of cardiac trans-lational motion, less angle dependent, andless affected by tethering from neighboringsegments than velocities by tissue Dopplerimaging, it presently must be seen as anexperimental parameter of relaxation (9).Similar to e0, LV untwisting velocity is deter-mined by LV relaxation and restoring forcesand has the potential to become an importantmeasure of diastolic function (10,11). Due tomethodological challenges, however, assess-ment of untwisting velocity is not recom-mended for routine clinical examinations atthe present time (9,10).

Because restoring forces are created bysystolic contraction, both e0 and untwistingvelocity are determined in part by systolicfunction (12,13), reflecting the tight coupling
between systolic and diastolic function.
There is no clinical method that differentiates be-tween myocardial relaxation and restoring forces. Inheart failure, both mechanisms are affected andcontribute to slowing of LV isovolumic pressure falland elevation of minimal LV diastolic pressure.Restoring forces may play a unique role by creatingnegative LV diastolic pressure, which represents thedriving force for diastolic suction (13,14). Restoringforces are generated in systole by a complex set ofmechanisms (15). A key element is storage of energydue to contraction of myocytes and ventricles belowresting length (unstressed dimension), and analogousto a compressed elastic spring, themyocardium recoilsin early diastole when the ventricle relaxes. Due toventricular filling, the negative pressures that resultfrom restoring forces may not be observed clinicallywhen LV pressure is recorded, and there is no easynoninvasive method to measure diastolic suction.Rapid early diastolic mitral-to-apical flow propagationhas been proposed as amarker of diastolic suction (16).

In ventricles with increased end-systolic volumes,restoring forces are attenuated or lost, as indicated byelevation of minimal LV diastolic pressure, and my-ocardial relaxation becomes an increasingly impor-tant determinant of e0.

Mitralflowvelocities andfilling patternsmay also beused to evaluate LV relaxation. Reduced decay rate ofLV pressure due to slowing of relaxation and loss ofrestoring forces leads to elevation of minimal LVpressure and therefore a reduction in the early dia-stolic transmitral pressure gradient and the earlytransmitral velocity (E) and a compensatory increase inthe A wave. The result is a reduction in the E/A ratio

CENTRAL ILLUSTRATION Schematic Drawing of Diastolic LV Filling

Diastolic left ventricular (LV) filling is shown with simultaneous recordings of left atrial (LA) and LV pressures, the diastolic transmitral pressure gradient, pulmonary

venous, transmitral, and transtricuspid flow velocity, and septal tissue (or mitral annular) velocity. Flow velocity and tissue velocity data, together with LA size and

systolic pulmonary pressure, are used in echocardiography to infer diastolic LV pressures and hence diastolic LV function.



1074

along with prolongation of the deceleration time ofE, and thisfilling pattern is named impaired relaxation.With further progression of diastolic function, how-ever, this feature may disappear due to compensatoryrise in LA pressure, which serves tomaintain LV filling,

and result in a pattern named pseudonormal fillingwith normal transmitral E/A ratio. In contrast to “true”normal filling, there is typically an associated reduc-tion in e0 and often a more marked pulmonary venousreversed velocity during atrial contraction. Such

TABLE 1 Normal Values for e0 and E/e0

e0 , cm/s E/e0 Ratio

Women, <40 yrs 10.0–19.2 3.0–8.2

Women, 40–59 yrs 6.5–16.1 3.2–10.4

Women, $60 yrs 1.8–14.6 3.1–14.3

Women, overall 5.4–18.2 2.5–10.9

Men, <40 yrs 8.7–19.5 2.5–8.5

Men, 40–59 yrs 6.1–15.3 3.0–9.4

Men, $60 yrs 4.4–12.0 3.1–12.3

Men, overall 4.8–16.8 2.4–10.4

Normal reference ranges for left ventricular regional lengthening velocity (e0 ; bypulse-wave Doppler) and E/e0 ratio (mean � 2 SD) in 1,266 randomly selectedhealthy individuals. Average values from all 4 left ventricular walls. The findingswere similar when e0 was calculated as an average of septal and lateral e0 . Modifiedfrom Dalen et al. (114).

FIGURE 1 The 3 Independent Determinants of LV Early

Diastolic Lengthening Velocity

(Top) Active relaxation, a decay in active fiber force by reduction

of actin-myosin interaction as cytosolic calcium is taken up by the

sarcoplasmic reticulum. (Middle) Restoring forces represent

passive elastic recoil, which occurs during left ventricular (LV)

relaxation and causes the ventricle to return to its resting posi-

tion. Restoring forces are the result of systolic contraction to

myocardial dimensions less than resting length (L0). The

magnitude of restoring forces varies inversely with minimal

length (Lmin). In the normal LV, restoring forces result in negative

early diastolic pressure, which sucks blood into the ventricle.

During heart failure, when the LV does not contract below L0,

restoring forces are lost and there is a loss of diastolic suction

and secondarily reduction in early diastolic lengthening velocity

(e0). (Bottom) Left atrial (LA) pressure at the onset of filling

determines the force with which the blood fills the LV and causes

it to lengthen. AO ¼ aorta. Modified, with permission from

Smiseth et al. (115).


1075

pseudonormalization due to increasing LA pressuresgenerated by advanced disease also occurs with otherblood flow–based parameters like isovolumic relaxa-tion time, E-wave deceleration time, and pulmonaryvenous diastolic flow, and constitutes a fundamentalproblem in the assessment of diastolic function. Insummary, impairment of early diastolic function dueto slowing of relaxation and loss of restoring forces canbe detected noninvasively in clinical routine practiceas reduction in e0 and as a transmitral filling pattern ofimpaired relaxation (Figure 2). Due to age dependencyof both parameters, it is important to use age-adjustedreference values (Table 1).

INDEXES OF REDUCED DIASTOLIC COMPLIANCE. LV dia-stolic compliance is the relationship between changein LV volume and change in pressure in diastole, or thelocal slope of the pressure-volume relationship. It isdetermined by LV myocardial compliance and byelastic properties of extraventricular structures (i.e.,pericardium and lungs) and by right ventricular pres-sure. Because the LV pressure-volume relationshipdoes not reflect myocardial compliance, the termchamber compliance is often used. Importantly, be-cause the LV pressure-volume relationship is cur-vilinear, chamber compliance is a function of theoperative LV diastolic pressure and consequently ismarkedly load dependent. Therefore, a change in LVchamber compliance does not necessarily mean therehas been a change in myocardial elastic properties, itmay just be a change in loading conditions that movesthe pressure-volume coordinate to a part of the curvethat has a different slope. There is no noninvasivemethod that measures diastolic compliance, butthere are indexes that are related to LV compliance.There is an association between reduced LV diastoliccompliance and short E-wave deceleration time (17–19). Values <150 ms are consistent with reduction inLV diastolic compliance in particular when there isalso a mitral filling pattern of restrictive physiology(tall E and high E/A ratio). The relationship betweenE-wave deceleration time and compliance is consistentwith theoretical predictions (18–20) (Figure 3). Anadditional feature of this filling pattern is abbreviatedisovolumic relaxation time due to premature openingof themitral valve caused by elevated LA pressure (21).Importantly, the relationship between E-wave decel-eration time and degree of diastolic dysfunction isnonlinear; therefore, in mild diastolic dysfunctiondominated by impaired relaxation, there is prolonga-tion of E-wave deceleration time because of ongoingrelaxation during flow deceleration. Furthermore, in

FIGURE 2 Extreme Example of “Impaired Relaxation”

Pattern in a Patient With Hypertrophic Cardiomyopathy

(Top) Transmitral flow pattern with near-absent E wave despite

sinus rhythm and normal heart rate. (Bottom) Tissue Doppler

recording from the septum with very low early diastolic

lengthening velocity.



1076

young healthy individuals with a highmitral E, there isoften a deceleration time of <150 ms (22). Therefore, ashort deceleration time should not be used as a stand-alone index of reduced LV compliance. Reduction inLV chamber compliance is also reflected in attenuationand abbreviation of the transmitral A velocity (23) andis typically combined with accentuation and prolon-gation of the pulmonary vein reversed A velocity(24,25). When LV compliance is reduced, a smallchange in volume results in a large change in pressure;therefore, atrial contraction results in a more markedincrease in LV pressure. When the atrium contractsagainst a ventricle with reduced compliance, littleblood moves forward across the mitral valve, ante-grade mitral flow is interrupted prematurely, and

instead blood regurgitates into the more compliantpulmonary veins. In the early phase of diastolic dys-function and with a well-functioning LA, there maybe a relatively normal LV pre–A-wave pressure and amarked rise in LV pressure during vigorous atrialcontraction, particularly when there is reduced LVcompliance. Low amplitude and short duration of thetransmitral A velocity in combination with a highamplitude and long duration of the reverse pulmonaryvenous velocity are typical for a ventricle with re-duced chamber compliance and elevated end-diastolicpressure (23–26).

Limitations of the difference in A-wave duration asan index of diastolic compliance include atrial me-chanical failure and problems with the quality of thepulmonary venous flow signal. Furthermore, reverseA may be shortened in tachycardia and when the PRinterval is prolonged due to atrioventricular (AV)block I or if there is significant antegrade pulmonaryvenous flow at the time of atrial contraction, reflect-ing the effect of blood inertia. In these cases, therelation of the difference in A duration betweenmitral and pulmonary venous flow has less predictivevalue for an increase in diastolic pressures.

Thus, a short E-wave deceleration time togetherwith a small and abbreviated mitral A and a largereverse A of long duration are consistent with reducedLV compliance.

ECHOCARDIOGRAPHIC DETECTION OF

ELEVATED LV FILLING PRESSURES IN

CLINICAL PRACTICE

LV filling pressure is invasively measured as left ven-tricular end-diastolic pressure (LVEDP) or as pulmo-nary capillary wedge pressure as an indirect measureof mean LA pressure. There is often a slight differencebetween the 2, but this is rarely of clinical significanceas long as there is no mitral stenosis. Filling pressure isconsidered elevated when the mean pulmonarycapillary wedge pressure is >12 mm Hg or when theLVEDP is >16 mm Hg (1).

