Early diastolic strain rate in relation to systolic and ... - CiteSeerX

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CLINICAL RESEARCHImaging

Early diastolic strain rate in relation to systolicand diastolic function and prognosis in acutemyocardial infarction: a two-dimensionalspeckle-tracking studyMads Ersbøll1*, Mads J. Andersen1, Nana Valeur2, Ulrik M. Mogensen1, Yama Fahkri1,Jens J. Thune1, Jacob E. Møller1, Christian Hassager1, Peter Søgaard3, and Lars Køber1

1The Heart Centre, Department of Cardiology, University Hospital Rigshospitalet, Copenhagen, Denmark; 2Department of Cardiology, University Hospital Herlev, Herlev, Denmark;and 3Department of Cardiology, Aalborg University, Aalborg, Denmark

Received 8 January 2013; revised 16 April 2013; accepted 30 April 2013; online publish-ahead-of-print 26 May 2013

See page 605 for the editorial comment on this article (doi:10.1093/eurheartj/eht217)

Aims Diastolic dysfunction in acute myocardial infarction (MI) is associated with adverse outcome. Recently, the ratio of earlymitral inflow velocity (E) to global diastolic strain rate (e′sr) has been proposed as a marker of elevated LV filling pressure.However, the prognostic value of this measure has not been demonstrated in a large-scale setting when existing para-meters of diastolic function are known. We hypothesized that the E/e′sr ratio would be independently associatedwith an adverse outcome in patients with MI.

Methodsand results

We prospectively included patients with MI and performed echocardiography with comprehensive diastolic evaluationincluding E/e′sr. The relationship between E/e′sr and the primary composite endpoint (all-cause mortality, hospitalizationfor heart failure (HF), stroke, and new onset atrial fibrillation) was analysed with Cox models. A total of 1048 patients(mean age 63+ 12, 73% male) were included and 142 patients (13.5%) reached the primary endpoint (medianfollow-up 29 months). A significant prognostic value was found for E/e′sr [hazard ratio (HR) per 1 unit change: 2.36,95% confidence interval (CI): 2.02–2.75, P , 0.0001]. After multivariable adjustment E/e′sr remained independentlyrelated to the combined endpoint (HR per 1 unit change, 1.50; CI: 1.05–2.13, P ¼ 0.02). The prognostic value ofE/e′sr was driven by mortality (HR per 1 unit change, 2.52; CI: 2.09–3.04, P , 0.0001) and HF admissions (HR per 1unit change, 2.79; CI: 2.23–3.48, P , 0.0001).

Conclusion Deformation-based E/e′sr contributes important information about global myocardial relaxation superior to velocity-based analysis and is independently associated with the outcome in acute MI.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Keywords Acute myocardial infarction † Echocardiography † Diastolic dysfunction † Strain echocardiography

IntroductionImpaired myocardial relaxation assessed with Doppler echocardiog-raphy in the aftermath of acute myocardial infarction (MI) is an im-portant predictor of major adverse events.1 –5 Recently, themeasurement of the regional diastolic strain rate has been proposedas a novel marker of elevated LV filling pressure.6– 8 The advantage ofthis marker is potentially that the regional early velocity of diastolicdeformation (strain rate) more accurately reflects the diastolic

performance of all myocardial segments. Furthermore, the ratio ofearly mitral inflow velocity to early diastolic strain rate of all the myo-cardial segments (E/e′sr) can be obtained with two-dimensionalspeckle tracking avoiding the limitations and angle dependency oftissue Doppler imaging.

The early active relaxation, the rate of recoil and the effectivechambercompliance may be affected by myocardial ischaemia and in-farction through different mechanisms. Active relaxation being anenergy-dependent process is affected immediately with the onset

* Corresponding author. Tel: +45 35459641, Fax: +45 35457705, Email: [email protected]

Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2013. For permissions please email: [email protected]

European Heart Journal (2014) 35, 648–656doi:10.1093/eurheartj/eht179

by guest on March 6, 2016

http://eurheartj.oxfordjournals.org/D

ownloaded from

of myocyte ischaemia similarly to systolic contractile function.Impaired myocyte contraction reduces the elastic energy stored intitin proteins, leading to reduced early diastolic untwisting and ven-tricular suction. Chamber compliance is also negatively affected inMI through peri-infarct interstitial oedema and in the later phase fi-brotic changes in the myocardium.9 While the extent of the infarctedmyocardium in itself will be related to the magnitude of diastolic im-pairment, pre-existing co-morbid conditions, such as hypertensionand diabetes, may adversely affect myocardial relaxation throughcompensatory myocardial hypertrophy and triglyceride depos-ition.10,11 All of these factors would be anticipated to affect E/e′srwhich may be a sensitive marker of LV diastolic function post-MI.

To the best of our knowledge, no large-scale studies have exam-ined the prognostic importance of E/e′sr in addition to existing echo-cardiographic indices of diastolic dysfunction in patients with MI.Accordingly, in this study we hypothesized that in patients with MIE/e′sr would be independently related to an adverse outcomewhenexistingconventional measuresof diastolic functionareknown.

Methods

Study design and patient populationWe conducted a prospective study of patients referred for invasive cor-onary angiography (CAG) due to either STEMI or non-STEMI at two ter-tiary cardiac centres in the Copenhagen region. All the patients providedwritten informed consent prior to transthoracic echocardiographicexamination.12 Exclusion criteria were age ,18 years, non-cardiacdisease with a life expectancy ,1 year, or inability to provide writteninformed consent. Furthermore, echocardiograms obtained in patientswith atrial fibrillation (AF), paced rhythm, severe aortic stenosis, orsevere mitral regurgitation (MR) were excluded from the analyses.

On the basis of hospital records, admission information on diabetesmellitus, hypertension, a history of ischaemic heart disease, and priorMI was registered. Findings in relation to CAG including culprit lesion,three-vessel disease (3VD), left main involvement (LM), and type ofrevascularization [percutaneous coronary intervention (PCI), coronaryartery bypass grafting (CABG), or no intervention] were registered.Objective signs of heart failure (HF) at presentation or during hospitaliza-tion were scoredaccording to the Killip classification scheme class I– IV,13

and episodes of AF recorded. Additional biochemical work-up includedcreatinine and peak troponin during the hospital stay. Estimated glomeru-lar filtration rate (eGFR) was measured from the four-variable MDRDformula.14 Information on a left or right bundle branch block (LBBBand RBBB) was obtained from 12-lead electrocardiograms at discharge.The study was approved by the Regional Scientific Ethics Committee(reference number H-D-2009-063).

