Ccf

YKTT

F

Btns

Oacp

MYei(w�f

RTsuiow

ICpadotCc

ADMac

1

ombined assessment of left ventricular dyssynchrony andontractility by speckled tracking strain imaging: A novel indexor predicting responders to cardiac resynchronization therapy

asuya Inden, MD, PhD, Rieko Ito, MT, Naoki Yoshida, MD, Hiromi Kamiya, MD,azuhisa Kitamura, MD, Tomoya Kitamura, MD, PhD, Masayuki Shimano, MD, PhD,omohiro Uchikawa, MD, Yukiomi Tsuji, MD, PhD, Rei Shibata, MD, PhD, Makoto Hirai, MD, PhD,oyoaki, Murohara, MD, PhD

rom the Department of Cardiology, Nagoya University Graduate School of Medicine, Nagoya, Japan.

rtP

Cbtf

KEm

AEttdtpfv

(

ACKGROUND Mechanical dyssynchrony is an important factor inhe response to cardiac resynchronization therapy (CRT). However,o echocardiographic measure can improve prediction of caseelection for CRT.

BJECTIVE The purpose of this study was to assess the efficacy ofnewly combined echocardiographic index for ventricular dyssyn-

hrony and contractility using speckled tracking strain analysis toredict responders to CRT.

ETHODS Forty-seven patients with severe heart failure in Nework Heart Association functional class III/IV, left ventricularjection fraction �35%, and QRS duration �130 ms were includedn the study. Echocardiography was performed, and a novel indexi-Index), the product of radial dyssynchrony and radial strain,as calculated. Responder to CRT was defined as a patient with a15% decrease in left ventricular end-systolic volume at 6-month

ollow-up.

ESULTS Thirty-two patients (68%) were classified as responders.he i-Index was significantly higher in responders than in nonre-ponders (3,450 � 1180 vs 1,481 � 841, P �.001). The areander receiver operator characteristic curve was 0.92 for the-Index, which was better than the index of radial dyssynchronynly (0.74). A cutoff value of i-Index �2,000 predicted responders

fQmfddasgileelepted January 12, 2010.)

547-5271/$ -see front matter © 2010 Heart Rhythm Society. All rights reserved

adial dyssynchrony had 81% sensitivity and 53% specificity. Fur-hermore, i-Index decreased in responders (1,985 � 1261,�.001) but not in nonresponders (1,684 � 866, P � .48).

ONCLUSION Our findings suggest that a novel combined indexy radial strain echocardiography might be a predictor of responseo CRT. The value of this novel echocardiographic index requiresurther assessment in larger studies.

EYWORDS Cardiac resynchronization therapy; Dyssynchrony;chocardiography; Heart failure; Myocardial contractility; Pace-aker; Responder

BBREVIATIONS CRT � cardiac resynchronization therapy;F � ejection fraction; LV � left ventricular; LVEDV � left ven-ricular end-diastolic volume; LVESV � left ventricular end-sys-olic volume; NYHA � New York Heart Association; RD � radialyssynchrony; RS � radial strain; SDt � standard deviation ofime to peak systolic strain of six LV segments; SPWMD � septal–osterior wall-motion delay; Ts-SD � standard deviation of timerom QRS to peak systolic velocity in ejection phase for 12 leftentricular segments

Heart Rhythm 2010;7:655–661) © 2010 Heart Rhythm Society. All

ith 94% sensitivity and 80% specificity. The index using only rights reserved.

ntroductionardiac resynchronization therapy (CRT) is a new thera-eutic modality for improving symptoms, cardiac function,nd prognosis in patients with severe heart failure.1–3 Ran-omized clinical trials have demonstrated that the majorityf patients receive benefit from CRT; however, at the sameime, approximately 30% of patients are nonresponders toRT.1,4,5 Although the criteria for inclusion in randomizedlinical trials were New York Heart Association (NYHA)

ddress reprint requests and correspondence: Dr. Yasuya Inden,epartment of Cardiology, Nagoya University Graduate School ofedicine, 65 Tsurumaicho, Showaku, Nagoya, 466-8550, Japan. E-mail

ddress: inden@med.nagoya-u.ac.jp. (Received November 23, 2009; ac-

unctional class, cardiac ejection fraction (EF), and ECGRS complex, several studies have suggested that indices ofechanical dyssynchrony by cardiac imaging are superior

or predicting responder to CRT compared with ECG QRSuration. Therefore, assessment of left ventricular (LV)yssynchrony by echocardiography has been emphasized aspredictive marker.6–15 However, in the Predictors of Re-

ponse to CRT (PROSPECT) trial, no single echocardio-raphic measure of dyssynchrony was recommended tomprove patient selection for CRT beyond current guide-ines.16 These echocardiographic measures can evaluate thexistence and site of dyssynchrony but cannot evaluate thexistence of scar or local myocardial contractility, particu-

arly in patients with ischemia.