Like all imaging modalities, echocardiography hasno direct way to measure pressures in the LV. Thus,indirect signs of pathological pressure-volume re-lationships in the LV during diastole must be reliedupon. Becausemitral flow velocities are determined bythe transmitral pressure gradient, peak mitral E is avaluable parameter in the estimation of LA pressure.However, due to a relatively large volume of bloodcontained within a normal mitral valve, a substantialfraction of the transmitral pressure gradient is neededto overcome inertia. Therefore, mitral velocity can-not be converted to pressure difference using the

FIGURE 3 Flow From Chamber A to B Measured by Doppler Echocardiography

In a hydraulic model, the flow from chamber A to B (arrow) was measured by Doppler echocardiography, mimicking transmitral flow during

passive left ventricular filling. A reduction in compliance in the receiving chamber (B) by placement of a solid object that reduced chamber

compliance led to shorter Doppler deceleration time. Reproduced with permission from Flachskampf et al. (18).


1077

simplified Bernoulli equation, like in mitral stenosis,and the real pressure gradient is larger than that pre-dicted by the simplified Bernoulli equation. Never-theless, peak transmitral filling rate correlates wellwith the transmitral pressure gradient (27,28). Animportant limitation of peak mitral E as an index of LApressure is that it reflects only the AV pressure differ-ence. Because LA pressure is the sum of LV diastolicpressure and the transmitral pressure difference,elevated mitral E velocity alone is not sufficient in-formation to conclude that LA pressure is elevated.Therefore, in normal hearts that have low or evenslightly negative early diastolic LV pressures, a tallmitral E obviously does not imply a high LA pressure.However, in ventricles with reduced ejection fractionor when e0 is low, impaired relaxation, and thereforeelevated LV early diastolic pressure, can be expected.This is why the combination of a tall E and a low e0 in-dicates elevated LV diastolic pressure and is expressedas a high E/e0 ratio. Because the normal range of e0

is wide, with values between approximately 5 and18 cm/s (Table 1), it is obvious that the E/e0 ratio cannotbe an accurate index of LA pressure. When used incombination with other indexes, however, it hasproven useful. Furthermore, an E/e0 ratio larger than 15is generally abnormal, regardless of age, and is strongevidence of diastolic dysfunction. Restrictive mitralfilling pattern with a tall peak E, short E-wave decel-eration time, and small A wave is consistent with

elevated LA pressure if e0 is reduced (Figure 4). Asindicated in Table 1, the lower normal limit for e0 is 6 to6.5 cm/s for individuals under the age of 60 years. Inolder individuals who are apparently healthy, evenlower values of e0 may be found, but this may reflectimpaired diastolic reserve with aging or undetectedcoronary artery disease. Furthermore, if e0 exceeds 8cm/s, reduced LV relaxation can be excluded andfurther studies of diastolic function may not beneeded. One important exception are patients withconstrictive pericarditis, who often have a normal e0. Inaddition, patients with significant mitral regurgitationtend to have elevated e0, and this should be kept inmind when the flow chart proposed in Figure 5 is usedto conclude that a patient has normal diastolic func-tion. Because e0 is used as a parameter of global LVdiastolic physiology, but is measured locally, localchanges may influence e0 in a way not reflecting globalLV diastolic function. In particular, septal infarction,ventricular pacing, mitral annular calcification, andleft bundle branch block may lead to low septal e0. Inthese cases, e0 should be measured at the basal lateralwall. In general, using an average value from septaland lateral e0 measurements (which typically areapproximately 10% higher than septal e0 measure-ments) has been recommended as a surrogate formeasuring e0 in all 6 walls; however, in some cases(e.g., very dilated ventricles) it is nearly impossible tocorrectly measure lateral longitudinal velocity, and

FIGURE 4 Typical Transmitral Restrictive Pattern

(Top) Patient with hypertension and severely depressed systolic and diastolic LV function. Note steep deceleration slope (dashed line), very

short E-wave deceleration time (102 ms), minimal A wave (E/A ratio ¼ 100/20), and very low septal tissue Doppler velocities both in systole

and diastole (e0, arrow), with E/e0 ratio ¼ 100/4 ¼ 25 (Bottom). Abbreviations as in Figure 1.



1078

the sample volume records a mix of longitudinal andradial tissue velocities. In these cases, lateral e0 shouldnot be used.

Systolic pulmonary artery pressure calculated as thesum of the estimated right atrial pressure and thesystolic tricuspid pressure gradient can be obtained inmost patients and is one of themost valuablemeasuresin the evaluation of LV diastolic function. Providedthere is no pulmonary arterial disease, an elevatedpulmonary artery pressure is strongly suggestive ofelevated LA pressure. Mitral regurgitation is anotherpossible cause of elevated pulmonary artery pressureand needs to be considered.

As explained in the section on indexes of diastolicstiffness, the magnitudes and durations of the ante-grade mitral A and the reversed pulmonary venous Acan be used to roughly estimate LVEDP. A reversepulmonary venous A wave >30 ms longer in durationthan the antegrade A wave is consistent with elevatedLVEDP (10).

Another measure from the pulmonary venous flowtracing is the ratio between peak systolic velocity andpeak diastolic velocity. A ratio of systolic to diastolicpeak velocity of <1 is consistent with, but not specificfor, elevated LV filling pressure (10); other conditionsassociated with such a ratio are atrial fibrillation,

FIGURE 5 Simplified Flow Chart for the Integrative Echocardiographic Evaluation of LV Diastolic Function

IFLAVI < 34 mL/m2 ANDe’ sep > 8 cm/s ANDe’ lat > 10 cm/sNo Constrictive Pericarditis

Grade IE/A < 0.8dec. time > 200 msAvg E/e’ < 9(Or Hypovolemiawithout DiastolicDysfunction)

Grade IIE/A 0.8 – 1.5dec. time 160 – 200 msAvg E/e’ 9 – 12

Grade IIIE/A > 2dec. time < 160 msAvg E/e’ > 12

IFLAVI ≥ 34 mL/m2 ANDe’ sep ≤ 8 cm/s ORe’ lat ≤ 10 cm/s

No Diastolic Dysfunction

Diastolic Dysfunction

If inconclusive, consider:SPAP , pulmonary venous S/D < 1, onset e’ later than onset E, LV wall hypertrophy, and other signs

Modified from current recommendations (10). Avg ¼ average (from basal septal and basal lateral LV wall); dec ¼ deceleration; lat ¼ lateral;

LAVI ¼ left atrial volume index; sep ¼ septal; SPAP ¼ systolic pulmonary artery pressure; other abbreviations as in Figure 1.


1079

mitral regurgitation, and hypernormal early diastolicrelaxation in young patients. Further signs of ele-vated LA pressure include the occurrence of a trans-mitral L wave (Figure 6), which is due to continuingantegrade pulmonary venous flow into late diastolein the presence of delayed LV relaxation and alater onset of e0 than E wave, indicating that mitralflow volume is “pushed” rather than “pulled” intothe LV.

LA volume can be measured by echocardiographyand reflects in part the cumulative effects of fillingpressures over time. Therefore, an enlarged LA is amarker of elevated LA pressure and a small leftatrium suggests normal LA pressure. Importantly,atrial volume is not an index of instantaneous pres-sure because enlargement of the atrium takes time.Furthermore, an enlarged atrium may persist longafter normalization of LA pressures. An upper normalvalue of 34 ml/m2 is commonly used; a dilated LAhigher than this value may also be seen in younghealthy athletes and in patients with bradycardia,anemia, and other high-output states; atrial arrhyth-mias; and mitral valve disease (10). On the otherhand, mild diastolic dysfunction without substantial

pressure elevation at rest may coexist with LA size inthe upper range of normal. Thus, LA volume shouldnot be used as a stand-alone parameter to rule in orrule out diastolic dysfunction.

Current recommendations (10) can be summarizedas follows. Echocardiographic evaluation of LV fillingpressure should include estimation of pulmonary ar-tery systolic pressure, peak mitral E, and E/e0 ratio.Duration difference between antegrade and retrogradeA wave, LA volume, and E-wave deceleration time areadditional useful measures. The only parameter in thislist that provides a direct estimate of pressure isthe tricuspid regurgitation velocity. An advantage ofpulmonary artery systolic pressure and difference inA-wave duration is that they are essentially age inde-pendent, and this also applies to the E/e0 ratio, whichis <15 in healthy individuals. Except for estimatedpulmonary artery pressure from tricuspid regurgita-tion velocity, all other indexes are predictive of LVfilling pressure due to associations and therefore needto be interpreted with several limitations in mind. Inspite of this, when indexes are consistent, it is possibleinmost patients to concludewhether the level of fillingpressure is normal or elevated (Figure 5). In some

FIGURE 6 Transmitral L Wave

This is a sign of elevated LA and LV filling pressure caused by continuing mid to late pulmonary venous flow. Abbreviations as in Figure 1.



1080

patients, however, there is inconsistency between in-dexes; therefore, the noninvasive study becomesinconclusive with regard to filling pressure. The diag-nosis of diastolic dysfunction by echocardiographyhas limited reliability in particular in the followingcircumstances: rapidly changing hemodynamics(29); presence of other local factors influencing e0

(infarction, bundle branch block, pacing); additionalsubstantial valvular heart disease (e.g., mitral regur-gitation); mitral annular calcification, which mayreduce e0 and augment E; and initial stages of diastolicdysfunction.

NEW ECHOCARDIOGRAPHIC TECHNIQUES

DIASTOLIC STRESS TEST. In some patients, LV diastolicpressures are normal at rest but become abnormalunder exercise; this is opposed to the response of ahealthy LV to physical exercise, during which diastolicpressures aremaintainedwithin normal ranges, in partdue to enhanced myocardial relaxation (shorter tau)(30). Moreover, the invasive finding of elevated pul-monary wedge pressure at rest or at exercise in pa-tients with suspected HFpEF seems to identify sickerpatients with highermortality, as a retrospective studysuggested in 355 patients (31), with a 10-year mortalityrate of 7% for those with normal pulmonary wedgepressure at rest and with exercise, 28% with elevationof exercise pressures only, and 32% with elevation atrest and with exercise (there was no difference inejection fraction in these groups). Therefore, acquiring

echocardiographic data under or after physical stress(treadmill or bicycle) is an attractive concept to un-mask diastolic dysfunction absent at rest. The typicalechocardiographic parameters acquired during exer-cise or immediately thereafter are the E/e0 ratio andpeak tricuspid regurgitant velocity. Because the testis not primarily intended to provoke ischemia, it canbe kept to modest degrees of exertion, although thetest can also be integrated into a classic stress echo-cardiographic examination investigating ischemia,with minimal additional time and expense. If there isE-A and e0-a0 fusion under exercise tachycardia, mea-surement of the E/e0 ratio is performed after the heartrate falls, taking advantage of the delayed normaliza-tion of exercise-induced diastolic dysfunction aftertermination of exercise compared with heart rate.