EchocardiographyEchocardiography wasperformed within 48 hof admission to the tertiarycentre. Patient presenting with STEMI were examined after acute CAGand patients with NSTEMI were examined on admission at our institutionprior to CAG. Echocardiographic cine loops were obtained by recordingthree consecutive heart cycles. All examinations were performed on aVivid e9 (General Electric, Horten, Norway). Images were obtained ata frame rate of 60–90 fps and digitally transferred to a remote worksta-tion for offline analysis (Echopac BT 11.1.0, General Electric, Horten,Norway). All the analyses were performed by a single experienced oper-ator (M.E.) blinded to clinical characteristics and outcome.

Two-dimensional parasternal images were used to determine LVdimensions and wall thickness. The left atrial maximum volume index(LA max) was determined from the biplane area length method, andleft ventricular ejection fraction (LVEF) was determined using thebiplane Simpson model. The LV mass was calculated from the LV lineardimensions in the parasternal view. Volumetric and dimensional mea-surements of the LV and left atrium were indexed to the body surfacearea when appropriate. All volumetric analyses were performed in ac-cordance with EAE/ASE recommendations.15

Doppler recordings of mitral inflow were performed by placing a2.5 mm sample volume at the tip of the mitral valve (MV) leafletsduring diastole. Peak velocity of early (E) and atrial (A) diastolic fillingand MV deceleration time (DT) were measured, and the E/A-ratio calcu-lated. Pulsed wave TDI recordings were performed at the lateral andmedial mitral annulus using a 2.5-mm sample volume with the measure-ments of myocardial peak early (e′). The mean E/e′ ratio was calculatedfrom an average of lateral and medial values of e′. Mitral regurgitationwas quantified by calculating the effective regurgitant orifice (ERO)using the proximal isovelocity surface area method. An ERO,0.20 cm2 was considered mild, 0.20–0.40 cm2 moderate, and.0.40 cm2 severe MR.

Speckle-tracking deformation analysisTwo-dimensional speckle tracking was performed in the three apicalviews by first manually tracing the endocardium at the onset of systoleafter which the software tracked the myocardial speckle patternframe-by-frame. The region of interest was adjusted to cover the thick-nessof themyocardium andadequate trackingwasverified andcorrectedif necessary. The aortic valve closure was identified on continuous waveDoppler recording through the aortic valve. The LV was subsequentlydivided by the software into 17 segments covering the entire myocar-dium. Global longitudinal strain (GLS) was calculated for each of thethree apical projections and overall GLS as the average value of allthree projections. Global systolic strain rate (SRs) and global earlydiastolic SR were calculated from the average of all 17 segments. TheE/e′sr ratio was calculated as the E velocity (m/s) divided by the globale′sr value (s21) (Figure 1).

Follow-up and endpoint definitionThe primary outcome was a composite of death from any cause, hospi-talization for HF, new onset AF, or stroke whichever came first andwhich have all been linked to LV dysfunction after MI.3,4,16,17 Informationon all-cause mortality was obtained from the Danish Civil RegistrationSystem. Secondary endpoints were death from any cause, hospitalizationfor HF, new onset AF, and stroke. Information on HF hospitalizations,new onset AF, and stroke was obtained from a systematic review of allhospital admissions after the index MI. Hospitalization from HF wasdefined as admission due to dyspnoea with chest radiographic evidenceof pulmonary congestion and treatment with i.v. diuretics. New onsetAF was defined as first-time admission to hospital due to AF withoutprior knowledge of AF. Stroke was defined as hospitalization with aneurological deficit of cerebrovascular cause that persisted beyond24 h.18 Verification of endpoints was performed by two independentreviewers blinded to echocardiographic information relating to theindex MI, and in case of disagreement a third reviewer would classifythe outcome.

Statistical analysisAll data are reported as mean+ SD or median (first- and third-quartile,Q1–Q3). Baseline clinical and echocardiographic data were analysedaccording to E/e′sr quartiles by the Cochran-Armitage trend test for

Early diastolic strain rate in relation to systolic and diastolic function 649

by guest on March 6, 2016

http://eurheartj.oxfordjournals.org/D

ownloaded from

categorical variables and ANOVA for continuous variables, with all testsbeing two-sided, and statistical significance was defined as P , 0.05. Thelinear relationship between echocardiographic measures was tested withcorrelation analysis using Pearson correlation coefficient. Missing valueswere imputed using multiple imputations five times with predictive meanmatching for the continuous covariates with the assumption of data beingmissing at random (R package mice). Statistical analyses were performedon all of the imputed data sets and results pooled using the methoddescribed by Rubin.19 All the analyses were verified with completecase analyses without imputations utilizing only the observations withcomplete data for all covariates.20

The optimal cut-off value of E/e′sr was found by maximizing the partiallikelihood in a univariable Cox model with the hazard ratio (HR) givenwith 95% confidence interval (CI). The Kaplan–Meier estimate of event-free survival was then calculated and plotted according to this cut-off.Multivariable Cox proportional hazards regression analyses were per-formed to analyse the independent and incremental value of E/e′sr inthe following manner: First, a model was fitted with clinical informationincluding age, diabetes, hypertension, Killip class . 1, bundle branchblock, LAD involvement, 3VD and/or LM involvement, CABG treatment,and eGFR; secondly, we added to this model LVEF, LA max, E/e′, MV DT,140 ms, GLS, SRs, and MR grade II, and finally, E/e′sr was added. Directcomparison between E/e′ and E/e′sr was performed in both univariableand multivariable models. Incremental model performance was assessedby change in the Waldx2 value. Noviolations of proportionalityor linear-ity were found in the overall model when analysing the Martingale- andSchoenfeld residuals.