. doi:10.1016/j.hrthm.2010.01.015

ctnctdic

MSFisf(fiwa6tts

DTtaapcwvotwMiigp

EAmwmwbaSDsPflL

eveb(puCa

SSilsmooatcdTssrfeicoedstTcS

DR�i

SCpdaaLbCcvd

656 Heart Rhythm, Vol 7, No 5, May 2010

A novel method of speckle tracking radial strain imagingan evaluate dyssynchrony15 and quantify myocardial sys-olic function and viability.17,18 Accordingly, we propose aovel index (i-Index), which includes both LV dyssyn-hrony and contractility. The objective of this study was toest the hypotheses that a novel combined assessment of LVyssynchrony and contractility by radial strain imaging candentify responders to CRT from among patients with echo-ardiographic dyssynchrony.

ethodstudy populationorty-seven consecutive patients who underwent CRT were

ncluded in the study. Selection criteria used for CRT in thistudy were (1) drug-refractory severe heart failure in NYHAunctional class III or IV, (2) depressed LVEF (�35%), and3) wide QRS complex (�130 ms). No patients had atrialbrillation. All patients received a CRT device with orithout automatic defibrillator function and were fully ex-

mined by echocardiography before CRT implantation andmonths after CRT. All patients received optimal drug

reatment, including angiotensin-converting enzyme inhibi-ors or angiotensin receptor blockers, beta-blockers, andpironolactone, if tolerated.

evice implantation and optimizationhe LV lead was placed on the posterior or lateral wall

hrough the coronary sinus in all patients. The right atrialnd ventricular leads were positioned in the right atrialppendage and right ventricular mid or apical septum. Therocedure for implantation of the right ventricular leadonsisted of advancing the lead out the pulmonary artery,ithdrawing the lead until it dropped below the pulmonicalve, and then advancing the lead into the midventricularr apical septum with the use of stylet. Proper right ven-ricular lead was confirmed by fluoroscopy. The leadsere connected to CRT devices (InSync III or Concerto,edtronic, Inc., Minneapolis, MN, USA). Atrioventricular

nterval and interventricular delay were optimized to max-mize the left ventricular outflow tract velocity–time inte-ral by echocardiography within 1 week after device im-lantation.

chocardiographyll echocardiographic studies were performed using a com-ercially available system (Vivid 7, GE Healthcare, Mil-aukee, WI, USA). Patients were examined before and 6onths after CRT implantation. Echocardiographic dataere obtained in the apical (two-chamber and four-cham-er) and parasternal (short-axis and long-axis) views using1.5- to 4.0-MHz transducer at depths of 12 to 16 cm.

tandard M-mode, gray-scale two-dimensional, and tissueoppler images were obtained during breath-hold and

tored in cineloop format from three consecutive beats.ulsed-wave Doppler of the LV and right ventricular out-ow tract also was recorded. LV end-diastolic diameter and

V end-systolic diameter were measured from M-mode c

chocardiography of the parasternal long-axis view. Leftentricular end-diastolic volume (LVEDV), left ventricularnd-systolic volume (LVESV), and EF were assessed byiplane Simpson’s rule. Septal–posterior wall-motion delaySPWMD)6 and standard deviation of time from QRS toeak systolic velocity in ejection phase for 12 left ventric-lar segments (Ts-SD)7,9 were calculated before and afterRT, and other conventional echocardiographic predictorslso were calculated, if needed, before CRT.

peckle tracking strain analysistandard two-dimensional gray-scale images were acquired

n the parasternal short-axis view at the papillary muscleevel and analyzed using the software for speckle trackingtrain echocardiography (EchoPAC 6.0, GE Healthcare). Ainimum rate of 50 Hz was required for reliable operation

f this program, and images were recorded with frame ratesf 50 to 80 Hz in this study (63.6 � 4.8 Hz). This customcoustic software tracks movement of stable acoustic pat-erns, called speckles, in myocardial tissues. Tracking oc-urs frame by frame throughout the cardiac cycle. Myocar-ial radial thickening was represented as a positive value.he short-axis image was then divided into six standardegments with corresponding time–strain curves from eachegment. After these echocardiographic measurements wereecorded by operators, all strain measurements were per-ormed by a single observer blinded to the clinical andchocardiographic information without two-dimensionalmages used for strain analysis. For discriminating dyssyn-hrony indices from strain images, several parameters werebtained from the time–strain curves: maximal time differ-nce of peak systolic strain among six LV segments [radialyssynchrony (RD)], standard deviation of time to peakystolic strain of six LV segments (SDt), and mean radialhickening of six LV segments [radial strain (RS); Figure 1].o add assessment of contractility to dyssynchrony, wealculated a novel index (i-Index), which is the product ofDt and RS.

