tissue doppler derived mechanical dyssynchrony does not change after cardiac resynchronization...

DOI: 10.1111/j.1540-8175.2011.01498.xC© 2011, Wiley Periodicals, Inc.

Tissue Doppler Derived Mechanical DyssynchronyDoes Not Change after Cardiac ResynchronizationTherapy

Rebecca Perry, B.Sc., D.M.U. (cardiac), A.M.S., Carmine G. De Pasquale, B.M.B.S., F.R.A.C.P., Ph.D.,Derek P. Chew, M.B.B.S., M.P.H., F.R.A.C.P., Andrew D. McGavigan, M.D., F.R.A.C.P.,and Majo X. Joseph, M.B.B.S., F.R.A.C.P.

Cardiac Services, Flinders Medical Centre, Bedford Park, South Australia, Australia

Background: Mechanical left ventricular (LV) dyssynchrony, as determined by tissue Doppler imaging(TDI), predicts response to cardiac resynchronization therapy (CRT). However, changes in TDI me-chanical dyssynchrony after CRT implantation have only limited investigation. Our objective was todetect changes in the extent and location of TDI mechanical dyssynchrony pre- and post-CRT, andto explore their relationship in response to CRT. Methods: Thirty-nine consecutive patients undergo-ing CRT implantation for chronic heart failure underwent TDI analysis pre-CRT and up to 12 monthspost-CRT. Regional dyssynchrony was determined by the time to systolic peak velocity of opposingLV walls. Dyssynchrony was defined as a difference in time to peak contraction of >105 msec. Twopatients were excluded, as suitable coronary venous access was not available. Results: Of the 37 pa-tients, 28 (76%) had significant mechanical dyssynchrony pre-CRT. Of those with dyssynchrony, 18(64%) had septal delay and 10 (36%) had LV free wall delay. Post-CRT, 29 (78%) patients had sig-nificant mechanical dyssynchrony, 17 (59%) with septal delay, and 12 (41%) with LV free wall delay.There was no difference in both the amount of dyssynchrony (P = 0.8) or the location of the dyssyn-chrony (P = 0.5), before and after CRT, even though 28 (76%) were considered responders basedon symptomatic and echocardiographic parameters. Conclusion: The TDI-derived dyssynchrony doesnot change with CRT despite significant symptomatic and echocardiographic improvement in cardiacfunction. The TDI is of limited utility for monitoring response to CRT. (Echocardiography 2011;28:961-967)

Key words: tissue Doppler echocardiography, cardiac pacing, congestive heart failure

Severe chronic heart failure (CHF) is associatedwith increased morbidity and mortality.1–3 Trialsof cardiac resynchronization therapy (CRT) havedemonstrated improvement in both, and it is nowa standard therapy for CHF.4–7 The majority ofthese trials have used the presence of a prolongedelectrocardiographic (ECG) QRS duration (≥120msec) as an electrical corollary of mechanicaldyssynchrony, and this has been shown to predictresponse in more than half of the CHF patients inCRT trials.4–7 However, this response rate is stillsuboptimal and represents the major limitationof using QRS duration for prediction of response.This has fueled interest in the development ofother modalities to select patients for CRT ther-

Conflicts of Interest: None.

Disclosure: None.

Address for correspondence and reprint requests: RebeccaPerry, B.Sc., D.M.U. (Cardiac), A.M.S., Cardiac Investiga-tions, Level 6, Flinders Medical Centre, Flinders Drive, Bed-ford Park, SA 5042, Australia. Fax: +61882044907; E-mail:[email protected]

apy. Tissue Doppler imaging (TDI), using variousmethods, has been shown to directly measure leftventricular (LV) mechanical dyssynchrony, and isuseful in the diagnosis and prediction of respon-ders.8–14 However, only limited data are availableon TDI-derived mechanical dyssynchrony post-CRT implantation, and whether changes in TDIdyssynchrony parallel response to CRT.15

Our objectives were to describe changes overtime in the extent and location of TDI mechanicaldyssynchrony in an unselected series of patientswho underwent CRT for standard heart failure in-dications, and to explore their relationship in re-sponse to CRT.

Methods:This study was approved by the Flinders MedicalCentre Research Ethics Committee.

Subjects:The subject population consisted of 39 consecu-tive CHF patients referred for CRT implantation.

961

Perry, et al.