The relative importance of diastolic function indices was assessed in amodel with quartiles of E/e′sr and clinically validated cut-off values of E/e′,LA max, and MV DT. The relationship between E/e′sr and the secondaryendpoints was analysed in separate uni- and multivariable Cox modelswith backwards elimination to final parsimonious models due to thelownumberof individual events. Inter- and intra-observer reproducibilitywas assessed in 20 randomly selected patients with the calculation of themean difference and 95% limits of agreement (+1.96 SD). The 20patients originated from both centres and for inter-observer studieswere analysed independently by a reader from both centres. All analyses

were performed using R version 2.15.0 (R Development Core Team2011, http://www.R-project.org/. library: Survival, Hmisc, Publish, mice).

Results

Baseline characteristicsA total of 1110 patients with MI were prospectively included,56 patients were excluded due to AF (n ¼ 40), ventricularpaced rhythm (n ¼ 8), severe aortic stenosis (n ¼ 8), and severe MR(n ¼ 6). Out of the 1048 patients remaining (mean age 63 years,73% male), missing values were encountered for E/e′sr [n ¼ 58(5.5%)], GLS [n ¼ 50 (4.7%)], E/e′ [n ¼ 12 (1%)], LVEF [n ¼ 8 (1%)],LA max [n ¼ 13 (1%)], and MV DT [n ¼ 10 (1%)] for a total of 62(5.9%) patients with one or more missing values. The reasons for themissing data were poor image quality (n ¼ 50), missing spectralDoppler from the MV inflow (n ¼ 10), and poor ECG traces preclud-ing strain calculations (n ¼ 2). The missing values were imputed and allsubsequent analyses were pooled estimates on the imputed data sets.

The baseline clinical characteristics of the patients are givenaccording to quartiles of E/e′sr (Table 1). Elevated E/e′sr was asso-ciated with older age and an increasing prevalence of hypertension,previous MI, diabetes, and reduced renal function. Increasing E/e′srwas associated with a higher prevalence of multivessel disease,lower prevalence of primary PCI, and higher likelihood of subsequentrevascularization by CABG. However, there was no difference in theproportion of medically managed patients without mechanical revas-cularization. Intra- and inter-observer mean difference and 95% limitsof agreement for E/e′sr were 20.07+0.19 and 0.04+ 0.25, re-spectively.

The median value of E/e′sr was 0.91 (Q1–Q3, 0.73–1.19), and ele-vated E/e′sr was significantly associated with increasing LV dilatation,reduced LVEF, and increased LVMi. Furthermore, E/e′sr was asso-ciated with worsening diastolic function, reduced RV function,

Figure 1 Representative example of the measurement of early diastolic strain rate from two-dimensional speckle tracking (MVO, mitral valveopening).

M. Ersbøll et al.650

by guest on March 6, 2016

http://eurheartj.oxfordjournals.org/D

ownloaded from

increasing prevalence of MR grade II, and both GLS and systolic SR(Table 2). There was a significant correlation between E/e′sr and E/e′ (r ¼ 0.74, P , 0.0001), LA max (r ¼ 0.26, P , 0.0001), MV DT(r ¼ 20.13, P , 0.0001), and LVEF (r ¼ 20.46, P , 0.0001).

Relationship between E/e′sr and outcomeDuring a median follow-up of 29 months (Q1–Q3, 22.0–32.6), atotal of 142 patients reached the combined endpoint (13.5%) ofwhich 68 patients died (47.8%), 39 patients were hospitalized forHF (27.5%), 15 patients were admitted with stroke (10.6%), and20 were hospitalized with new onset AF (14.1%). The individualsecondary endpoints were distributed as follows: 80 patients died(7.6%), 42 patients were hospitalized for HF (4.0%), 16 patients

were admitted with stroke (1.5%), and 20 were hospitalized withnew onset AF (1.9%). There was no difference in event-free survivalduring the follow-up according to earlier (late 2009–2010) vs. laterdate (2010–early 2012) of inclusion into the study population(x2 ¼ 2.3, P ¼ ns). In univariable analysis, E/e′sr was significantlyrelated to the primary endpoint (HR: 2.36; CI: 2.02–2.75, P ,

0.0001) and outperformed the model with E/e′ in direct compari-son (x2 ¼ 124.8 vs. x2 ¼ 70.9, P , 0.0001). With the optimalcut-off value of E/e′sr .1.25, a total of 214 patients (20.4%) wereidentified of which 68 (31.7%) reached the primary endpoint atend of the follow-up (HR: 4.38; CI: 3.05–5.93, P , 0.0001, x2 ¼

75.1) (Figure 2). A total of 143 patients (13.6%) had E/e′ .15of which 42 (29.4%) reached the primary composite endpoint

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 1 Baseline clinical characteristics according to quartiles of E/e′sr