efinition of responders to CRTesponders to CRT were defined as patients displaying a15% reduction in LVESV at 6-month follow-up after

mplantation.

tatistical analysisontinuous variables are expressed as mean � SD and com-ared using two-tailed Student t-test for paired and unpairedata. Categorical data are expressed as number and percent-ge and compared using Chi-square test. Linear regressionnalysis was performed to assess the relationship betweenVESV and baseline dyssynchrony indices. The extent ofaseline LV dyssynchrony needed to predict response toRT was determined by receiver operator characteristicurve analysis. The variables that were significant in uni-ariate analysis were used in logistic regression analysis toetermine independent predictors of evolution. P �.05 was

onsidered significant.

RPT6pIas3te((r(g(

Lw6

VAescr4e2s(7

FtdasSsofapmds

T

c�Nd

657Inden et al Combined Dyssynchrony and Contractility Index for CRT

esultsatient baseline characteristicshe baseline characteristics of the 47 patients (27 men; age7 � 12 years) in this study are summarized in Table 1. Allatients had severe heart failure (40 NYHA functional classII, 7 class IV), severe LV dysfunction (EF 26.6% � 7.9%),nd wide ECG QRS complex (162 � 31 ms). ECG QRShapes were as follows: left bundle branch block type (n �9), right bundle branch block type (n � 4), and intraven-ricular conduction disturbance (n � 4). Underlying dis-ases in this study population were dilated cardiomyopathy72%), ischemic heart disease (13%), cardiac sarcoidosis6%), and valvular heart disease (2%). All patients hadeceived optimal medical therapy, including beta-blockers87%), angiotensin-converting enzyme inhibitors or an-iotensin-receptor blockers (72%), and spironolactone77%). CRT systems were implanted successfully, and

igure 1 Measurements of speckledracking radial strain analysis. Left: Two-imensional echocardiography from short-xis view at the papillary muscle level andix left ventricular (LV) segments. Right:peckle tracking time–strain curves fromix LV segments. Open squares � averagef peak systolic strains (radial thickening)rom six LV segments (RS [%]). Solidrrowed line � maximum difference ofeak systolic strain among six LV seg-ents (RD). Dashed arrowed lines � stan-

ard deviation of time to peak systolictrain for six LV segments (SDt).

able 1 Patient baseline characteristic and left ventricular dys

All (n � 47) Resp

Age (years) 67 � 12 7Male (n, %) 27 (57) 17NYHA class (III/IV) 40/7 26/Etiology (D/I/S/V) 34/9/3/1 25/QRS interval (ms) 162 � 31 16ECG (L/R/IV) 39/4/4 28/BNP (pg/mL) 424 � 404 41LVEDV (mL) 215 � 87 21LVESV (mL) 161 � 79 15LVDd (mm) 69 � 10 6LVDs (mm) 59 � 12 5LVEF (%) 26.6 � 7.9 27SPWMD (ms) 184 � 94 20Ts-SD (ms) 54 � 28 6RD (ms) 354 � 159 38SDt (ms) 147 � 69 16RS (%) 21.2 � 11.7 24i-Index (ms%) 2,821 � 1,420 3,45

BNP � brain natriuretic peptide; ECG IV � intraventricular conduction disturbanardiomyopathy; Etiology I � ischemic heart disease; Etiology S � cardiac sarcoidos

left ventricular end-systolic diameter; LVEDV � left ventricular end-diastolic voluYHA � New York Heart Association; SPWMD � septal–posterior wall-motion delay; RD

eviation of time to peak systolic strain; Ts-SD � standard deviation of time from QRS to p

V leads were positioned in the posterior or lateral LVall in all patients. No complications occurred during-month follow-up.

entricular function response to CRTll patients were reevaluated at 6-month follow-up. In the

ntire patient group, a significant improvement in clinicaltatus was observed, with a reduction of NYHA functionallass (35 class II, 10 class III, 2 class IV, P �.001) and aeduction in plasma brain natriuretic peptide levels from24 � 404 pg/mL to 263 � 301 pg/mL (P �.001). Onchocardiography, LVEDV was significantly reduced from15 � 87 mL to 174 � 79 mL (P �.001), and LVESV wasignificantly reduced from 161 � 79 mL to 119 � 75 mLP �.001). LVEF was significantly improved from 26.6% �.9% to 34.4% � 13.0% (P �.001).