Published studies of this test have shown thefollowing. Increases in the E/e0 ratio and pulmonarypressure, assessed by peak tricuspid regurgitant ve-locity, occur in a subset of patients with suspectedstress-inducible diastolic LV dysfunction (4,32). Thefrequency of exercise-inducible diastolic dysfunctionobviously depends on the selection of patients un-dergoing the test; in 1 study of 559 patients under-going stress echocardiography, it was 20% (32).Interestingly, only 9% of patients with post-exerciseE/e0 ratio >13, which in this study was taken as evi-dence of inducible diastolic dysfunction, also hadevidence of inducible ischemia by wall motion ab-normality assessment. An increase in the E/e0 ratiounder physical exercise indicates a concomitant


1081

increase in LVEDP (33). Increases in the E/e0 ratioindicate a worse prognosis (34,35).

However, many patients in whom such a dysfunc-tion would be suspected (e.g., patients with LV hy-pertrophy or multiple risk factors but no clear-cutechocardiographic signs of diastolic dysfunction atrest) have a normal diastolic stress test; although apositive test does seem to identify diastolic dys-function and impaired prognosis with high specificity(4,34), the sensitivity and negative predictive valueof the diastolic stress test are less clear, largely due tothe difficulty in obtaining an independent standard(i.e., pressure measurements) during exercise. In astudy with invasive hemodynamic validation, an ex-ercise E/e0 ratio >13 had a sensitivity of 73% and aspecificity of 96% for an exercise LVEDP >15 mm Hg(33). In 498 Korean patients with preserved ejectionfraction (34) who were specifically referred for a dia-stolic stress test, mainly to evaluate unexplaineddyspnea on exertion or “LV functional reserve,”171 (34%) developed exercise-induced pulmonaryhypertension, of whom only 49 (29%) developed anE/e0 ratio >15 under exercise; this again confirms theapparent low sensitivity of the diastolic stress test forinducible diastolic dysfunction. Exercise-inducedpulmonary hypertension had a distinctly worse out-come if associated with E/e0 ratio >15, whereas theoutcome of patients with pulmonary hypertensionwithout E/e0 ratio elevation was similar to that ofpatients without inducible pulmonary hypertension.By study criteria, none of these patients had anexercise-inducible ischemic wall motion response. Inan Australian study of 493 patients with an indicationfor exercise stress echocardiography and preservedejection fraction, 436 patients had a normal E/e0 ratioat rest, but only 41 (9%) developed an E/e0 ratio >13with exercise (35), and there was no relationship be-tween E/e0 ratio and the development of ischemia,which occurred in more than one-half of the patients.Follow-up over 1 year showed no cardiovasculardeaths, but there were more hospitalizations in thegroup with exercise-inducible increases in E/e0 ratiothan in patients without such increases. Publishedstudies of the diastolic stress test have examined dif-ferently selected groups of patients, which is evidentfrom the widely differing (between none and one-halfof patients) rates of exercise tests positive for ischemicwall motion abnormalities. This adds to the uncer-tainty regarding the diagnostic accuracy of the test.

STRAIN IMAGING. Longitudinal LV function has beenfound to be reduced in patients with HFpEF. This hasbeen assessed both by tissue Doppler and speckle-tracking strain imaging. A recent study comparing

longitudinal and circumferential global strain in pa-tients with manifest HFpEF, patients with hyperten-sion without heart failure, and control patients foundlower longitudinal and circumferential peak strainin patients with HFpEF (36).

Strain rate during the isovolumic relaxation time(IVRsr) or early diastolic strain rate (e0sr) (Figure 7),derived from global longitudinal speckle-trackingstrain, recently were proposed as a better substitutefor e0 to estimate LV filling pressures (37). AlthoughE/IVRsr correlated better with filling pressures thanE/e0sr, both parameters are noisy and IVRsr is espe-cially challenging to measure and thus is experimentalat best. Nevertheless, the parameter E/e0sr was shownto contain independent prognostic information in alarge longitudinal study of 1,048 survivors of myocar-dial infarction (38).

ECHOCARDIOGRAPHIC ASSESSMENT OF DIASTOLIC

FUNCTION DURING OR IMMEDIATELY AFTER ACUTE

ISCHEMIA. Echocardiographic signs of diastolic dys-function typically do not imply acute myocardialischemia. However, during acute supply ischemia(e.g., balloon inflation during coronary angioplasty),an abrupt reduction in the transmitral flow velocityintegral (affecting both E and A waves) is observed(39,40). Further, diastolic tissue velocities decrease orare even reversed in the ischemic region, and e0sr de-creases drastically. These diastolic changes in defor-mation during acute ischemia have been proposed asmarkers of acute ischemia and have been shown to lastlonger than wall motion abnormalities after reperfu-sion. Ishii et al. (41) showed that delayed relaxationcould be detected using a radial myocardial strainindex during and up to 10 min after induction ofischemia by treadmill stress echocardiography. Inmore recent work (42), a modification of this index,now based on global longitudinal strain, was corre-lated with simultaneous LV pre–A-wave pressure inpatients assessed for coronary disease; these patients,however, did not have acute ischemia.

TORSION. The clinical evaluation of the twisting mo-tion of the LV has become feasible as a by-product ofspeckle tracking–based circumferential strain mea-surements. Being largely still a research tool, there islittle standardization of measurements and even ofterminology. Recommendations from internationalechocardiographic societies (9) define rotation asrotational displacement around the long axis of the LV(in degrees); twist as the net difference between (small)rotation of the LV base and the larger rotation ofthe apical portion of the LV (in degrees)—in systole,the apex rotates counterclockwise (looking fromapex toward base) and the base clockwise; torsion as

FIGURE 7 Diastolic Strain Rate

Speckle tracking–based early diastolic longitudinal strain rate (red double arrow; 1/s) from the mid-septum (blue dot in left 4-chamber view) of

a patient with normal left ventricular function. AVC ¼ aortic valve closure.



1082

the ratio (or gradient) of twist and long-axis length(in degrees/cm).

Animal studies (11) have shown that the peak LVuntwisting rate in early diastole is a marker of diastolicfunction; the rate depends on LA pressure, activemyocardial relaxation, and elastic “restoring forces”stored during the buildup of systolic twist. Twist andtorsion increase with LV hypertrophy (43), volumeloading (44), physical endurance training (45), and age(43). However, the untwisting rate during isovolumicrelaxation reflects active relaxation and passiverestoring forces only, thus being largely load inde-pendent, although not independent of systolic func-tion because this feeds the passive restoring forces ofthe LV (46). Peak untwisting rate, which occurs in earlydiastole, has been calculated in healthy volunteers at115 � 40�/s, more than doubling during physical exer-cise (45,47) and increasing after volume loading (44),perhaps due to higher restoring forces built up byhigher previous systolic twist. Although hypertrophy,due to hypertension or cardiomyopathy, leads tohigher twist, patients with hypertrophic cardiomyop-athy (46) and patients with HFpEF have been reportedto have lower diastolic untwisting velocity (48)(Figure 8). Thus, the loss of rapid untwisting, due todecreased LV restoring forces, deprives the LV of itsearly diastolic suctionmechanism, leads to an increase

in LV diastolic pressures, and is a crucial event in theevolution to heart failure symptoms. Published reportsare not entirely consistent, however, and increasedsystolic twisting and diastolic untwisting rates wereobserved in a group of patients with mild diastolicdysfunction compared with healthy control patients,potentially related to LV hypertrophy and smallerend-systolic volumes in these patients (49).

In spite of their theoretical attractiveness, however,these rotational events have very small amplitudes—inhealthy humans, peak LV twist has been estimated atjust 11� 4� (47). Thus, the robustness of suchmeasuresin clinical practice remains to be determined.

DIASTOLICFUNCTIONASSESSEDBYECHOCARDIOGRAPHY

IN LARGE CLINICAL INTERVENTION TRIALS. Echocardio-graphic measures in clinical trials can be used toeither select the eligible participants or evaluate thetreatment effect. Echocardiographic parameters of LVdiastolic function have not been used as vigorouslyas systolic function measures for inclusion into clin-ical trials or as surrogate endpoints in clinical thera-peutic interventional trials. Several smaller phase IIclinical trials have used diastolic measures for par-ticipant selection and/or assessment of treatmenteffect. These trials have primarily focused on 2 dis-ease categories, hypertension and HFpEF, in whichdiastolic dysfunction is frequently present.

FIGURE 8 Left Ventricular Rotation

Rotation curve of patient with heart failure with preserved ejection fraction (left) and control patient (right) showing reduced magnitude of

peak apical rotation and percentage of early untwist (25% of total untwist duration), derived from speckle tracking 2-dimensional echocar-

diography. AVC ¼ aortic valve closure. Reproduced with permission from Tan et al. (48).


1083

USING DIASTOLIC ECHOCARDIOGRAPHIC PARAMETERS

IN PARTICIPANT SELECTION. Parameters of diastolicfunction are strong predictors of future cardiovascu-lar morbidity and mortality (50–54). Because diastolicfunction is not evaluated by a single measure, butthrough the interpretation of several echocardio-graphic parameters, using diastolic function as anentry criterion for trials generally requires sophisti-cated interpretation of data and is best suited for acore laboratory. The use of diastolic function asa measure for entry has been minimal in trials ofHFpEF, despite HFpEF initially being called “diastolicheart failure” (55–59). Diastolic echocardiographicmeasures have been used in patient selection only insmaller clinical trials (54,55). In the SIDAMI (Sildenafiland Diastolic Dysfunction After Acute MyocardialInfarction) trial, the investigators hypothesized thatsildenafil would reduce filling pressure during exer-cise in patients with diastolic dysfunction after my-ocardial infarction (60). The investigators used only2 measures of diastolic function as inclusion criteria:E/e0 ratio and LA volume index. Even so, they had toscreen approximately 1,300 patients by echocardiog-raphy to identify approximately 300 patients with an

E/e0 ratio $8 and an LA volume index $34 ml/m2. Thisillustrates that the complexity of diastolic functiongrading can make it challenging to identify appro-priate candidates to include in trials.

In the Use of Exercise and Medical Therapies toImprove Cardiac Function Among Patients WithExertional Shortness of Breath Due to Lung Congestiontrial (59), the investigators were aware of the lowspecificity of exertional shortness of breath to di-agnose diastolic dysfunction. They sought to over-come this problem by performing the evaluation ofdiastolic dysfunction during exercise. The inves-tigators therefore identified patients who had exer-tional shortness of breath and a coherent increase inthe E/e0 ratio to >13. Again, however, due to the diffi-culties cited in identifying patients with congruentsymptoms and echocardiographic parameters, the re-searchers only randomized 61 patients in this trial.