Characteristic E/e′sr

<0.73 (n 5 262) 0.73–0.91 (n 5 262) 0.91–1.20 (n 5 262) >1.20 (n 5 262) P-value

Age, years 57.3+11.5 59.9+11.6 64.0+11.0 69.0+10.5 ,0.0001

Male sex 197 (75.2) 205 (78.2) 185 (70.6) 179 (68.3) 0.0465

BMI, kg/m2 26.1+4.1 26.8+4.4 27.4+4.4 27.0+4.6 0.0042

Medical history

Hypertension 85 (32.4) 101 (38.5) 135 (51.5) 169 (64.5) ,0.0001

Previous MI 31 (11.8) 25 (9.5) 30 (11.5) 46 (17.6) 0.0363

Diabetes 16 (6.1) 28 (10.7) 34 (13.0) 65 (24.8) ,0.0001

Smoking 196 (74.8) 184 (70.2) 173 (66.0) 175 (66.8) 0.1147

Pre MI medical treatment

ACEI/ARB 40 (15.3) 57 (21.8) 62 (23.7) 102 (38.9) ,0.0001

Beta-blocker 25 (9.5) 24 (9.2) 27 (10.3) 45 (17.2) 0.0116

ASA/clopidogrel 36 (13.7) 33 (12.6) 38 (14.5) 53 (20.2) 0.0712

eGFR, mL/min/1.73 m2 93.6+26.0 91.6+26.4 90.7+27.4 79.8+27.4 ,0.0001

Killip class . 1 9 (3.4) 22 (8.4) 36 (13.7) 88 (33.6) ,0.0001

Heart rate, b.p.m. 70.0+12.2 71.6+12.3 73.5+13.1 76.1+13.9 ,0.0001

RBBB 9 (3.4) 6 (2.3) 10 (3.8) 19 (7.3) 0.0141

LBBB 0 (0.0) 4 (1.5) 6 (2.3) 10 (3.8) 0.0304

Blood pressure, mmHg

Systolic 128.8+17.7 128.2+20.0 131.0+20.9 132.0+24.5 0.3000

Diastolic 79.0+11.4 79.3+12.0 80.4+11.8 79.5+14.1 0.7614

Infarct classification

non-STEMI 80 (30.5) 75 (28.6) 78 (29.8) 104 (39.7) 0.0253

STEMI 182 (69.5) 187 (71.4) 184 (70.2) 158 (60.3) 0.0253

Troponin T (mg/L) 2.8+3.9 3.67+4.24 4.41+5.03 4.68+6.09 0.0030

Troponin I (mg/L) 83.4+151.8 110.4+161.4 129.1+171.5 144.8+210.3 0.2508

Angiographic findings

LAD involvement 76 (29.1) 92 (35.1) 114 (43.5) 141 (53.8) ,0.0001

Three-vessel disease or LM 21 (8.0) 30 (11.5) 44 (16.8) 75 (28.6) ,0.0001

Intervention

No PCI 55 (21.0) 47 (17.9) 52 (19.8) 76 (29.0) 0.0032

PCI 41 (15.6) 38 (14.5) 48 (18.3) 49 (18.7) 0.5032

Primary PCI 166 (63.4) 177 (67.6) 162 (61.8) 137 (52.3) 0.0123

Additional CABG 12 (4.6) 15 (5.7) 26 (9.9) 39 (14.9) ,0.0001

Early diastolic strain rate in relation to systolic and diastolic function 651

by guest on March 6, 2016

http://eurheartj.oxfordjournals.org/D

ownloaded from

(HR: 3.15; CI: 2.20–4.52, P , 0.0001, x2 ¼ 39.0). With completecase analysis (n ¼ 988), a total of 134 events were modelled andthe HR for continuous E/e′sr remained unchanged (HR: 2.38;CI: 2.05–2.77, P , 0.0001).

In the multivariable Cox model, E/e′sr provided independent andincremental prognostic information in relation to the primary end-point when added to a model including age, diabetes, hypertension,

Killip class.1, bundle branch block, LAD involvement, 3VD and/orLM involvement, CABG treatment, eGFR, troponin, LVEF, LA max,E/e′, MV DT ,140 ms, GLS, SRs, and MR grade II (adjusted HR:1.50; CI: 1.05–2.13, P ¼ 0.02) (Table 3). In direct comparison,the full multivariable model without E/e′ or E/e′sr (x2 ¼ 169.2) wasnon-significantly improved by adding E/e′ (x2 ¼ 170.7, P ¼ 0.27),whereas the addition of E/e′sr instead was a significant improvement(x2 ¼ 181.4, P ¼ 0.01). With complete case analysis E/e′sr remainedsignificant in the model (adjusted HR: 1.53; CI: 1.07–2.17, P ¼ 0.02).Comparing the added value of E/e′ vs. E/e′sr to a model consisting ofclinical and echocardiographic variables widely used in clinical prac-tice, resulted in significant improvement with E/e′sr as opposed toE/e′ (Figure 3).

There was a highly significant prognostic value of E/e′sr in relationto the secondary outcomes mortality (HR: 2.52; CI: 2.09–3.04, P ,

0.0001) and HF admission (HR: 2.79; CI: 2.23–3.48, P , 0.0001),whereas the prognostic utility was borderline non-significant in rela-tion to stroke (HR: 7.74; CI: 0.97–3.11, P ¼ 0.06) and new onset AF(HR: 1.55; CI: 0.85–2.83, P ¼ 0.15). After multivariable adjustmentand backwards elimination, E/e′sr maintained an independent valuein relation to mortality (HR: 1.65; CI: 1.24–2.29, P ¼ 0.0006),whereas relationship with HF admissions (HR: 1.41; CI: 0.95–2.09,P ¼ 0.090), stroke (HR: 0.89; CI: 0.37–2.12, P ¼ 0.79), and newonset AF (HR: 1.06; CI: 0.45–2.49, P ¼ 0.90) was attenuated.

Prognostic value of E/e′sr in relationto diastolic dysfunction indicesThe relationship between outcome and quartiles of E/e′sr adjustedfor clinically validated cut-off values of E/e′ (,8; 8–15; .15), LAmax (,34, 34–40, and .40 mL/m2), and MV DT (,140, 140–240, and .240 ms) are given in Figure 4. In this model, the third-

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 2 Baseline echocardiographic characteristics according to quartiles of E/e′sr

Characteristic E/e′sr

<0.73 (n 5 262) 0.73–0.91 (n 5 262) 0.91–1.20 (n 5 262) >1.20 (n 5 262) P-value

LV systolic function

LV EDV, mL 82.2+21.5 84.9+23.5 86.1+27.5 99.3+36.6 ,0.0001

LV ESV, mL 37.1+13.2 40.6+14.6 43.8+19.4 58.7+31.3 ,0.0001

LVEF, % 55.4+8.1 52.8+8.6 50.4+9.6 43.4+12.2 ,0.0001

LVMI, g/m2 84.5+19.6 86.2+21.2 92.0+25.3 103.6+31.0 ,0.0001

LV diastolic function

LA max (mL/m2) 32.5+9.7 32.3+9.7 33.1+10.4 39.4+11.9 ,0.0001

E/e′ ratio 7.7+2.4 9.1+2.2 11.0+2.8 15.4+5.7 ,0.0001

E/A ratio 1.04+0.35 1.01+0.30 0.99+0.41 1.19+0.69 ,0.0001

MV deceleration time (ms) 193.5+50.7 188.4+48.6 189.2+47.8 182.1+61.8 0.1020

TAPSE (cm) 2.3+0.4 2.3+0.5 2.1+0.4 2.0+0.5 ,0.0001

Deformation parameters

GLS (%) 216.3+2.8 214.7+2.8 212.9+2.8 210.6+3.2 ,0.0001

Systolic SR 20.89+0.16 20.80+0.15 20.70+0.16 20.57+0.17 ,0.0001

Early diastolic SR 1.05+0.28 0.87+0.20 0.71+0.18 0.53+0.17 ,0.0001

MR grade II– III 1 (0.4) 4 (1.5) 9 (3.4) 15 (5.7) 0.0011

Figure 2 Kaplan–Meier estimate of event-free survival(n ¼ 1054) stratified by the optimal cut-off value of E/e′sr .1.25.