ny indices

(n � 32) Nonresponder (n � 15) P value

60 � 14 .00410 (67) .3812/3 .759/4/1/1 .35146 � 30 .02

11/2/2 .480 442 � 383 .82

227 � 96 .52175 � 93 .4372 � 11 .1663 � 12 .14

25.2 � 7.9 .42145 � 84 .04940 � 14 .02

8 283 � 143 .03100 � 48 �.001

.7 15.1 � 5.9 .0180 1,481 � 841 �.001

� left bundle branch block; ECG R � right bundle branch block; Etiology D � dilatedgy V � valvular heart disease; LVDd � left ventricular end-diastolic diameter; LVDs

F � left ventricular ejection fraction; LVESV � left ventricular end-systolic volume;imum difference of peak systolic strain; RS � mean radial thickening; SDt � standard

synchro

onders

0 � 9(53)65/2/09 � 29

2/25 � 420 � 835 � 737 � 107 � 11.3 � 8.02 � 951 � 308 � 159 � 67.1 � 120 � 1,1

ce; ECG Lis; Etiolome; LVE� max

eak systolic velocity in ejection phase for 12 left ventricular segments.

RArrLLL(t

21n9

1mysp(eat

e(Tmassrlrrrctp

rasWas

PRfcac

Fscmesns(

Tt

w

658 Heart Rhythm, Vol 7, No 5, May 2010

esponders versus nonresponders to CRTt 6-month follow-up, 32 (68%) patients were classified as

esponders to CRT according to the criterion of �15%eduction in LVESV. Responders showed a reduction inVEDV (from 210 � 83 mL to 150 � 56 mL, P �.001) andVESV (from 155 � 73 mL to 95 � 49 mL, P �.001).VEF also improved significantly after CRT in responders

from 27.2% � 8.0% to 38.5% � 12.4%, P �.001). In con-

igure 2 Relationship between baseline echocardiographic mea-urements and left ventricular reverse remodeling after 6 months ofardiac resynchronization therapy. Linear regression analysis de-onstrated modest but significant relationships between baseline

chocardiographic measurements and reduction in left ventricular end-ystolic volume (�LVESV) during follow-up after cardiac resynchro-ization therapy. A: i-Index. B: Maximum difference of peak systolictrain (RD). C: Standard deviation of time to peak systolic strainSDt).

rast, nonresponders showed no changes in both LVEDV (from v

27 � 96 mL to 234 � 97 mL, P � .70) and LVESV (from75 � 93 mL to 171 � 95 mL, P � .68), and they showedo improvement in LVEF (from 25.2% � 7.9% to 25.7% �.8%, P � .83).

Baseline characteristics of both groups are given in Table. Clinical status of heart failure and echocardiographiceasurements were similar, but nonresponders were

ounger than responders in this study population and had ahorter QRS interval. No significant differences in LV leadositions were seen between responders and nonrespondersLV anterolateral/lateral/posterolateral: 10/14/8 in respond-rs and 5/6/4 in nonresponders). Atrioventricular intervalnd interventricular delay also did not differ between thewo groups (data not shown).

LV dyssynchrony parameters at baseline showed differ-nces between responders and nonresponders to CRTTable 1). The conventional measurements (SPWMD ands-SD) were larger in responders than in nonresponders. Theeasurements by speckle tracking analysis (RD, SDt, i-Index)

lso were significantly larger in responders than in nonre-ponders. Furthermore, the index of wall thickening (RS)howed a significantly larger value in responders. Linearegression analysis demonstrated modest but significant re-ationships between baseline indices (RD, SDt, i-Index) andeduction in LVESV during follow-up (Figure 2). Logisticegression analysis showed that the independent predictor ofesponse to CRT was only i-Index (odds ratio 1.004, 95%onfidence interval 1.000–1.008, P � .043; Table 2), andhat other factors including age and QRS interval did notrove independent.