USING DIASTOLIC ECHOCARDIOGRAPHIC MEASURES TO

ASSESS TREATMENT EFFECT. Evaluation of a possibletreatment effect with diastolic echocardiographic pa-rameters has important limitations. Doppler velocitycurves, especially obtained from blood flow, show



1084

substantial interindividual and day-to-day variabilitywith changes in afterload, pre-load, and sympathe-tic tone. Additionally, they are highly susceptible tospectral broadening of the time-velocity curve,makingit difficult to accurately assess true peak velocities andto identify small changes in diastolic function withtreatment. Further, there is considerable interob-server and test-retest variability in diastolic echocar-diographic parameters, necessitating relatively largesample sizes to detect a treatment effect. Thus, echo-cardiographic core laboratories are crucial to assessdiastolic echocardiographic parameters in multicentertrials (61–63). Furthermore, even if a therapeuticintervention leads to an improved diastolic function,in the absence of clear-cut symptomatic or prognos-tic clinical benefit, the improvement in echocardio-graphic parameters alone would not be sufficientevidence for drug or device approval. Moreover, somemeasures are very age dependent (10,64,65). Hence,for example, in the VALIDD (Valsartan in DiastolicDysfunction) trial, patients with diastolic dysfunctionbased on lateral mitral e0 were enrolled—the exactcutoff that allowed entry was dependent on patientage (66). Furthermore, the U-shaped nature of the E/Aratio, isovolumic relaxation time, and decelerationtime in the progression through the diastolic functiongrades, with similar values seen in healthy individualsand patients with cardiac disease, make it impossibleto assess the point in the progress of diastolicdysfunction a specific participant is, if other diastolicmeasures are not included in the assessment. Lastly,sinus tachycardia and first-degree AV block can resultin partial or complete fusion of the mitral E and Awaves, making it difficult to assess the E/A ratio andthe deceleration time. These limitations make bloodflow Doppler measures especially poor as surrogateendpoints. On the other hand, e0, E/e0 ratio, and LAvolume incrementally change with worsening dia-stolic function, making them easier to use as surrogatediastolic endpoints in interventional trials.

Whereas Doppler velocities and time intervalsreflect filling pressures and diastolic relaxation at thetime of measurement, the LA volume reflects the cu-mulative effects of diastolic dysfunction and increasedfilling pressures over time and has been called the he-moglobin A1c of diastolic function. Trials using thismarker as a surrogate endpoint should therefore beplanned for a sufficient amount of time for the LA toundergo remodeling to a smaller volume. This isprobably one of the reasons why most trials testing atreatment of diastolic function have mainly focusedon e0 and/or the E/e0 ratio (59,66–68). In the PARA-MOUNT (Prospective Comparison of the AngiotensinReceptor Neprilysin Inhibitor With Angiotensin

Receptor Blocker on Management of Heart FailureWith Preserved Ejection Fraction) trial (69), theinvestigators assessed the effect of LCZ696 on echo-cardiographic measures as secondary endpoints; theprimary endpoint was the change in the levels of N-terminal pro–B-type natriuretic peptide (NT-proBNP).The investigators found that the only echocardio-graphic measure that improved, coherent with animprovement in the levels of NT-proBNP, was LA size(both defined by the diameter and the volume). Incomparison, the authors found no change in e0 or theE/e0 ratio. This emphasizes that LA size, which reflectssustained increase in LVfilling pressure,may be amorerobust metric of diastolic function than Doppler-derived measures, which are subject to greater vari-ability and are more sensitive to changes in afterload,pre-load, and sympathetic tone.

Recent studies have focused on assessing LA func-tion instead of maximal size as a sensitive marker ofdiastolic dysfunction and morbidity (70–72). Becausethe LA is a thin-walled structure, continuouslyexposed to LV pressure and function, it may provideinformation on pre-clinical and early impairment ofthe LV. Measures of LA function have been demon-strated to identify minuscule impairment of LV func-tion even when ejection fraction is normal (72) and toidentify patients with hidden paroxysmal atrial fibril-lation following a cryptogenic ischemic stroke as theonly echocardiographic parameter (70). However, in-vestigators using LA measures of structure and func-tion have to bear in mind that the pathophysiologicalfactors that affect function and maximal size aredifferent. Parameters of LA function (e.g., longitudinalshortening, minimal size, emptying fraction) aredetermined by LA contractility and the instantaneousloading conditions, which strongly depend in turnon LV properties and pulmonary vein compliance,whereas the maximal LA size is a marker of chronicpressure overload causedmainly by severe LVdiastolicdysfunction.

LARGE INTERVENTION TRIALSASSESSING THE EFFECTOF

TREATMENT ON DIASTOLIC FUNCTION. Only a few largeinterventional phase II trials have assessed the treat-ment effect on diastolic dysfunction as a surrogateprimary endpoint. They generally fall into the categoryof diastolic dysfunction due to hypertension (66,67) ordiastolic function in HFpEF (59,68). The aim of theVALIDD trial was to determinewhether lowering bloodpressure with an angiotensin receptor blocker wouldimprove diastolic function to a greater extent thanwould pharmacological approaches not based on in-hibition of the renin-angiotensin-aldosterone axis.The investigators found that lowering blood pressure


1085

improved diastolic function, defined by an increasein e0 at follow-up, irrespective of the type of antihy-pertensive agent used (66). They also found that theparticipants in the valsartan group improved the iso-volumic relaxation time and longitudinal systolic peaktissue velocity. Whether the improvement in earlyrelaxation was due to an actual improvement in thediastolic function of the heart or due to decreasedafterload, which is well known to improve the rate ofearly relaxation (73,74), remains unknown.

In the EXCEED (Exforge Intensive Control of Hy-pertension to Evaluate Efficacy in Diastolic Dysfunc-tion) trial, the aim was to test whether intensive bloodpressure control among patients with uncontrolledhypertension and diastolic dysfunctionwould bemoreeffective in improving diastolic function than standardblood pressure control (70). The investigators foundthat the degree of blood pressure reduction observedwas the most important determinant of the degreeof improvement in diastolic function, determinedby e0. However, they found no difference in the im-provement in e0 between the intensive blood pressuretreatment group and the standard blood pressuretreatment group. Isovolumic relaxation time wasthe only diastolic measure that showed a statisticallysignificant improvement in the intensive treatmentgroup compared with the standard treatment group,which is comparable to the results of the VALIDD trial(66). Again, whether these improvements in the dia-stolic function actually represent improved diastolicfunction or rather changes in the diastolic measuresaccompanying decreased afterload is unclear.

The aim of the Aldo-DHF (Aldosterone ReceptorBlockade in Diastolic Heart Failure) trial was to eval-uate the effect of spironolactone on diastolic functionand exercise capacity in patients with HFpEF (68). Theinvestigators found that spironolactone improveddiastolic function, as determined by the E/e0 ratio (andnatriuretic peptide levels). This improvement infunction (including e0, E, LV ejection fraction andmassindex, and LV end-diastolic diameter) was not associ-ated with improvement in the maximal exercise ca-pacity, patient symptoms, or quality of life. Again, as inthe VALIDD and EXCEED trials, the improvement inthe E/e0 ratio was accompanied by a highly significantreduction in blood pressure. However, the in-vestigators demonstrated that after adjustment forbaseline and follow-up blood pressure values, the ef-fects of spironolactone on diastolic function (E/e0 ratio)remained statistically significant, suggesting that thereverse remodeling effects of spironolactone were in-dependent of blood pressure reduction. Nevertheless,blood pressure is only an indirect measure of LVafterload, and correction for blood pressure reduction

in the statistical model may not be adequate to refute aprimary effect of changes in loading conditions toexplain the differences seen between the 2 groups.Additionally, even though spironolactone improvedan echocardiographic measure of diastolic function,the TOPCAT (Treatment of Preserved Cardiac FunctionHeart Failure With an Aldosterone Antagonist) trialrecently demonstrated no benefit of spironolactone inpreventing adverse outcomes in HFpEF (75). In thistrial, the E/e0 ratio was among the strongest prognos-ticators in patients with HFpEF, re-emphasizing theimportance of diastolic impairment in the progressionof the disease.

ASSESSMENT OF DIASTOLIC FUNCTION:

NUCLEAR IMAGING

From gated radionuclide ventriculography data, LVfilling rates can be calculated by comparing sequentialLV volumes during diastole, which can be calculatedwith frame rates of approximately 20/s. From these,time-volume filling curves, peak filling rates, and timeto peakfilling rates can be determined (Figure 9), whichdecrease with slowed LV relaxation. Low time resolu-tion and pre-load dependence of filling rates affecttheir interpretation; thus, nuclear imaging is rarelyused to assess diastolic function.


CARDIAC MAGNETIC RESONANCE

Cardiac magnetic resonance (CMR) is generally regar-ded as the gold standard for assessing the volumeof cardiac chambers and, as such, is well suited toprovide LV volume changes over time, which can beconverted to filling curves routinely by analysispackages (Figure 9). Similar to the less accurate andtemporally coarser nuclear ventriculography data,peak filling rates and time to peak filling are thefundamental parameters. However, in a study exam-ining the use of early and late LV peak filling ratesin assessing iron load in thalassemia, which can bedirectly estimated by T2* CMR (76), these parameterswere not found to be useful. Blood flow velocity canbe measured directly (i.e., not via volumetric fillingrates) by velocity encoding or “phase-contrast” CMR.Hence, transmitral and pulmonary vein flow canbe measured and used in a manner analogous toechocardiographic Doppler values (77). In addition,tissue velocity (e.g., of the ventricular walls) can bemeasured, although only small validation studiesexist (78). Of course, CMR is also an established tech-nique to measure LA volume (79); however, system-atically and substantially higher LA volume values byCMR than by echocardiography have to be taken into

FIGURE 9 Volume Curves From Cardiac Magnetic Resonance of the LV

Blood volumeED ES

ml

ms0 100 200 300 400 600 800500 700 900

0

50

100

150

200

250

ED ES

ml

ms

00

50

100

150

200

250

300

350

400

450

500

50100

150200

250300

350400

450500

550600

650700

750

LV volume on the y-axis and time after R wave on the x-axis. (Top) Normal filling in a

young woman without cardiovascular disease. (Bottom) Slowed diastolic filling in a patient

with long-standing severe hypertension and LV hypertrophy (mass, 159 g/m2; note the

increased LV volume). The red squares identify inflection points of the curve, which can be

used for quantification. Such curves can also be generated from nuclear or computed

tomography data. ED ¼ end-diastole, ES ¼ end-systole; LV ¼ left ventricular.