M. Ersbøll et al.652

by guest on March 6, 2016

http://eurheartj.oxfordjournals.org/D

ownloaded from

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 3 Hazard ratio (95% CI) for the composite endpoint death, admission for heart failure, stroke, and new onset atrial fibrillation in univariable and multivariableCox models with sequential addition of covariates

Covariate Univariable Model 1 Model 2 Model 3

HR 95% CI P-value HR 95% CI P-value HR 95% CI P-value HR 95% CI P-value

Age 1.07 1.05–1.08 ,0.0001 1.05 1.03–1.07 ,0.0001 1.05 1.03–1.07 ,0.0001 1.05 1.04–1.07 ,0.0001

Diabetes 2.20 1.50–3.22 0.0001 1.70 1.14–2.55 0.0099 1.60 1.06–2.43 0.0261 1.59 1.04–2.42 0.0309

Hypertension 1.82 1.30–2.55 0.0005 1.15 0.81–1.64 0.4324 1.11 0.78–1.59 0.5494 1.10 0.77–1.57 0.6026

Killip class.1 3.35 2.36–4.75 ,0.0001 2.08 1.43–3.01 0.0001 1.30 0.87–1.95 0.1980 1.31 0.88–1.97 0.1850

LAD involvement 1.12 0.88–1.41 0.3617 1.02 0.80–1.29 0.8982 0.91 0.70–1.18 0.4687 0.90 0.69–1.16 0.4129

Three-vessel disease /LM 2.07 1.42–3.00 0.0001 1.30 0.84–2.02 0.2380 1.03 0.64–1.64 0.9148 0.98 0.61–1.57 0.9235

CABG 1.36 0.96–1.92 0.0874 1.00 0.67–1.49 0.9935 0.90 0.59–1.37 0.6284 0.93 0.61–1.42 0.7388

eGFR (10 mL/m2 decrease) 1.20 1.13–1.29 ,0.0001 1.07 1.01–1.15 0.0335 1.06 1.00–1.13 0.0687 1.05 0.98–1.12 0.1525

Troponin 1.18 1.02–1.38 0.0290 1.16 0.99–1.36 0.0599 1.03 0.87–1.21 0.7489 1.04 0.88–1.22 0.6332

RBBB 2.73 1.54–4.83 0.0006 1.19 0.65–2.18 0.5793 0.96 0.5–1.85 0.9059 1.05 0.54–2.02 0.8845

LBBB 3.85 1.88–7.87 0.0002 3.13 1.49–6.56 0.0026 2.25 1.04–4.89 0.0393 2.27 1.05–4.9 0.0372

LVEF 0.94 0.93–0.96 ,0.0001 0.97 0.95–0.99 0.0128 0.98 0.95–1.00 0.0251

LA max 1.03 1.02–1.05 ,0.0001 1.01 1.00–1.02 0.1303 1.01 1.00–1.02 0.1591

E/e′ 1.10 1.08–1.13 ,0.0001 1.02 0.98–1.06 0.2715 0.99 0.95–1.04 0.6794

MV deceleration time ,140 ms 2.30 1.61–3.28 ,0.0001 1.19 0.78–1.82 0.4100 1.18 0.77–1.81 0.4529

SRs (100/ms) 1.03 1.02–1.04 ,0.0001 0.99 0.97–1.01 0.1746 0.99 0.97–1.01 0.1938

GLS 1.23 1.18–1.29 ,0.0001 1.13 1.01–1.26 0.0272 1.10 0.99–1.23 0.0882

MR grade II 3.21 1.69–6.10 0.0004 1.17 0.56–2.44 0.6759 1.19 0.57–2.48 0.6375

E/e′sr 2.36 2.02–2.75 ,0.0001 1.49 1.05–2.12 0.0266

Earlydiastolic

strainrate

inrelation

tosystolic

anddiastolic

function653

by guest on March 6, 2016 http://eurheartj.oxfordjournals.org/ Downloaded from

and fourth-quartile of E/e′sr remained significant along with LA max.40 mL/m2 and MV DT ,140 ms, while none of the E/e′ groups wassignificant. A majority of patients (n ¼ 626, 59.7%) had E/e′ values in

the indeterminate rangeof 8–15. Adding E/e′srmeasurements to thisgroup and stratifying according to the cut-off of 1.25 resulted in a sig-nificant separation of the event-free survival curves (P , 0.0001)(Figure 5).

DiscussionIn the so far largest prospective study of 1054 patients with MI under-going comprehensive diastolic evaluation, we demonstrated that (i)E/e′sr was independently associated with the composite endpointwhen adjusting for both diastolic dysfunction indices and systolicdeformation-based measures; (ii) E/e′sr outperformed conventionalvelocity-based E/e′ both in univariable and multivariable models.Taken together, these findings suggest that deformation-based E/e′sr contributes important information about global myocardial re-laxation properties superior to velocity-based analysis, and inde-pendently of indices of chronically elevated filling pressure andrestrictive filling pattern in patients with MI.

Prognostic value of E/e′sr comparedwith E/e′

Annular myocardial velocities assessed by TDI may not fully captureearly regional or global myocardial relaxation due to several factors.These include well-known limitations of Doppler-based methodssuch as the angle dependency with the potential for significanterrors with angulations .208. Regional wall motion abnormalitiesin the sampling region may lead to low annular velocities on tissueDoppler despite overall near normal LV relaxation. Furthermore,early diastolic relaxation is an active energy-dependent processthat is rapidly initiated in the basal segments of the LV and propagatestowards the apex. This base to the apex gradient of SR creates a waveof relaxation that, together with untwisting and myocardial thinningdue to its incompressible nature, results in chamber dilatation.21,22

Velocity-based assessment of early relaxation using e′ obtained inthe basal medial and lateral annulus may not accurately reflect region-al disturbances in relaxation.