After 6 months of CRT, responders showed a significanteduction in SPWMD, Ts-SD, RD, SDt, and i-Index asssessed by echocardiography, but nonresponders did nothow a significant reduction except for SPWMD (Figure 3).

all thickening assessed RS did not change significantlyfter 6-month follow-up in both responders and nonre-ponders.

rediction of response to CRTeceiver operator characteristic curve analysis was per-

ormed to calculate an optimal cutoff value for dyssyn-hrony indices to predict response to CRT (Figure 4). Therea under the curve for i-Index was 0.92, and the optimalutoff value to predict response to CRT was 2,000 ms%,

able 2 Predictors of responders to cardiac resynchronizationherapy in logistic regression analysis

Odds ratio 95% Confidence interval P value

Age (years) 0.934 0.805–1.082 .363QRS interval (ms) 1.062 0.988–1.143 .103SPWMD (ms) 1.015 0.996–1.035 .131Ts-SD (ms) 1.081 0.989–1.181 .088RD (ms) 0.994 0.979–1.009 .442i-Index (ms%) 1.004 1.000–1.008 .043

RD � maximum difference of peak systolic strain; SPWMD � septal–posteriorall-motion delay; Ts-SD � standard deviation of time from QRS to peak systolic

elocity in ejection phase for 12 left ventricular segments.

yacsdiI

hoRd

DTLnfpcrais

SItcsp

FcAovsM�

Tp

I

Swv

FwDS(ws

659Inden et al Combined Dyssynchrony and Contractility Index for CRT

ielding 94% sensitivity and 80% specificity (Table 3). Therea under the curve for SDt was 0.81, and the optimalutoff value was 105 ms, yielding 88% sensitivity and 73%pecificity. The other indices (RD, RS, SPWMD, Ts-SD)id not show higher area under the curve compared with the-Index. The positive and negative predictive values of i-ndex were 91% and 86%, respectively, and both were the

igure 3 Changes in echocardiographic measurements and i-Index afterardiac resynchronization therapy in responders and nonresponders.: Septal–posterior wall-motion delay (SPWMD). B: Standard deviationf time from QRS to peak systolic velocity in ejection phase for 12 leftentricular segments (Ts-SD). C: Maximum difference of peak systolictrain (RD). D: Standard deviation of time to peak systolic strain (SDt). E:

ean radial thickening (RS., F: i-Index. Open bars � baseline. Solid bars6-month follow-up.

igure 4 Receiver operating characteristic curve analysis for predictionith other radial strain echocardiographic measurements. Right: Comparisoppler echocardiography. Area under the curve (AUC) was significantlyensitivity and the specificity were 94% and 80%, respectively, for i-Index.RD); green line � standard deviation of time to peak systolic strain (SDall-motion delay (SPWMD); dark green line � standard deviation of ti

egments (Ts-SD).

ighest among the indices. Figure 5 shows the distributionf SDt and RS that composed the i-Index before CRT.esponders and nonresponders to CRT were approximatelyivided by the curving line y � 2,000/x.

iscussionhis is the first study to assess radial LV dyssynchrony andV contractility for predicting response to CRT using aovel combined index. The name i-Index, which is derivedrom Inden (the author’s name), shows a higher value inatients with both positive dyssynchrony and concealedontractile property by speckle tracking strain echocardiog-aphy and can predict mid-term responders to CRT frommong patients with severe heart failure with high sensitiv-ty and high specificity compared with the published dys-ynchrony measurements.

peckle tracking radial strain imaging and i-Indexn this study, we used speckle tracking radial strain imagingo evaluate heart failure patients with LV dyssynchrony. Wealculated two measurements from time to peak systolictrain of six LV segments: RD (maximal time difference ofeak systolic strain among six LV segments) and SDt (stan-

able 3 Sensitivity, specificity, and area under the curve toredict responders to cardiac resynchronization therapy

ndex (cutoff value) Sensitivity Specificity Area under the curve P value

SPWMD (130 ms) 0.72 0.53 0.68 .047Ts-SD (32 ms) 0.97 0.27 0.76 .005RD (258 ms) 0.81 0.53 0.70 .027SDt (105 ms) 0.88 0.73 0.81 .01RS (19%) 0.66 0.73 0.74 .008i-Index (2,000 ms%) 0.94 0.80 0.92 �.001

RD � maximum difference of peak systolic strain; RS � mean radial thickening;Dt � standard deviation of time to peak systolic strain; SPWMD � septal–posteriorall-motion delay; Ts-SD � standard deviation of time from QRS to peak systolicelocity in ejection phase for 12 left ventricular segments.

onders to cardiac resynchronization therapy. Left: Comparison of i-IndexIndex with major measurements by M-mode echocardiography and tissuewith i-Index (0.92) than with other measurements by echocardiography.

es � i-Index; dark blue line � maximum difference of peak systolic strainle line � mean radial thickening (RS); light blue line � septal–posterior