All subjects had to have an ejection fraction (EF)<35%, QRS duration ≥120 msec, be on opti-mized medical therapy for heart failure, and haveNew York Heart Association (NYHA) class III or IVheart failure to be referred for CRT. Echocardio-graphy was performed before CRT implantationand every 3 months up to 12 months postim-plant. All studies were performed on a Philips iE33(Philips Medical Systems, Bothell, WA, USA). Pa-tients with an existing CRT device, a pacemakerin situ or in atrial fibrillation were excluded fromthis study.

Subjects were considered responders to theCRT if there was a decrease in NYHA class of atleast one level and an increase in EF of >10%.

Echocardiography:Biplane Simpson’s EF was calculated from the api-cal two- and four-chamber views. Diastolic func-tion was assessed using the early diastolic velocityfrom the mitral inflow (e) and the septal tissueDoppler profile (e’) to give a ratio of e:e’.16

During the echocardiography study, each api-cal view (two-chamber, four-chamber, and longaxis) was overlayed by the TDI color map, opti-mized to give the highest frame rate by reducingdepth and sector size, and then three or morecardiac cycles for each view were digitally stored.All images were analyzed offline using dedicatedsoftware (SQ application, Qlab 6.0 Philips Medi-cal Systems). Only TDI images, with a frame rateof greater than 90 cycles/sec, were considered foranalysis, in accordance with the American Societyof Echocardiography (ASE) recommendations forthe performance of dyssynchrony analysis.17

ECG Analysis:An ECG was taken at each study visit (poste-chocardiography guided CRT optimization forfollow-up visits) using a MAC 5000 ECG ma-chine (Marquette Medical Systems, Milwaukee,WI, USA) to determine the QRS duration. The QRSduration was determined electronically using soft-ware installed in the ECG machine and was con-firmed manually using calipers. The QRS durationwas measured from the start of the Q-wave (or ifno Q-wave present, from the start of the R-wave)to the end of the S-wave (the “J” point). The ECGdata was analyzed independently of the echocar-diography data.

Tissue Doppler Assessment:The timings of mitral and aortic valve openingand closure were determined using the pulsedwave (PW) spectral tracings from each of thesevalves; these timings were entered into the soft-ware package to be displayed on the graphs. TheTDI images were analyzed with the quantitativeanalysis package using tissue Doppler graphs. The

TDI color overlay was removed to allow easieridentification of the LV myocardium. For each api-cal view, the region of interest (or m-line measur-ing approximately 5 mm × 10 mm) was placedat a basal and midwall level in opposing wallsegments. These regions of interest were trackedwithin the myocardium over the cardiac cycle.

Tissue Doppler data was obtained from thebasal and mid level of the following walls: sep-tal, lateral, inferior, anterior, posterior, and an-teroseptal. The time between the start of the QRSand the peak positive systolic velocity (occurringbetween aortic valve opening and closure, labeledas s’) was measured (Fig. 1) giving 12 measuresof time to peak velocity. The time to peak ve-locities of opposing walls at the same level (i.e.,basal to basal and midwall to midwall) were sub-tracted from one another to determine the delaybetween the opposing walls giving six measure-ments of delay. These six measures of delay werethen averaged together to give an average levelof delay both pre- and post-CRT implantation.

Lead Placement:Device implantation was successful in all patientswithout major complications; however, of the 39patients, 2 were excluded, as the coronary ve-nous access was unavailable or no suitable tribu-tary identified, and the LV lead was instead posi-tioned in the high right ventricular outflow tractto produce multisite right ventricular pacing. Inthe remaining 37 patients, the LV pacing leadwas inserted transvenously to a tributary of thecoronary sinus dependent on venous anatomy.Lead position was determined from posteroante-rior and left lateral chest x-rays taken post-CRT.

Assessment of Dyssynchrony:We have previously reported our method of as-sessing significant dyssynchrony.18 In brief, a co-hort of 100 volunteers with normal LV systolicfunction and a normal QRS was studied to de-termine normal levels of dyssynchrony betweenopposing wall segments (basal to basal and mid-wall to midwall). Abnormal dyssynchrony was de-fined as greater than 95% confidence interval ofthe mean (105 msec). The dyssynchrony was onlyconsidered significant if it was above this abnor-mal level between two or more opposing wallsegments.

The most significant areas of delay were notedand categorized into either septal delay, or LV freewall delay, or no significant dyssynchrony.