1086

account. Newer reports also have looked at the pre-dictive value of CMR parameters of LA function (e.g.,minimal volume or LA strain by CMR feature tracking)and found predictive value of such measures for thedevelopment of heart failure (80).

Strain, strain rate, and torsion can be extractedfrom CMR data by 3 entirely different approaches:1) tagging, in which material points in the myocardiumare marked by “nulling” the signal from these pointsat baseline (e.g., in a grid pattern) and then followingthese points in the myocardium over time (Figure 10);2) feature tracking, which is akin to speckle tracking

by echocardiography—this is a post-processing tech-nique applied to CMR or computed tomography (CT)images that primarily depends on the characteristicsof the images (e.g., contrast, noise), not the under-lying imaging technique; and 3) velocity encoding(or phase-contrast) CMR; this technique, which isstandard in CMR for blood flow velocity measurement,can be applied to tissue as well, analogous to tissueDoppler imaging in echocardiography, and the datacan be processed to estimate strain (81).

Higher torsion by CMR has been found in womenand older individuals, which is ascribed to theconcomitant increase in concentric remodeling andrelative wall thickness (43). These researchers alsosaw lower circumferential strain in older individuals.

The prognostic utility of CMR indexes of diastolicstrain was evaluated in a substudy from the MESA(Multi-Ethnic Study of Atherosclerosis) project, alarge-scale prospective, population-based observa-tional study (82). From CMR tagging, circumferentialLV strain and strain rate were calculated; in addition topeak torsion rate and early diastolic circumferentialstrain rate, a new index of relaxation analogous to theisovolumic relaxation time was computed from thesecurves in 1,544 patients. Although torsion rate and e0srwere of no or little value, the new index predicted(independently from clinical variables, ejection frac-tion, and natriuretic peptide levels) the occurrence ofheart failure and/or atrial fibrillation over a meanfollow-up of 8 years. Similar findings were reportedmore than a decade agowith simple echocardiographicmeasurements of LV systolic and diastolic function,particularly peak E velocity (83). Similarly, a baselineechocardiographic pattern of impaired relaxation,combined with LA enlargement, was found to be pre-dictive of heart failure or atrial fibrillation duringfollow-up (84).

Recently, in a small study of women with “signsand symptoms” of ischemia but no epicardial coro-nary stenosis, CMR was used to evaluate both systolicand diastolic LV function by strain and torsion(Figure 10). Although no differences in systolic strainwere found, peak (early) diastolic strain rate and peakuntwisting rate seemed to differentiate these “syn-drome X” patients from healthy individuals (85).

An entirely different CMR-specific, albeit indirect,way to assess diastolic function is based on theconcept of “T1 mapping,” which calculates T1 (orlongitudinal) proton relaxation times (in ms) after anelectromagnetic pulse and depicts them as pixel in-tensity (86,87). Pathological changes of T1 times inmyocardial space may be due to excess water, iron, orprotein deposition. T1 mapping can be carried outwith or without gadolinium contrast (“native” T1

FIGURE 10 LV Circumferential and Torsional Strain Derived From CMR Tagging

(A) Representative cardiac magnetic resonance (CMR) image obtained at the level of the papillary muscles, with a tissue tagging sequence

(Top). Analysis was performed using commercially available software (HARP, Diagnosoft, Durham, North Carolina), which constructs a mesh

tracking tissue deformation over a single cardiac cycle, using the harmonic phase. (B) Representative tracing of circumferential strain (solid

line) and circumferential strain rate (dashed line) over a single cardiac cycle. (C) Representative data tracing of apical (red line) and basal (blue

line) rotation, net torsion (solid black line), and the rate of ventricular twisting (dashed line) over a single cardiac cycle. Reproduced with

permission from Nelson et al. (85).


1087



1088

mapping vs. extracellular volume imaging). Whereasnative T1 mapping encompasses both intracellularand extracellular changes, the so-called extracellularvolume technique, which requires gadolinium-basedcontrast application, allows quantification of the ex-tracellular volume fraction of myocardium, which is asurrogate for myocardial fibrosis, interstitial edema,and others (Figure 11). Different than late enhance-ment after gadolinium contrast injection, which vi-sualizes localized increases in extracellular volumeand has become commonplace for myocardial scarimaging, T1 mapping does not require focal differ-ences and thus allows quantification of diffuse my-ocardial changes such as nonfocal edema or fibrosis.T1 mapping of extracellular volume gives no specificinformation why the extracellular space is increasedand what kind of material fills it—it only measures thefraction of the total myocardial volume that is extra-cellular. In a small series of post–heart transplant

FIGURE 11 Determination of ECV Fraction by T1 Imaging With

Contrast Cardiac Magnetic Resonance

T1-based extracellular volume map in which signal intensity is

proportional to extracellular volume fraction (the percentage of

total tissue volume that is extracellular). The map was calculated

from T1 times before and after contrast application, and extra-

cellular volume fraction was calibrated by measuring extracel-

lular volume (ECV) in a blood sample as (1 � hematocrit) (86). In

the sample region of interest in the myocardium, the average

ECV fraction was 22%.

patients, a modest correlation of T1 times with theconstant of chamber stiffness was found. Impor-tantly, in this small group of patients, there was nosignificant relation of T1 relaxation times to inva-sively measured LV diastolic pressure, the relaxationconstant tau, or near simultaneous echocardiographicE/e0 ratio, thus cautioning against a too simplisticequation between T1 times and diastolic function(88). In another study comparing patients with heartfailure with reduced or preserved ejection fractionwith control patients, patients with HFpEF had largerextracellular volume fractions (29%) than controlpatients (27%) but smaller than patients with reducedejection fraction (31%); however, the differenceswere very small (89). Another study found a weakcorrelation of the E/e0 ratio with myocardial T1 timesafter contrast application and in a small subgroupwith myocardial biopsy, a good inverse correlationbetween T1 time after contrast application andextracellular volume fraction (90). Thus, there areconflicting results for T1-related techniques, fibrosis,and diastolic function. The technique still has tobe considered experimental, with a clear need formore validation and establishment of procedural andanalytical standards. Thus, although CMR techniquesat present allow the characterization of tissue to adegree and thus allow for unique diagnostic infor-mation, their utility for diagnosing and followingdiastolic LV function remains limited and impractical.

Finally, CMR is a well-established modality for thedifficult diagnosis of constrictive pericarditis, forwhich it not only can directly assess pericardial thick-ening but also help diagnose the amount of respiratoryvariation in left and right ventricular filling (91).


CARDIAC CT

Because of radiation exposure and the availability ofother techniques for this purpose, CT has only spo-radically been used to assess diastolic function. Similarto nuclear imaging and CMR, it provides LV volumedata, albeit at a modest temporal resolution. An im-portant application of CT is the detailedmorphologicaldiagnosis of constrictive pericarditis.

ROLE OF AGE AND PHYSICAL EXERCISE FOR

THE LIMITS OF NORMALCY

A vexing question is the extent to which age-relatedchanges in diastolic function, and their correlates onimaging, are truly related to aging and thus unavoid-able and which changes are due to modifiable factors.The relaxation constant tau has been found by some(92), but not all (93), investigators to increase with age.


1089

LV chamber stiffness, the inverse of LV compliance,increases with age in sedentary, but not in highlyendurance-trained, individuals (94). Therefore, itis not clear how much of the changes observed indiastolic imaging parameters in older adults (e.g.,decreased transmitral E/A ratio) are due to aging aloneand how much is caused by either diseases, such ashypertension, or physical deconditioning (“sedentarylifestyle”). It may well be that aging plays a relativelymodest role in the changes observed in elderly in-dividuals. For example, higher transmitral peak E/Aratios have been found in athletes than in sedentaryindividuals of the same age, particularly in olderadults (95–98). A study of nearly 3,000 middle-agedindividuals participating in a fitness program (98)found these effects not only in high-level athletes butalso in less trained individuals. Individualswith higherexercise capacity had less concentric LV remodeling(but higher LV mass in men), smaller relative wallthickness, lower E/e0 ratios, and higher E/A ratios thanless fit individuals. Interestingly, LA and LV sizeincreased with exercise capacity, indicating that it isnot unambiguously related to diastolic function.Ejection fraction, on the other hand, was not related toexercise capacity. Several trials have shown positiveeffects of exercise training on patients with HFpEFand indirectly demonstrated improved diastolic func-tion (99–101).

An important question is what significance imagingsigns of diastolic dysfunction have in patientswithout clear-cut symptoms. This situation has beentermed pre-clinical diastolic dysfunction, a suggestivename in need of objective evidence from longitudinalstudies. Several large studies have suggested thateven mild diastolic abnormalities in the presence ofpreserved LV ejection fraction imply an impairedprognosis independently from age and baselinesymptoms (102–104). On the other hand, population-based studies have shown a high prevalence ofasymptomatic diastolic dysfunction based on echo-cardiographic parameters. Abhayaratna et al. (105)showed that in an Australian population aged $60years with preserved ejection fraction, 23.5% hadmild and 5.6% had moderate or severe diastolic dys-function by echocardiography; even of the lattergroup, 36% were asymptomatic. In contrast, only0.5% of the group with reduced ejection fraction wasasymptomatic. Similar numbers have been reportedfrom the Framingham Heart Study in the UnitedStates (106), with 36% of elderly participants havingasymptomatic diastolic dysfunction and 5% asymp-tomatic systolic dysfunction. From Italy, similarvalues for silent diastolic dysfunction (35% of a pop-ulation 65 to 84 years old), again far more than

for systolic dysfunction, have been documented(107). A retrospective study following patients who atbaseline had moderate or severe diastolic dysfunctionon echocardiography, but no heart failure symptoms,showed that only 12% went on to develop overt heartfailure over the next 3 years (108). However, in pa-tients with diabetes, asymptomatic diastolic dys-function and, in particular, E/e0 ratio >15 (septal e0)have been shown to predict more than a doubling ofthe rate of overt heart failure over the next 5 years(109,110).

OPEN QUESTIONS AND UNMET NEEDS

There continue to be large gaps in our understandingof diastolic heart failure, and delayed relaxation,reduced compliance, and atrial dysfunction may havevery separate causes and yet all produce the clinicalsyndrome of diastolic dysfunction. Further, whetherthis condition is due to a specific myocardial diseaseor results from a myocardial response to unfavorableworking conditions, in particular arterial stiffening,remains unclear.