QRS to peak systolic velocity in ejection phase for 12 left ventricular

of respon of i-greaterRed lin

t); purpme from

dsrgylS7aafgowarars(rd

(lr3cFiphi

ITimLdRsannreipeeimtdIwwoRsfCt(e

CtOmtsmdhlg

gvrenrtoma1

Frdrsb

660 Heart Rhythm, Vol 7, No 5, May 2010

ard deviation of time to peak systolic strain of six LVegments). Suffoletto et al15 reported that RD predictedesponder with 89% sensitivity and 83% specificity. Del-ado et al19 showed that speckle tracking radial strain anal-sis (83% sensitivity and 80% specificity) was superior toongitudinal and circumferential strain analyses and thatDt by radial strain analysis also detected responders with6% sensitivity and 60% specificity. We used radial strainnalysis (rather than longitudinal or circumferential strainnalysis) because radial strain analysis provided larger dif-erences compared with other analyses as shown by Del-ado et al.19 Thickening of myocardium toward the center isbserved by radial strain analysis, whereas shortening to-ard the tangential projection is evaluated by longitudinal

nd circumferential strain analyses. This difference of di-ection might provide the differences of percent shorteningnd timing among these three analyses. In our study, SDt byadial strain analysis could predict responders with 88%ensitivity and 73% specificity, and SDt was superior to RDTable 3). These data also showed that speckle trackingadial strain analysis was a good method for evaluating LVyssynchrony and predicting responders to CRT.

We showed that LVEF did not differ, but wall thickeningRS) in responders before implantation was significantlyarger than in nonresponders (Table 1), and RS predictedesponders with 66% sensitivity and 73% specificity (Table). These findings indicate that measurement of RS possiblyan be evaluated as an index for predicting responders.urthermore, we proposed a new index, the i-Index, which

s the product of the two measurements SDt and RS. i-Indexredicted responders with higher sensitivity (94%) andigher specificity (80%) compared with other indices used

igure 5 Distribution of two measurements composing the i-Index inesponders and nonresponders. Scatter diagram of measurements of stan-ard deviation of time to peak systolic strain (SDt; open circles) and meanadial thickening (RS; filled circles) shows that responders and nonre-ponders to cardiac resynchronization therapy are approximately dividedy the curving line y � 2000/x.

n this study. e

mplication and usefulness of the i-Indexhe indices for LV dyssynchrony used in previous studies,

ncluding the PROSPECT trial, all were time measure-ents. This is reasonable because dyssynchrony means thatV wall motion does not synchronize and the time lag isetected in LV wall motion. We used the measurement ofS, which is the extent of LV wall thickness. We demon-

trated that LV global EF did not differ before implantationnd that RS in responders was significantly higher than inonresponders. After follow-up, LVEF increased but RS didot increase in responders. This finding indicates that inesponders with high RS, low LVEF is attributed to thexistence of dyssynchrony and that CRT contributes tomprovement of LVEF by reducing LV dyssynchrony inatients with good LV contractility. Vartdal et al17 and Parkt al18 reported that speckle tracking strain imaging couldvaluate LV viability in patients with myocardial infarction.-Index is the product of two echocardiographic measure-ents and includes both LV dyssynchrony and LV contrac-

ility. Lafitte et al20 showed that a multiparametric echocar-iographic strategy was a better indicator of CRT response.t is possible that nonresponders with low contractility whoere judged to have LV dyssynchrony by other indicesere distinguished by i-Index. In addition, based on the dataf SDt and RS, i-Index is computed by multiplying SDt byS. Therefore, this new noninvasive method could be fea-

ible, reliable, and reproducible. i-Index decreased afterollow-up in responders, and we could follow the efficacy ofRT using this index. Because a modest correlation be-

ween i-Index and reduction of LVESV was demonstratedFigure 2), we could anticipate the extent of reverse remod-ling by CRT based on the value of i-Index.

omparison of responders and nonresponderso CRTne reason for unresponsiveness to CRT is believed to be aismatch between LV lead position and latest LV activa-

ion site due to limited venous anatomy.21 Existence of LVcar tissue in ischemic heart disease has been shown toake the clinical and structural response to CRT more

ifficult.22 Previous studies have suggested that ischemiceart disease, severe mitral regurgitation, and severely di-ated LV end-diastolic diameter are the underlying etiolo-ies associated with failed response to CRT.23–25