Optimization of CRT Device:All subjects underwent CRT optimization pre-discharge and every 3 months, in an attemptto improve optimal functioning of the de-vice according to the ASE guidelines.17 The

962

Tissue Doppler Derived Mechanical Dyssynchrony

Figure 1. Tissue Doppler graph demonstrating method of detection of dyssynchrony at the basal level in the four-chamber view.Basal septal s’ is shown by the arrow labeled “A” and basal lateral s’ is shown by the arrow labeled “B.”

atrioventricular (AV) delay was optimized usingthe iterative method. The PW Doppler at the levelof the mitral annulus was performed to assess themitral E- and A-waves. The AV delay was set to along delay (e.g., 250 msec) and then checked toensure biventricular pacing capture. The AV delaywas then decreased in 20 msec until the mitral E-and A-waves have adequate separation, and theA-wave was not truncated. The ventricular to ven-tricular delay on the device was set to LV offset of–50 msec (LV lead 50 msec before the right ven-tricular [RV] lead), and then reduced by 10 msecincrements up to a LV offset of +50 msec (RV lead50 msec before LV lead) until the maximum aorticvelocity time integral was reached.

Statistical Analysis:Continuous variables are displayed as median(25th–75th percentiles) for nonparametric data,mean ± standard deviation for parametric data,and discrete variables as number and percentage.Statistical analysis was performed with SPSS forWindows 15.0 (SPSS Inc., Chicago, IL, USA). Thechange in EF, e:e’, and QRS duration was ana-lyzed over all the time points using a Friedmantest, with Bonferroni correction for repeated mea-sures. Specific changes at each time point were

analyzed using a Wilcoxon signed-rank test. Dif-ferences between responders and nonresponderswere analyzed using a Mann–Whitney test. Allvariables related to response to CRT with P < 0.1were included in a multivariate model. Statisticalsignificance was assumed if P < 0.05.

Results:Of the 37 patients, mean age was 66 ± 10 years,27 (73%) were males, and 26 (70%) had ahistory of ischemic heart disease. The averagetime of follow-up was 8.3 ± 5.6 months. Therewere 3 (8%) patients who had renal impairment,but none had a creatinine level greater than200 μmol/L (normal range 50–100 μmol/L). Pa-tient characteristics are shown in Table I.

Echocardiographic Parameters:Pre-CRT, the median EF was 24% (19.5–30), theLV end diastolic dimension was 6.4 cm (6.1–7.7),LV end systolic dimension was 5.9 cm (5.4–6.4),and e:e’ was 20 (13.2–27.7). Post-CRT, the me-dian EF increased to 37% (25–45; P < 0.001), theLV end diastolic dimension decreased to 5.6 cm(4.8–7.1; P < 0.001), the LV end systolic dimen-sion decreased to 5.2 cm (4.5–5.7; P < 0.001),

963

Perry, et al.

TABLE I

Baseline Patient Characteristics Pre-CRT Implantation

BaselineCharacteristic Value

Age (years) 66 ± 10Males 27 (73%)AF 7 (19%)IHD 26 (70%)Location of MI 23 (62%) LAD, 2 (5%) RCA, 1 (3%)

LCx, 11 (30%) non-IHDWall motion

abnormality10 (27%) anterior, 2 (5%) inferior, 1

(3%) lateral, 13 (35%) global, 11(30%) non-IHD

Area of delay 18 (49%) septal, 10 (27%) LV freewall, 9 (24%) none

NYHA class 3 (8%) class II, 26 (70%) class III, 8(22%) class IV

Renal impairment 3 (8%)Hypertension 3 (8%)Diabetes mellitus 4 (11%)Beta-blocker 33 (89%)ACE inhibitor 33 (89%)Statin 23 (62%)Loop diuretic 30 (81%)Amiodarone 12 (32%)Spironolactone 16 (43%)

BMI = body mass index; AF = atrial fibrillation; IHD = ischemicheart disease; MI = myocardial infarction; LAD = left ante-rior descending coronary artery; RCA = right coronary artery;LCx = left circumflex coronary artery; NYHA class = NewYork Heart Association class; ACE inhibitors = angiotensin-converting enzyme inhibitors.

and e:e’ trended to decrease to 16 (13.7–22.5;P = 0.17; Table II).