The issue of diastolic function grading remains hotlydebated. A large observational study reported thatpatients frequently (17% of patients examined at aclinical echocardiography laboratory) had intermedi-ate features between grades 1 and 2 (E/A ratio #0.75,deceleration time>140ms, and E/e0 ratio$10) and hada distinctly worse prognosis than those with classicgrade 1 dysfunction (differing in that E/e0 ratio #8),which was similar to the prognosis of grade 2 dys-function (111). This classification purports to reflectdifferent severities and advancement of diastolicdysfunction, and for some diseases, like amyloidosis,this has been convincingly shown (112). However, it isbased on echocardiographic patterns and thus indirectsigns of elevated diastolic LV pressures, not any spe-cific disease pathology. It is also unclear whetherthe existing stages are optimal for clinical use. Anadditional subdivision of stage 1 (defined by E/Aratio #0.75, deceleration time >140 ms, and E/e0

ratio <10) has been suggested, with the same limits forE/A ratio and deceleration time but E/e0 ratio $10 (111).Others have proposed to add reversibility of stage 3after therapy as a stratifying criterion. More epidemi-ological data are needed to better understand theclinical and prognostic value and perhaps to modifythis classification.

Mitral regurgitation considerably complicates theinterpretation of imaging data with regard to LVdiastolic function: it elevates diastolic pressures inthe LA and LV, causes chronic dilation of the LA, in-creases pulmonary capillary pressure and pulmonary



1090

pressures, and leads to the clinical picture of heartfailure. Most of the routinely evaluated parameters ofdiastolic function, such as transmitral and pulmonaryvenous flow patterns, E/e0 ratio, systolic pulmonarypressure, LA size, and others, are affected by mitralregurgitation indistinguishably from primarily LVmyocardial dysfunction. In the presence of grossmorphological abnormalities of the mitral valve (e.g.,a flail leaflet), the differential diagnosis is relativelyeasy; however, this is not the case in functional mitralregurgitation, and separation of valvular andmyocardial etiology is particularly difficult in patientswith atrial fibrillation and degenerative valvularchanges like annular calcification, among otherconditions.

No established drug or other therapy for HFpEFexists, perhaps with the exception of physical exercisetraining. Given the heterogeneity of HFpEF pop-ulations, it may well be that only targeted therapies(113), addressing specific subgroups of such patients,will be successful. This underscores the need of “deepphenotyping” these patients, a task for which cardiacimaging with its ever more differentiated, multi-pronged armamentarium appears well suited.

REPRINT REQUESTS AND CORRESPONDENCE: Dr.Frank A. Flachskampf, Uppsala Universitet, Institu-tionen för Medicinska Vetenskaper, AkademiskaSjukhuset, Ingång 40, Plan 5, 751 85 Uppsala, Sweden.E-mail: [email protected].

RE F E RENCE S

1. Paulus WJ, Tschope C, Sanderson JE, et al. Howto diagnose diastolic heart failure: a consensusstatement on the diagnosis of heart failure withnormal left ventricular ejection fraction by theHeart Failure and Echocardiography Associationsof the European Society of Cardiology. Eur Heart J2007;28:2539–50.

2. Zile MR, Baicu CF, Gaasch WH. Diastolic heartfailure—abnormalities in active relaxation andpassive stiffness of the left ventricle. N Engl J Med2004;350:1953–9.

3. Kitzman DW, Higginbotham MB, Cobb FR,Sheikh KH, Sullivan MJ. Exercise intolerance in pa-tients with heart failure and preserved left ventric-ular systolic function: failure of the Frank-Starlingmechanism. J Am Coll Cardiol 1991;17:1065–72.

4. Ha JW, Oh JK, Pellikka PA, et al. Diastolic stressechocardiography: a novel noninvasive diagnostictest for diastolic dysfunction using supine bicycleexercise Doppler echocardiography. J Am SocEchocardiogr 2005;18:63–8.

5. Nagueh SF, Middleton KJ, Kopelen HA,Zoghbi WA, Quiñones MA. Doppler tissue imaging:a noninvasive technique for evaluation of leftventricular relaxation and estimation of fillingpressures. J Am Coll Cardiol 1997;30:1527–33.

6. Opdahl A, Remme EW, Helle-Valle T, et al. De-terminants of left ventricular early-diastolic length-ening velocity: independent contributions from leftventricular relaxation, restoring forces, and length-ening load. Circulation 2009;119:2578–87.

7. Hasegawa H, Little WC, Ohno M, et al. Diastolicmitral annular velocity during the development ofheart failure. J Am Coll Cardiol 2003;41:1590–7.

8. Iwano H, Pu M, Upadhya B, Meyers B,Vlachos P, Little WC. Delay of left ventricularlongitudinal expansion with diastolic dysfunction:impact on load dependence of e0 and longitudinalstrain rate. Physiol Rep 2014;2:e12082.

9. Mor-Avi V, Lang RM, Badano LP, et al. Currentand evolving echocardiographic techniques for thequantitative evaluation of cardiac mechanics: ASE/EAE consensus statement on methodology andindications endorsed by the Japanese Society of

Echocardiography. Eur J Echocardiogr 2011;12:167–205.

10. Nagueh S, Appleton C, Evangelista A, et al.Recommendations for the echocardiographicdiagnosis of diastolic dysfunction. Eur J Echo-cardiogr 2009;10:165–93.

11. Opdahl A, Remme EW, Helle-Valle T,Edvardsen T, Smiseth OA. Myocardial relaxation,restoring forces, and early-diastolic load are in-dependent determinants of left ventricularuntwisting rate. Circulation 2012;126:1441–51.

12. Popovi�c ZB, Desai MY, Buakhamsri A, et al.Predictors of mitral annulus early diastolic veloc-ity: impact of long-axis function, ventricular fillingpattern, and relaxation. Eur J Echocardiogr 2011;12:818–25.

13. Nikolic S, Yellin EL, Tamura K, et al. Passiveproperties of canine left-ventricle—diastolic stiff-ness and restoring forces. Circ Res 1988;62:1210–22.

14. Remme EW, Opdahl A, Smiseth OA. Mechanicsof left ventricular relaxation, early diastoliclengthening, and suction investigated in a math-ematical model. Am J Physiol Heart Circ Physiol2011;300:H1678–87.

15. Rademakers FE, Buchalter MB, Rogers WJ,et al. Dissociation between left ventricularuntwisting and filling. Accentuation by catechol-amines. Circulation 1992;85:1572–81.

16. Firstenberg MS, Smedira NG, Greenberg NL,et al. Relationship between early diastolic intra-ventricular pressure gradients, an index of elasticrecoil, and improvements in systolic and diastolicfunction. Circulation 2001;104 12 Suppl 1:I330–5.

17. Appleton CP, Hatle LK, Popp RL. Demonstra-tion of restrictive ventricular physiology byDoppler echocardiography. J Am Coll Cardiol1988;11:757–68.

18. Flachskampf FA, Weyman AE, Guerrero JL,Thomas JD. Calculation of atrioventricularcompliance from the mitral flow profile: analyticand in vitro study. J Am Coll Cardiol 1992;19:998–1004.

19. Little WC, Ohno M, Kitzman DW, Thomas JD,Cheng CP. Determination of left ventricularchamber stiffness from the time for decelerationof early left ventricular filling. Circulation 1995;92:1933–9.

20. Garcia MJ, Firstenberg MS, Greenberg NL,et al. Estimation of left ventricular operatingstiffness from Doppler early filling decelerationtime in humans. Am J Physiol Heart Circ Physiol2001;280:H554–6.

21. Myreng Y, Smiseth OA. Assessment of leftventricular relaxation by Doppler echocardiogra-phy. Comparison of isovolumic relaxation time andtransmitral flow velocities with time constant ofisovolumic relaxation. Circulation 1990;81:260–6.

22. Schirmer H, Lunde P, Rasmussen K. Mitral flowderived Doppler indices of left ventricular diastolicfunction in a general population: the Tromsostudy. Eur Heart J 2000;21:1376–86.

23. Myreng Y, Smiseth OA, Risøe C. Left ventric-ular filling at elevated diastolic pressures: rela-tionship between transmitral Doppler flowvelocities and atrial contribution. Am Heart J1990;119:620–6.

24. Appleton CP, Galloway JM, Gonzalez MS,Gaballa M, Basnight MA. Estimation of left ven-tricular filling pressures using two-dimensionaland Doppler echocardiography in adult patientswith cardiac disease. Additional value of analyzingleft atrial size, left atrial ejection fraction and thedifference in duration of pulmonary venous andmitral flow velocity at atrial contraction. J Am CollCardiol 1993;22:1972–82.

25. Rossvoll O, Hatle L. Pulmonary venous flowvelocities recorded by transthoracic Doppler ul-trasound: relation to left ventricular diastolicpressures. J Am Coll Cardiol 1993;21:1687–96.

26. Nishimura RA, Abel MD, Hatle LK, Tajik AJ.Relation of pulmonary vein to mitral flow veloc-ities by transesophageal Doppler echocardiogra-phy. Effect of different loading conditions.Circulation 1990;81:1488–97.

27. Ishida Y, Meisner JS, Tsujioka K, et al. Leftventricular filling dynamics: influence of left

mailto:[email protected]

http://refhub.elsevier.com/S1936-878X(15)00509-4/sref1






































































































































1091

ventricular relaxation and left atrial pressure. Cir-culation 1986;74:187–96.

28. Ohno M, Cheng CP, Little WC. Mechanism ofaltered patterns of left ventricular filling duringthe development of congestive heart failure. Cir-culation 1994;89:2241–50.

29. Mullens W, Borowski AG, Curtin RJ,Thomas JD, Tang WH. Tissue Doppler imaging inthe estimation of intracardiac filling pressure indecompensated patients with advanced systolicheart failure. Circulation 2009;119:62–70.

30. Miura T, Miyazaki S, Guth BD, Indolfi C,Ross J Jr. Heart rate and force-frequency effectson diastolic function of the left ventricle in exer-cising dogs. Circulation 1994;89:2361–8.

31. Dorfs S, Zeh W, Hochholzer W, et al. Pulmo-nary capillary wedge pressure during exercise andlong-term mortality in patients with suspectedheart failure with preserved ejection fraction. EurHeart J 2014;35:3103–12.