In this study, no significant differences in echocardio-raphic measurements, LV volumes, and LVEF before de-ice implantation were seen between responders and non-esponders (Table 1). LV end-diastolic diameter and LVnd-systolic diameter showed a tendency to be larger inonresponders (P � .16 and .14), similar to a previouseport.25 The population of this study had a small distribu-ion of ischemic heart disease (13%) compared with previ-us reports.23–25 In our study, LV dyssynchrony measure-ents showed significant differences between responders

nd nonresponders before device implantation study (Table), as reported in previous studies.6,7,9 SPWMD in respond-

rs was longer than in nonresponders, and Ts-SD, an index

fiiatuststupCb

SIorrsdptgr

COwcaicmcr

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

661Inden et al Combined Dyssynchrony and Contractility Index for CRT

rom tissue Doppler imaging, also was significantly longern responders. RD and SDt by speckle tracking radial strainmaging also was longer in responders than nonresponders,s reported previously.15,19,26 However, wall thickening byhe speckle tracking radial strain method in patients whonderwent CRT has not been reported. Septal reboundtretch,27 internal stretch fraction (mechanical dyscoordina-ion),28 and septal flash29 have been shown to predict re-ponse. In this study, RS was larger in responders to CRThan in nonresponders, and RS is believed to be one of theseful predictors of responders. Lower contractility is sup-osed to be one of the reasons for nonresponsiveness toRT, although it has been masked because of similar LVEFetween responders and nonresponders.

tudy limitationsn this study, all strain evaluation was performed by a singlebserver in order to avoid interobserver variation. We choseeduction of LVESV at 6-month follow-up as a marker foresponder to CRT because that measurement was used ineveral previous studies.7,16,19,26,30 This study was con-ucted in a single center and consisted of a small number ofatients. The follow-up period was relatively short. Thus,he present pilot study provides preliminary evidence sug-esting that the i-Index is a dependable indicator of CRTesponder. Future large multicenter studies are required.

onclusionur study findings suggest that a novel index (i-Index),hich includes combined assessment of radial dyssyn-

hrony and LV contractility by speckle tracking radial strainnalysis, might be a predictor of mid-term response to CRTn patients with heart failure. Assessment of left ventricularontractility using radial strain imaging will be an importanteans for evaluating LV function with mechanical dyssyn-

hrony. The value of this novel echocardiographic indexequires further assessment in larger studies.

eferences1. Abraham WT, Fisher WG, Smith AL, et al. Cardiac resynchronization in chronic

heart failure. N Engl J Med 2002;346:1845–1853.2. Bristow MR, Saxon LA, Boehmer J, et al. Cardiac-resynchronization therapy

with or without an implantable defibrillator in advanced chronic heart failure.N Engl J Med 2004;350:2140–2150.

3. Cleland JG, Daubert JC, Erdmann E, et al. The effect of cardiac resynchroni-zation on morbidity and mortality in heart failure. N Engl J Med 2005;352:1539–1549.

4. Young JB, Abraham WT, Smith AL, et al. Combined cardiac resynchronizationand implantable cardioversion defibrillation in advanced chronic heart failure:the MIRACLE ICD Trial. JAMA 2003;289:2685–2694.

5. Saxon LA, Boehmer JP, Hummel J, et al. Biventricular pacing in patients withcongestive heart failure: two prospective randomized trials. The VIGOR CHFand VENTAK CHF Investigators. Am J Cardiol 1999;83:120D–123D.

6. Pitzalis MV, Iacoviello M, Romito R, et al. Cardiac resynchronization therapytailored by echocardiographic evaluation of ventricular asynchrony. J Am CollCardiol 2002;40:1615–1622.

7. Yu CM, Fung WH, Lin H, Zhang Q, Sanderson JE, Lau CP. Predictors of leftventricular reverse remodeling after cardiac resynchronization therapy for heart

failure secondary to idiopathic dilated or ischemic cardiomyopathy. Am JCardiol 2003;91:684–688.

8. Bax JJ, Bleeker GB, Marwick TH, et al. Left ventricular dyssynchrony predictsresponse and prognosis after cardiac resynchronization therapy. J Am CollCardiol 2004;44:1834–1840.

9. Yu CM, Fung JW, Zhang Q, et al. Tissue Doppler imaging is superior to strainrate imaging and postsystolic shortening on the prediction of reverse remodelingin both ischemic and nonischemic heart failure after cardiac resynchronizationtherapy. Circulation 2004;110:66–73.

0. Bax JJ, Marwick TH, Molhoek SG, et al. Left ventricular dyssynchrony predictsbenefit of cardiac resynchronization therapy in patients with end-stage heartfailure before pacemaker implantation. Am J Cardiol 2003;92:1238–1240.