Early diastolic SR is impaired in at-risk segments despite nearnormal systolic deformation and can differentiate between remoteuninjured segments and at-risk segments in animal models of acuteMI.23 Thus, our findings can be interpreted in the context thatdeformation-based diastolic evaluation apart from better reflectingall myocardial regions compared with annular velocity-based assess-ment, potentially also confer better discriminative information on theextent of myocardial ischaemic injury. Accordingly, indexing MVinflow velocity to the e′sr should yield more information on globalmyocardial relaxation properties than annular velocity-based mea-sures and suggest that E/e′sr may represent a sensitive and morephysiological approach of integrating haemodynamics and myocar-dial relaxation properties in patients with acute MI. Furthermore,the novel finding in this study of independence of and indeed strongerprognostic value of E/e′sr compared with GLS and SRs, which havepreviously been identified as significant prognostic markers in acuteMI24,25 and closely related with the infarct size,26 highlights the im-portance of LV diastolic impairment in acute MI. We found a lowerproportion of patients with E/e .15 (13.6 vs. 29.2%) comparedwith the study by Hillis et al.2 However, substantial advances in

Figure 3 Incremental model performance assessed by increasein Wald x2 by adding either E/e′ (red, model 2) or E/e′sr (green,model 3) to a clinical + echo model (blue, model 1) consisting ofage, hypertension, diabetes, Killip class .1, multivessel disease, esti-mated glomerular filtration rate, troponin, left ventricular ejectionfraction, left atrial maximum volume index, and mitral valve deceler-ation time ,140 ms.

Figure4 Kaplan–Meierestimateof event-free survival accordingto E/e′ ,8 (blue), E/e′ . 15 (green), and E/e′ between 8 and 15stratified according to E/e′sr .1.25 (red) and E/e′sr , 1.25 (black).

M. Ersbøll et al.654

by guest on March 6, 2016

http://eurheartj.oxfordjournals.org/D

ownloaded from

treatment have occurred between that study and ours which likelyexplains the differing prevalence of E/e′ .15.13,27 We were able todemonstrate that E/e′sr was independent of severely dilated LAand restrictive filling pattern, whereas the study by Hillis et al.2 didnot report LA volumes. Our findings expand on that study in thatwe demonstrate the incremental importance of an early assessmentof global myocardial relaxation properties using E/e′sr as a superioralternative to conventional E/e′. Evaluating diastolic function at asingle time point in the early phase after acute MI is subjected tothe risk that filling pressure and loading conditions may changeduring the peri-infarct period. A recent study suggested, that E/e′

obtained 2 weeks after the MI was the most important predictor ofHF and cardiac death.28 However, the authors did not provide adirect comparison with E/e′ obtained in the early phase after MI,which was reported by Hillis et al.2 Further studies are needed toassess whether and to which extent E/e′sr changes during thecourse of the MI.

E/e′sr as a measure of LV filling pressureA number of studies have demonstrated a closer correlationbetween invasively measured LV filling pressure and deformation-based E/e′ compared with E/e′ based on annular velocities in mixedpopulations referred forclinically indicated invasive LV pressure mea-surements.6– 8 Kimura et al. found a significantly better correlationbetween LV filling pressure and E/e′sr7 in patients with decompen-sated HF with an indication for invasive haemodynamic monitoring.

Dokainish et al.6 found similar correlation between LV filling pressureand E/e′sr as well as E/(Global diastolic strain) (E/Ds), whereas the re-lationship with conventional E/e′ was weaker. A study by Wang et al.using both animal experiments and patient studies found that E/e′srduring the isovolumic relaxation period (E/e′sr-ivr) was a bettermeasure of LV end-diastolic pressure8 than E/e′sr. In contrast astudy enrolling patients with HF and preserved LVEF and controlsfound no incremental benefit with E/e′sr or E/e′sr-ivr for the predic-tion of LV filling pressure.29 The results of these studies underscorethat no single echocardiographic measure may be appropriate in allthe patients for predicting LV filling pressure and rather an integrativeapproach of several measures should be pursued.30 A recent study in288 patients with acute MI found that E/e′sr-ivr contained prognosticvalue above E/e′, while E/e′sr was not prognostic which is in contrastto our findings.31 The study was smaller than ours; the outcome dif-fered in respect to composition and neither LA max, MV DT nor sys-tolic deformation measures were reported.

Interestingly, the limited correlation found between E/e′sr andfilling pressure in these studies suggests that the prognostic value ofE/e′sr may not be entirely driven by a strong relationship with LVfilling pressure at rest. However, a recent study found that patientswith MI and diastolic dysfunction, while having normal LV filling pres-sureat rest, exhibited markedlyelevated LV filling pressureduring ex-ercise compared with controls.32 We found that in the large group ofpatients with E/e′ in the indeterminate range of 8–15 addition of E/e′sr .1.25 led to the identification of a smaller subset with the

Figure 5 The forest plot of the multivariable Cox model consisting of E/e′sr, E/e′, LA max, and mitral valve deceleration time by clinical cut-offvalues. Reference levels are within each group the following: E/e′sr lowest quartile, E/e , 8, LA max ,34 mL/m2, and mitral valve decelerationtime between 140 and 240 ms (LA max, left atrial maximum volume index; MV, mitral valve).

Early diastolic strain rate in relation to systolic and diastolic function 655

by guest on March 6, 2016

http://eurheartj.oxfordjournals.org/D

ownloaded from

worse outcome. These findings suggest that deformation-basedmeasures of global relaxation may be of particular importance inthis group of patients. Whether E/e′sr, as a more precise measureof global myocardial relaxation is related to exercise-induced eleva-tions in LV filling pressure remains to be demonstrated.

LimitationsWhile the present study to the best of our knowledge is the largestever in acute MI to assess both comprehensive conventional anddeformation-based systolic and diastolic function, we only assesseddeformation in the longitudinal direction. Circumferential deform-ation and analyses of torsion and untwist have been linked to poordiastolic function and it is possible that these measures could moreaccurately reflect global myocardial relaxation properties. We didnot measure other previously reported diastolic deformationindices such as E/Ds, E/e′sr-ivr and cannot assess their relative im-portance in comparison with E/e′sr. We excluded patients with AF,paced rhythm, severe AS, and severe MR and the present resultsare therefore not necessarily applicable to these important sub-groups. The analyses done in this study were performed using a pro-prietary system (EchoPac, GE) and it is possible that using differentsoftware could have resulted in different conclusions. Image qualityis a significant obstacle in deformation analysis and it is possiblethat our feasibility and reproducibility would differ from othercentres. We included both STEMI and NSTEMI and although wefound no difference in the prognostic value of E/e′sr between thesetwo groups, this could be due to the low-sample size. We cannotaddress the potential impact of post-MI medication on changes inE/e′sr or any possible interactions between these in relation to theoutcome. New onset AF may be significantly underestimated inthis study due toclinically silent episodesofAF whichcannotbequan-tified due to the lack of 24 h monitoring;33 however, the prognosticvalue of E/e′sr was mainly driven by mortality and HF admission.Cut-off values reported in our study are data-driven and need tobe independently validated in new studies.