32. Gibby C, Wiktor DM, Burgess M, Kusunose K,Marwick TH. Quantitation of the diastolic stresstest: filling pressure vs. diastolic reserve. EurHeart J Cardiovasc Imaging 2013;14:223–7.

33. Burgess MI, Jenkins C, Sharman JE,Marwick TH. Diastolic stress echocardiography:hemodynamic validation and clinical significanceof estimation of ventricular filling pressure withexercise. J Am Coll Cardiol 2006;47:1891–900.

34. Shim CY, Kim SA, Choi D, et al. Clinical out-comes of exercise-induced pulmonary hyperten-sion in subjects with preserved left ventricularejection fraction: implication of an increase in leftventricular filling pressure during exercise. Heart2011;97:1417–24.

35. Holland DJ, Prasad SB, Marwick TH. Prognosticimplications of left ventricular filling pressure withexercise. Circ Cardiovasc Imaging 2010;3:149–56.

36. Kraigher-Krainer E, Shah AM, Gupta DK, et al.Impaired systolic function by strain imaging inheart failure with preserved ejection fraction. J AmColl Cardiol 2014;63:447–56.

37. Wang J, Khoury DS, Thohan V, Torre-Amione G, Nagueh SF. Global diastolic stain ratefor the assessment of left ventricular relaxationand filling pressures. Circulation 2007;115:1376–83.

38. Ersbøll M, Andersen MJ, Valeur N, et al. Earlydiastolic strain rate in relation to systolic anddiastolic function and prognosis in acute myocar-dial infarction: a two-dimensional speckle-trackingstudy. Eur Heart J 2014;35:648–56.

39. Kukulski T, Jamal F, D’Hooge J, Bijnens B, DeScheerder I, Sutherland GR. Acute changes insystolic and diastolic events during clinical coro-nary angioplasty: a comparison of regional veloc-ity, strain rate and strain measurement. J Am SocEchocardiogr 2002;15:1–12.

40. Werner GS, Sold G, Teichmann D, Andreas S,Kreuzer H, Wiegand V. Impaired relationship be-tween Doppler echocardiographic parameters ofdiastolic function and left ventricular filling pres-sure during acute ischemia. Am Heart J 1990;120:63–72.

41. Ishii K, Imai M, Suyama T, et al. Exercise-induced post-ischemic left ventricular delayed

relaxation or diastolic stunning: is it a reliablemarker in detecting coronary artery disease? J AmColl Cardiol 2009;53:698–705.

42. Chiang SJ, Daimon M, Ishii K, et al. Assess-ment of elevation of and rapid change in leftventricular filling pressure using a novel globalstrain imaging diastolic index. Circ J 2014;78:419–27.

43. Yoneyama K, Gjesdal O, Choi EY, et al. Age,sex, and hypertension-related remodeling in-fluences left ventricular torsion assessed by tag-ged cardiac magnetic resonance in asymptomaticindividuals: the Multi-Ethnic Study of Atheroscle-rosis. Circulation 2012;126:2481–90.

44. Weiner RB, Weyman AE, Khan AM, et al. Pre-load dependency of left ventricular torsion: theimpact of normal saline infusion. Circ CardiovascImaging 2010;3:672–8.

45. Weiner RB, Hutter AM Jr., Wang F, et al. Theimpact of endurance exercise training on leftventricular torsion. J Am Coll Cardiol Img 2010;3:1001–9.

46. Notomi Y, Popovic ZB, Yamada H, et al. Ven-tricular untwisting: a temporal link between leftventricular relaxation and suction. Am J PhysiolHeart Circ Physiol 2008;294:H505–13.

47. Notomi Y, Martin-Miklovic MG, Oryszak SJ,et al. Enhanced ventricular untwisting during ex-ercise: a mechanistic manifestation of elastic recoildescribed by Doppler tissue imaging. Circulation2006;113:2524–33.

48. Tan YT, Wenzelburger F, Lee E, et al. Thepathophysiology of heart failure with normalejection fraction: exercise echocardiography re-veals complex abnormalities of both systolic anddiastolic ventricular function involving torsion,untwist, and longitudinal motion. J Am Coll Car-diol 2009;54:36–46.

49. Park SJ, Miyazaki C, Bruce CJ, Ommen S,Miller FA, Oh JK. Left ventricular torsion by two-dimensional speckle tracking echocardiography inpatients with diastolic dysfunction and normalejection fraction. J Am Soc Echocardiogr 2008;21:1129–37.

50. Møller JE, Hillis GS, Oh JK, et al. Left atrialvolume: a powerful predictor of survival afteracute myocardial infarction. Circulation 2003;107:2207–12.

51. Meta-Analysis Research Group in Echocardi-ography (MeRGE) AMI Collaborators, Møller JE,Whalley GA, et al. Independent prognosticimportance of a restrictive left ventricular fillingpattern after myocardial infarction: an individualpatient meta-analysis: Meta-Analysis ResearchGroup in Echocardiography acute myocardialinfarction. Circulation 2008;117:2591–8.

52. Hillis GS, Møller JE, Pellikka PA, et al. Nonin-vasive estimation of left ventricular filling pres-sure by E/e0 is a powerful predictor of survivalafter acute myocardial infarction. J Am Coll Cardiol2004;43:360–7.

53. Wang M, Yip G, Yu CM, et al. Independent andincremental prognostic value of early mitralannulus velocity in patients with impaired leftventricular systolic function. J Am Coll Cardiol2005;45:272–7.

54. Bella JN, Palmieri V, Roman MJ, et al. Mitralratio of peak early to late diastolic filling velocityas a predictor of mortality in middle-aged andelderly adults: the Strong Heart Study. Circulation2002;105:1928–33.

55. Massie BM, Carson PE, McMurray JJ, et al.Irbesartan in patients with heart failure and pre-served ejection fraction. N Engl J Med 2008;359:2456–67.

56. Pitt B, Pfeffer MA, Assmann SF, et al. Spi-ronolactone for heart failure with preserved ejec-tion fraction. N Engl J Med 2014;370:1383–92.

57. Cleland JGF, Tendera M, Adamus J, et al. ThePerindopril in Elderly People With Chronic HeartFailure (PEP-CHF) study. Eur Heart J 2006;27:2338–45.

58. Yusuf S, Pfeffer MA, Swedberg K, et al. Effectsof candesartan in patients with chronic heart fail-ure and preserved left-ventricular ejection frac-tion: the CHARM-Preserved trial. Lancet 2003;362:777–81.

59. Kosmala W, Holland DJ, Rojek A, et al. Effectof If-channel inhibition on hemodynamic statusand exercise tolerance in heart failure with pre-served ejection fraction: a randomized trial. J AmColl Cardiol 2013;62:1330–8.

60. Andersen MJ, Ersbøll M, Axelsson A, et al.Sildenafil and diastolic dysfunction after acutemyocardial infarction in patients with preservedejection fraction: the Sildenafil and DiastolicDysfunction After Acute Myocardial Infarction(SIDAMI) trial. Circulation 2013;127:1200–8.

61. Hole T, Otterstad JE, St. John Sutton M,Frøland G, Holme I, Skjaerpe T. Differences be-tween echocardiographic measurements of leftventricular dimensions and function by local in-vestigators and a core laboratory in a 2-yearfollow-up study of patients with an acutemyocardial infarction. Eur J Echocardiogr 2002;3:263–70.

62. Baur LH, Schipperheyn JJ, van der Velde EA,et al. Reproducibility of left ventricular size, shapeand mass with echocardiography, magnetic reso-nance imaging and radionuclide angiography inpatients with anterior wall infarction. A plea forcore laboratories. Int J Card Imaging 1996;12:233–40.

63. Douglas PS, DeCara JM, Devereux RB, et al.Echocardiographic imaging in clinical trials:American Society of Echocardiography standardsfor echocardiography core laboratories: endorsedby the American College of Cardiology Founda-tion. J Am Soc Echocardiogr 2009;22:755–65.

64. Mogelvang R, Sogaard P, Pedersen SA,Olsen NT, Schnohr P, Jensen JS. Tissue Dopplerechocardiography in persons with hypertension,diabetes, or ischaemic heart disease: the Copen-hagen City Heart Study. Eur Heart J 2009;30:731–9.

65. Henein M, Lindqvist P, Francis D, Mörner S,Waldenström A, Kazzam E. Tissue Doppler analysisof age-dependency in diastolic ventricular behav-iour and filling: a cross-sectional study of healthyhearts (the Umeå General Population HeartStudy). Eur Heart J 2002;23:162–71.

66. Solomon SD, Janardhanan R, Verma A, et al.Effect of angiotensin receptor blockade and












































































































































































































1092

antihypertensive drugs on diastolic function in pa-tients with hypertension and diastolic dysfunction:a randomised trial. Lancet 2007;369:2079–87.

67. Solomon SD, Verma A, Desai A, et al. Effect ofintensive versus standard blood pressure loweringon diastolic function in patients with uncontrolledhypertension and diastolic dysfunction. Hyper-tension 2010;55:241–8.

68. Edelmann F, Wachter R, Schmidt AG, et al.Effect of spironolactone on diastolic function andexercise capacity in patients with heart failure withpreserved ejection fraction: the Aldo-DHF ran-domized controlled trial. JAMA 2013;309:781–91.

69. Solomon SD, Zile M, Pieske B, et al. Theangiotensin receptor neprilysin inhibitor LCZ696in heart failure with preserved ejection fraction: aphase 2 double-blind randomised controlled trial.Lancet 2012;380:1387–95.

70. Biering-Sørensen T, Christensen LM,Krieger DW, et al. LA emptying fraction improvesdiagnosis of paroxysmal AF after cryptogenicischemic stroke: results from the SURPRISE study.J Am Coll Cardiol Img 2014;7:962–3.

71. Fatema K, Barnes ME, Bailey KR, et al. Mini-mum vs. maximum left atrial volume for predictionof first atrial fibrillation or flutter in an elderlycohort: a prospective study. Eur J Echocardiogr2009;10:282–6.

72. Santos ABS, Kraigher-Krainer E, Gupta DK,et al. Impaired left atrial function in heart failurewith preserved ejection fraction. Eur J Heart Fail2014;16:1096–103.

73. Zile MR, Gaasch WH, Wiegner AW,Robinson KG, Bing OH. Mechanical determinantsof maximum isotonic lengthening rate in rat leftventricular myocardium. Circ Res 1987;60:815–23.

74. Zile MR, Blaustein AS, Gaasch WH. The effectof acute alterations in left ventricular afterloadand beta-adrenergic tone on indices of early dia-stolic filling rate. Circ Res 1989;65:406–16.