1. Notabartolo D, Merlino JD, Smith AL, et al. Usefulness of the peak velocitydifference by tissue Doppler imaging technique as an effective predictor ofresponse to cardiac resynchronization therapy. Am J Cardiol 2004;94:817–820.

2. Sogaard P, Egeblad H, Kim WY, et al. Tissue Doppler imaging predicts im-proved systolic performance and reversed left ventricular remodeling duringlong-term cardiac resynchronization therapy. J Am Coll Cardiol 2002;40:723–730.

3. Cazeau S, Bordachar P, Jauvert G, et al. Echocardiographic modeling of cardiacdyssynchrony before and during multisite stimulation: a prospective study.Pacing Clin Electrophysiol 2003;26:137–143.

4. Sogaard P, Egeblad H, Pedersen AK, et al. Sequential versus simultaneousbiventricular resynchronization for severe heart failure: evaluation by tissueDoppler imaging. Circulation 2002;106:2078–2084.

5. Suffoletto MS, Dohi K, Cannesson M, Saba S, Gorcsan J 3rd. Novel speckle-tracking radial strain from routine black-and-white echocardiographic images toquantify dyssynchrony and predict response to cardiac resynchronization ther-apy. Circulation 2006;113:960–968.

6. Chung ES, Leon AR, Tavazzi L, et al. Results of the Predictors of Response toCRT (PROSPECT) trial. Circulation 2008;117:2608–2616.

7. Vartdal T, Brunvand H, Pettersen E, et al. Early prediction of infarct size bystrain Doppler echocardiography after coronary reperfusion. J Am Coll Cardiol2007;49:1715–1721.

8. Park YH, Kang SJ, Song JK, et al. Prognostic value of longitudinal strain afterprimary reperfusion therapy in patients with anterior-wall acute myocardialinfarction. J Am Soc Echocardiogr 2008;21:262–267.

9. Delgado V, Ypenburg C, van Bommel RJ, et al. Assessment of left ventriculardyssynchrony by speckle tracking strain imaging comparison between longitu-dinal, circumferential, and radial strain in cardiac resynchronization therapy.J Am Coll Cardiol 2008;51:1944–1952.

0. Lafitte S, Reant P, Zaroui A, et al. Validation of an echocardiographic multipa-rametric strategy to increase responders patients after cardiac resynchronization:a multicentre study. Eur Heart J 2009;30:2880–2887.

1. Van de Veire NR, Schuijf JD, De Sutter J, et al. Non-invasive visualization ofthe cardiac venous system in coronary artery disease patients using 64-slicecomputed tomography. J Am Coll Cardiol 2006;48:1832–1838.

2. Bleeker GB, Kaandorp TA, Lamb HJ, et al. Effect of posterolateral scar tissueon clinical and echocardiographic improvement after cardiac resynchronizationtherapy. Circulation 2006;113:969–976.

3. Reuter S, Garrigue S, Barold SS, et al. Comparison of characteristics in respond-ers versus nonresponders with biventricular pacing for drug-resistant congestiveheart failure. Am J Cardiol 2002;89:346–350.

4. Gasparini M, Mantica M, Galimberti P, et al. Is the outcome of cardiac resyn-chronization therapy related to the underlying etiology? Pacing Clin Electro-physiol 2003;26:175–180.

5. Diaz-Infante E, Mont L, Leal J, et al. Predictors of lack of response to resyn-chronization therapy. Am J Cardiol 2005;95:1436–1440.

6. Gorcsan J 3rd, Tanabe M, Bleeker GB, et al. Combined longitudinal and radialdyssynchrony predicts ventricular response after resynchronization therapy.J Am Coll Cardiol 2007;50:1476–1483.

7. De Boeck BW, Teske AJ, Meine M, et al. Septal rebound stretch reflects thefunctional substrate to cardiac resynchronization therapy and predicts volumetricand neurohormonal response. Eur J Heart Fail 2009;11:863–871.

8. Kirn B, Jansen A, Bracke F, van Gelder B, Arts T, Prinzen FW. Mechanicaldiscoordination rather than dyssynchrony predicts reverse remodeling upon cardiacresynchronization. Am J Physiol Heart Circ Physiol 2008;295:H640–H646.

9. Parsai C, Baltabaeva A, Anderson L, Chaparro M, Bijnens B, Sutherland GR.Low-dose dobutamine stress echo to quantify the degree of remodelling aftercardiac resynchronization therapy. Eur Heart J 2009;30:950–958.

0. Knebel F, Schattke S, Bondke H, et al. Circumferential 2D-strain imaging for theprediction of long term response to cardiac resynchronization therapy. Cardio-

vasc Ultrasound 2008;6:28.