ConclusionDeformation-based analysis of global myocardial relaxation proper-ties using E/e′sr confers important prognostic information independ-ently of conventional diastolic function indices and systolicdeformation and is superior to the velocity-based E/e′ ratio inpatients with acute MI.

FundingThe authors would like to thank the following for providing financialsupport to echocardiographic equipment and analytical software:Fondation Juchum, Switzerland; Toyota Fonden, Denmark; BeckettFonden, Denmark, and Danielsen Fonden, Denmark.

Conflict of interest: none declared.

References1. Moller JE, Egstrup K, Kober L, Poulsen SH, Nyvad O, Torp-Pedersen C. Prognostic

importance of systolic and diastolic function after acute myocardial infarction. AmHeart J 2003;145:147–153.

2. Hillis GS, Moller JE, Pellikka PA, Gersh BJ, Wright RS, Ommen SR, Reeder GS, Oh JK.Noninvasive estimation of left ventricular filling pressure by e/e′ is a powerful

predictor of survival after acute myocardial infarction. J Am Coll Cardiol 2004;43:360–367.

3. Jons C, Joergensen RM, Hassager C, Gang UJ, Dixen U, Johannesen A, Olsen NT,HansenTF, Messier M, Huikuri HV, Thomsen PE. Diastolic dysfunction predicts new-onset atrial fibrillation and cardiovascular events in patients with acute myocardialinfarction and depressed left ventricular systolic function: a CARISMA substudy.Eur J Echocardiogr 2010;11:602–607.

4. Jons C, Jacobsen UG, Joergensen RM, Olsen NT, Dixen U, Johannessen A, Huikuri H,MessierM, McNitt S, Thomsen PE. The incidence andprognostic significance of new-onset atrial fibrillation inpatientswith acutemyocardial infarctionand left ventricularsystolic dysfunction: a CARISMA substudy. Heart Rhythm 2011;8:342–348.

5. Møller JE, Pellikka Pa, Hillis GS, Oh JK. Prognostic importance of diastolic functionand filling pressure in patients with acute myocardial infarction. Circulation 2006;114:438–444.

6. Dokainish H, Sengupta R, Pillai M, Bobek J, Lakkis N.Usefulness of newdiastolic strainand strain rate indexes for the estimation of left ventricular filling pressure. Am JCardiol 2008;101:1504–1509.

7. Kimura K, Takenaka K, Ebihara A, Okano T, Uno K, Fukuda N, Ando J, Fujita H,Morita H, Yatomi Y, Nagai R. Speckle tracking global strain rate e/e′ predicts lvfilling pressure more accurately than traditional tissue Doppler e/e′ . Echocardiog-raphy 2012;29:404–410.

8. Wang J, Khoury DS, Thohan V, Torre-Amione G, Nagueh SF. Global diastolic strainrate for the assessment of left ventricular relaxation and filling pressures. Circulation2007;115:1376–1383.

9. Thune JJ, Solomon SD. Left ventricular diastolic function following myocardial infarc-tion. Curr Heart Fail Rep 2006;3:170–174.

10. Høfsten DE, Løgstrup BB, Møller JE, Pellikka PA, Egstrup K. Abnormal glucose me-tabolism in acute myocardial infarction: influence on left ventricular function andprognosis. JACC Cardiovascular imaging 2009;2:592–599.

11. Ng AC, Delgado V, Bertini M, van der Meer RW, Rijzewijk LJ, Hooi Ewe S,Siebelink HM, Smit JW, Diamant M, Romijn JA, de Roos A, Leung DY, Lamb HJ,Bax JJ. Myocardial steatosis and biventricular strain and strain rate imaging in patientswith type 2 diabetes mellitus. Circulation 2010;122:2538–2544.

12. Thygesen K, Alpert JS, Jaffe AS, Simoons ML, Chaitman BR, White HD. Third univer-sal definition of myocardial infarction. Eur Heart J 2012;33:2551–2567.

13. Steg PG, James SK, Atar D, Badano LP, Blomstrom-Lundqvist C, Borger MA, DiMario C, Dickstein K, Ducrocq G, Fernandez-Aviles F, Gershlick AH, Giannuzzi P,Halvorsen S, Huber K, Juni P, Kastrati A, Knuuti J, Lenzen MJ, Mahaffey KW,Valgimigli M, van’t Hof A, Widimsky P, Zahger D. ESC guidelines for the managementof acute myocardial infarction in patients presenting with ST-segment elevation. EurHeart J 2012;33:2569–2619.

14. Lameire N, Adam A, Becker CR, Davidson C, McCullough PA, Stacul F, Tumlin J.Baseline renal function screening. Am J Cardiol 2006;98:21K–26K.

15. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH,RomanMJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MS, Stewart WJ.Recommendations for chamber quantification: a report from the American Societyof Echocardiography’s Guidelines and Standards Committee and the ChamberQuantification Writing Group, developed in conjunction with the European Asso-ciation of Echocardiography, a branch of the European Society of Cardiology. J AmSoc Echocardiogr 2005;18:1440–1463.

16. Sampson UK, Pfeffer MA, McMurray JJ, Lokhnygina Y, White HD, Solomon SD. Pre-dictors of stroke in high-risk patients after acute myocardial infarction: insights fromthe valiant trial. Eur Heart J 2007;28:685–691.

17. Loh E, Sutton MS, Wun CC, Rouleau JL, Flaker GC, Gottlieb SS, Lamas GA, Moye LA,Goldhaber SZ, Pfeffer MA. Ventricular dysfunction and the risk of stroke after myo-cardial infarction. N Engl J Med 1997;336:251–257.