75. Shah AM, Claggett B, Sweitzer NK, et al. Car-diac structure and function and prognosis in heartfailure with preserved ejection fraction: findingsfrom the echocardiographic study of the Treat-ment of Preserved Cardiac Function Heart FailureWith an Aldosterone Antagonist (TOPCAT) trial.Circ Heart Fail 2014;7:740–51.

76. Westwood MA, Wonke B, Maceira AM, et al.Left ventricular diastolic function compared withT2* cardiovascular magnetic resonance for earlydetection of myocardial iron overload in thalas-semia major. J Magn Reson Imaging 2005;22:229–33.

77. Hartiala JJ, Mostbeck GH, Foster E, et al.Velocity-encoded cine MRI in the evaluation of leftventricular diastolic function: measurement ofmitral valve and pulmonary vein flow velocitiesand flow volume across the mitral valve. Am HeartJ 1993;125:1054–66.

78. Paelinck BP, de Roos A, Bax JJ, et al. Feasi-bility of tissue magnetic resonance imaging: a pilotstudy in comparison with tissue Doppler imagingand invasive measurement. J Am Coll Cardiol2005;45:1109–16.

79. Maceira AM, Cosín-Sales J, Roughton M,Prasad SK, Pennell DJ. Reference left atrial

dimensions and volumes by steady state freeprecession cardiovascular magnetic resonance.J Cardiovasc Magn Reson 2010;12:65.

80. Habibi M, Chahal H, Opdahl A, et al. Associa-tion of CMR-measured LA function with heartfailure development: results from the MESA study.J Am Coll Cardiol Img 2014;7:570–9.

81. Heiberg E, Pahlm-Webb U, Agarwal S, et al.Longitudinal strain from velocity encoded cardio-vascular magnetic resonance: a validation study.J Cardiovasc Magn Reson 2013;15:15.

82. Ambale-Venkatesh B, Armstrong AC, Liu CY,et al. Diastolic function assessed from tagged MRIpredicts heart failure and atrial fibrillation over an8-year follow-up period: the Multi-Ethnic Study ofAtherosclerosis. Eur Heart J Cardiovasc Imaging2014;15:442–9.

83. Aurigemma GP, Gottdiener JS, Shemanski L,Gardin J, Kitzman D. Predictive value of systolicand diastolic function for incident congestiveheart failure in the elderly: the CardiovascularHealth Study. J Am Coll Cardiol 2001;37:1042–8.

84. Tsang TS, Barnes ME, Gersh BJ, Bailey KR,Seward JB. Risks for atrial fibrillation andcongestive heart failure in patients $65 years ofage with abnormal left ventricular diastolic relax-ation. Am J Cardiol 2004;93:54–8.

85. Nelson MD, Szczepaniak LS, Wei J, et al. Dia-stolic dysfunction in women with signs and symp-toms of ischemia in the absence of obstructivecoronary artery disease: a hypothesis-generatingstudy. Circ Cardiovasc Imaging 2014;7:510–6.

86. Ugander M, Oki AJ, Hsu LY, et al. Extracellularvolume imaging by magnetic resonance imagingprovides insights into overt and sub-clinicalmyocardial pathology. Eur Heart J 2012;33:1268–78.

87. Moon JC, Messroghli DR, Kellman P, et al.Myocardial T1 mapping and extracellular volumequantification: a Society for Cardiovascular Mag-netic Resonance (SCMR) and CMR Working Groupof the European Society of Cardiology consensusstatement. J Cardiovasc Magn Reson 2013;15:92.

88. Ellims AH, Shaw JA, Stub D, et al. Diffusemyocardial fibrosis evaluated by post-contrast T1mapping correlates with left ventricular stiffness.J Am Coll Cardiol 2014;63:1112–8.

89. Su MY, Lin LY, Tseng YH, et al. CMR-verifieddiffuse myocardial fibrosis is associated with dia-stolic dysfunction in HFpEF. J Am Coll Cardiol Img2014;7:991–7.

90. Mascherbauer J, Marzluf BA, Tufaro C, et al.Cardiac magnetic resonance postcontrast T1 timeis associated with outcome in patients with heartfailure and preserved ejection fraction. Circ Car-diovasc Imaging 2013;6:1056–65.

91. Thavendiranathan P, Verhaert D, Walls MC,et al. Simultaneous right and left heart real-time,free-breathing CMR flow quantification identifiesconstrictive physiology. J Am Coll Cardiol Img2012;5:15–24.

92. Popovi�c ZB, Prasad A, Garcia MJ, et al. Rela-tionship among diastolic intraventricular pressuregradients, relaxation, and preload: impact of ageand fitness. Am J Physiol Heart Circ Physiol 2006;290:H1454–9.

93. Yamakado T, Takagi E, Okubo S, et al. Effectsof aging on left ventricular relaxation in humans.Analysis of left ventricular isovolumic pressuredecay. Circulation 1997;95:917–23.

94. Arbab-Zadeh A, Dijk E, Prasad A, et al. Effectof aging and physical activity on left ventricularcompliance. Circulation 2004;110:1799–805.

95. Palka P, Lange A, Nihoyannopoulos P. Theeffect of long-term training on age-related leftventricular changes by Doppler myocardial veloc-ity gradient. Am J Cardiol 1999;84:1061–7.

96. Nottin S, Nguyen LD, Terbah M, Obert P.Long-term endurance training does not preventthe age-related decrease in left ventricular relax-ation properties. Acta Physiol Scand 2004;181:209–15.

97. Morra EA, Zaniqueli D, Rodrigues SL, et al.Long-term intense resistance training in menis associated with preserved cardiac structure/function, decreased aortic stiffness, and lowercentral augmentation pressure. J Hypertens 2014;32:286–93.

98. Brinker SK, Pandey A, Ayers CR, et al. Asso-ciation of cardiorespiratory fitness with left ven-tricular remodeling and diastolic function: theCooper Center Longitudinal Study. J Am CollCardiol HF 2014;2:238–46.

99. Kitzman DW, Brubaker PH, Morgan TM,Stewart KP, Little WC. Exercise training in olderpatients with heart failure and preserved ejectionfraction: a randomized, controlled, single-blindtrial. Circ Heart Fail 2010;3:659–67.

100. Edelmann F, Gelbrich G, Dungen HD, et al.Exercise training improves exercise capacity anddiastolic function in patients with heart failurewith preserved ejection fraction: results of the Ex-DHF (Exercise Training in Diastolic Heart Failure)pilot study. J Am Coll Cardiol 2011;58:1780–91.

101. Fujimoto N, Prasad A, Hastings JL, et al.Cardiovascular effects of 1 year of progressiveendurance exercise training in patients with heartfailure with preserved ejection fraction. Am HeartJ 2012;164:869–77.

102. Redfield MM, Jacobsen SJ, Burnett JC Jr.,Mahoney DW, Bailey KR, Rodeheffer RJ. Burden ofsystolic and diastolic ventricular dysfunction in thecommunity: appreciating the scope of the heartfailure epidemic. JAMA 2003;289:194–202.

103. Aljaroudi W, Alraies MC, Halley C, et al.Impact of progression of diastolic dysfunction onmortality in patients with normal ejection fraction.Circulation 2012;125:782–8.

104. Halley CM, Houghtaling PL, Khalil MK,Thomas JD, Jaber WA. Mortality rate in patientswith diastolic dysfunction and normal systolicfunction. Arch Intern Med 2011;171:1082–7.

105. Abhayaratna WP, Marwick TH, Smith WT,Becker NG. Characteristics of left ventricular dia-stolic dysfunction in the community: an echocar-diographic survey. Heart 2006;92:1259–64.

106. Lam CS, Lyass A, Kraigher-Krainer E, et al.Cardiac dysfunction and noncardiac dysfunction asprecursors of heart failure with reduced and pre-served ejection fraction in the community. Circu-lation 2011;124:24–30.









































































































































































































1093

107. Mureddu GF, Agabiti N, Rizzello V, et al.Prevalence of preclinical and clinical heart failurein the elderly. A population-based study in centralItaly. Eur J Heart Fail 2012;14:718–29.

108. Vogel MW, Slusser JP, Hodge DO, Chen HH.The natural history of preclinical diastolicdysfunction: a population-based study. Circ HeartFail 2012;5:144–51.

109. From AM, Scott CG, Chen HH. The develop-ment of heart failure in patients with diabetesmellitus and pre-clinical diastolic dysfunction apopulation-based study. J Am Coll Cardiol 2010;55:300–5.

110. Greenberg B. Pre-clinical diastolic dysfunc-tion in diabetic patients: where do we go fromhere? J Am Coll Cardiol 2010;55:306–8.

111. Kuwaki H, Takeuchi M, Chien-Chia Wu V,et al. Redefining diastolic dysfunction grading:

combination of E/A #0.75 and deceleration time>140 ms and E/ε0 $10. J Am Coll Cardiol Img2014;7:749–58.

112. Klein AL, Hatle LK, Taliercio CP, et al. SerialDoppler echocardiographic follow-up of left ven-tricular diastolic function in cardiac amyloidosis.J Am Coll Cardiol 1990;16:1135–41.

113. Senni M, Paulus WJ, Gavazzi A. New strate-gies for heart failure with preserved ejectionfraction: the importance of targeted therapies forheart failure phenotypes. Eur Heart J 2014;35:2797–815.

114. Dalen H, Thorstensen A, Vatten LJ, Aase SA,Stoylen A. Reference values and distribution ofconventional echocardiographic Doppler measuresand longitudinal tissue Doppler velocities in apopulation free from cardiovascular disease. CircCardiovasc Imaging 2010;3:614–22.

115. Smiseth OA, Remme EW, Opdahl A, Aakhus S,Skulstad H. Heart failure with normal left ventric-ular ejection fraction: basic principles and clinicaldiagnostics. In: Bartunek J, Vanderheyden M, edi-tors. Translational Approach to Heart Failure. NewYork, NY: Springer, 2013:25–62.

KEY WORDS cardiac magnetic resonance,computed tomography, diastolicdysfunction, echocardiography, heart failurewith preserved ejection fraction (HFpEF),left ventricular function

Go to http://www.acc.org/jacc-journals-cme to take theCME quiz for this article.
















































Documents

Cardiac Imaging to Evaluate Left Ventricular Diastolic ... · pressure and therefore a reduction in the early dia-stolic transmitral pressure gradient and the early transmitralvelocity(E)andacompensatoryincreasein