18. Bonita R. Epidemiology of stroke. Lancet 1992;339:342–344.19. van Buuren S, Boshuizen HC, Knook DL. Multiple imputation of missing blood pres-

sure covariates in survival analysis. Stat Med 1999;18:681–694.20. He Y. Missing data analysis using multiple imputation: getting to the heart of the

matter. Circ Cardiovasc Qual Outcomes 2010;3:98–105.21. Stoylen A, Skjelvan G, Skjaerpe T. Flow propagation velocity is not a simple index of

diastolic function in early filling. A comparative study of early diastolic strain rate andstrain rate propagation, flow and flow propagation in normal and reduced diastolicfunction. Cardiovasc Ultrasound 2003;1:3.

22. Stoylen A, Slordahl S, Skjelvan GK, Heimdal A, Skjaerpe T. Strain rate imaging innormal and reduced diastolic function: comparison with pulsed Doppler tissueimaging of the mitral annulus. J Am Soc Echocardiogr 2001;14:264–274.

23. Azevedo CF, Amado LC, Kraitchman DL, Gerber BL, Osman NF, Rochitte CE,Edvardsen T, Lima JA. Persistentdiastolicdysfunction despite complete systolic func-tional recoveryafter reperfusedacutemyocardial infarctiondemonstrated by taggedmagnetic resonance imaging. Eur Heart J 2004;25:1419–1427.

24. Ersboll M, Andersen MJ, Valeur N, Mogensen UM, Waziri H, Moller JE, Hassager C,Sogaard P, Kober L. The prognostic value of left atrial peak reservoir strain in acute

M. Ersbøll et al.656

by guest on March 6, 2016

http://eurheartj.oxfordjournals.org/D

ownloaded from

myocardial infarction is dependent on left ventricular longitudinal function and leftatrial size. Circ Cardiovasc Imaging 2013;6:26–33.

25. Hung C-L, Verma A, Uno H, Shin S-H, Bourgoun M, Hassanein AH, McMurray JJ,Velazquez EJ, Kober L, Pfeffer Ma, Solomon SD. Longitudinal and circumferentialstrain rate, left ventricular remodeling, and prognosis after myocardial infarction. JAm Coll Cardiol 2010;56:1812–1822.

26. Gjesdal O, Helle-Valle T, Hopp E, Lunde K, Vartdal T, Aakhus S, Smith HJ, Ihlen H,Edvardsen T. Noninvasive separation of large, medium, and small myocardial infarctsin survivors of reperfused st-elevation myocardial infarction: A comprehensivetissue Doppler and speckle-tracking echocardiography study. Circ Cardiovascimaging 2008;1:189–196, 182 p following 196.

27. Hamm CW, Bassand JP, Agewall S, Bax J, Boersma E, Bueno H, Caso P, Dudek D,Gielen S, Huber K, Ohman M, Petrie MC, Sonntag F, Uva MS, Storey RF, Wijns W,Zahger D. ESC guidelines for the management of acute coronary syndromes inpatients presenting without persistent ST-segment elevation: the task force forthe management of acute coronary syndromes (ACS) in patients presentingwithout persistent ST-segment elevation of the European Society of Cardiology(ESC). Eur Heart J 2011;32:2999–3054.

28. Ersboll M, Valeur N, Mogensen UM, Andersen M, Greibe R, Moller JE, Hassager C,Sogaard P, Kober L. Global left ventricular longitudinal strain is closely associatedwith increased neurohormonal activation after acute myocardial infarction inpatients with both reduced and preserved ejection fraction: a two-dimensionalspeckle tracking study. Eur J Heart Fail 2012;14:1121–1129.

29. Kasner M, Gaub R, Sinning D, Westermann D, Steendijk P, Hoffmann W,Schultheiss H-P, Tschope C. Global strain rate imaging for the estimation of diastolic

function in HFNEF compared with pressure-volume loop analysis. Eur J Echocardiogr2010;11:743–751.

30. Dokainish H, Nguyen JS, Sengupta R, Pillai M, Alam M, Bobek J, Lakkis N. Do addition-al echocardiographic variables increase the accuracy of e/e′ for predicting left ven-tricular filling pressure in normal ejection fraction? An echocardiographic andinvasive hemodynamic study. J Am Soc Echocardiogr 2010;23:156–161.

31. Shanks M, Ng AC, van de Veire NR, Antoni ML, Bertini M, Delgado V, Nucifora G,Holman ER, Choy JB, Leung DY, Schalij MJ, Bax JJ. Incremental prognostic value ofnovel left ventricular diastolic indexes for prediction of clinical outcome in patientswith st-elevation myocardial infarction. Am J Cardiol 2010;105:592–597.

32. Andersen MJ, Ersboll M, Bro-Jeppesen J, Gustafsson F, Hassager C, Kober L,Borlaug BA, Boesgaard S, Kjaergaard J, Moller JE. Exercise hemodynamics in patientswith and without diastolic dysfunction and preserved ejection fraction after myocar-dial infarction. Circ Heart Fail 2012;5:444–451.

33. Camm AJ, Lip GY, De Caterina R, Savelieva I, Atar D, Hohnloser SH, Hindricks G,Kirchhof P, Bax JJ, Baumgartner H, Ceconi C, Dean V, Deaton C, Fagard R,Funck-Brentano C, Hasdai D, Hoes A, Knuuti J, Kolh P, McDonagh T, Moulin C,Popescu BA, Reiner Z, Sechtem U, Sirnes PA, Tendera M, Torbicki A, Vahanian A,Windecker S, Vardas P, Al-Attar N, Alfieri O, Angelini A, Blomstrom-Lundqvist C,Colonna P, De Sutter J, Ernst S, Goette A, Gorenek B, Hatala R, Heidbuchel H,Heldal M, Kristensen SD, Le Heuzey JY, Mavrakis H, Mont L, Filardi PP,Ponikowski P, Prendergast B, Rutten FH, Schotten U, Van Gelder IC,Verheugt FW. 2012 focused update of the ESC guidelines for the management ofatrial fibrillation: an update of the 2010 ESC Guidelines for the Management ofAtrial Fibrillation. Developed with the special contribution of the European HeartRhythm Association. Eur Heart J 2012;33:2719–2747.

Early diastolic strain rate in relation to systolic and diastolic function 656a

by guest on March 6, 2016

http://eurheartj.oxfordjournals.org/D

ownloaded from