Level and Location of TDI Dyssynchrony:Significant TDI-derived mechanical dyssynchronywas seen in 28 (76%) patient’s pre-CRT. Of thosewith dyssynchrony, 18 (64%) had a significantseptal delay and 10 (36%) had a significant LV

TABLE II

Changes in QRS Duration, Ejection Fraction, e:e’, and LVEDdand Pre- and Postcardiac Resynchronization Therapy

Pre-CRT Post-CRT P Value

QRS duration(msec)

160 (136–174) 155 (133–179) 0.76

Ejectionfraction (%)

24 (19.5–30) 37 (25–45) <0.001

e:e’ 20 (13.2–27.7) 16 (13.6–22.5) 0.17LVEDd (cm) 6.4 (6.1–7.7) 5.6 (4.8–7.1) <0.001LVESd (cm) 5.9 (5.4–6.4) 5.2 (4.5–5.7) <0.001

CRT = cardiac resynchronization therapy; LVEDd = left ven-tricular end diastolic diameter; LVESd = left ventricular endsystolic diameter.

free wall delay. Seventeen of the 18 patients(94%) with septal delay were responders to CRT,9 of 10 (90%) with LV free wall delay were respon-ders, and 2 of 9 (22%) without baseline dyssyn-chrony were responders.

Post-CRT, 29 (78%) patients had significantmechanical dyssynchrony, 17 (59%) with sep-tal delay, and 12 (41%) with LV free wall delay.There was no difference in both the amount ofdyssynchrony (P = 0.8) or the location of thedyssynchrony (P = 0.5) before and after CRT,even though 28 (76%) patients were consideredresponders based on symptomatic and echocar-diographic parameters.

Lead Placement:The LV lead was positioned in the mid lateral veinin 24 (65%), in the basal lateral vein in 3 (8%),in the mid posterior vein in 7 (19%), the basalposterior vein in 1 (3%), and in the mid anteriorvein in 2 (5%) subjects. The LV lead was placedaccording to coronary anatomy and did not cor-relate with the region of greatest delay on TDIanalysis.

Responders versus Nonresponders:Our responder cohorts were older (68 ± 10 years)than the nonresponders (61± 7 years; P = 0.05),were more likely to have baseline dyssynchronyby TDI (93% vs. 22%; P = 0.03), and had adecrease in LV end systolic dimension with CRT(5.6 cm [4.9–6.6]–5.0 cm [4.5–5.7]; P < 0.001;Table III). Using multivariate analysis, only thepresence of dyssynchrony pre-CRT could predictresponse to CRT (Table IV). Both responders andnonresponders to CRT did not have any signifi-cant change in the level (P = 0.6 and 0.25, respec-tively) and location (P = 0.4 and 0.1, respectively)of TDI-derived mechanical dyssynchrony. Therewas no change in QRS duration pre- and post-CRT in the total cohort (160 msec [136–174] preand 155 msec [133–179] post; P = 0.8), and be-tween responders (158 msec [138–167] pre and160 msec [135–178] post; P = 0.9) and nonre-sponders (174 msec [120–177] pre and 154 msec[120–184] post; P = 0.4).

Discussion:Novel echocardiographic techniques, such as TDIdyssynchrony imaging, offer promise in guidingthe use of expensive and invasive device ther-apy in CHF. However, despite the confirmedvalue of TDI dyssynchrony in predicting clini-cal response to CRT in the CHF population, wefound no change in TDI dyssynchrony parameterspost-CRT.

The TDI dyssynchrony has become one of themore widely used methods for the assessmentof LV dyssynchrony, however, as the PROSPECT

964


TABLE III

Patient Characteristics of Responders versus Nonresponders

Responder (n = 28) Nonresponder (n = 9) P Value

Age 68 ± 10 61 ± 7 0.05Gender 20 (71%) males 7 (78%) males 0.81IHD 18 (64%) 8 (89%) 0.27QRS duration pre-CRT 158 (138–167) 174 (120–177) 0.79QRS duration post-CRT 160 (135–178) 154 (120–184) 0.62Level of delay pre-CRT 165 (92–238) 95 (51–140) 0.14Level of delay post-CRT 128 (55–273) 131 (61–170) 0.66Evidence of TDI dyssynchrony

pre-CRT26 (93%) 2 (22%) 0.03

Location of dyssynchronypre-CRT

17 (61%) septal9 (32%) LV free wall

2 (7%) none


7 (78%) none

0.01

Location of dyssynchronypost-CRT


6 (21%) none


2 (22%) none

0.89

EF pre-CRT (%) 25 (19.2–30.8) 20 (18.5–28.5) 0.63EF post-CRT (%) 45 (38.5–54) 29 (22.5–36) 0.01e:e’ pre-CRT 20 (13.7–26.7) 20 (10–28.7) 0.81e:e’ post-CRT 14.5 (12.7–15) 18 (16.5–27.5) 0.04

∗

LVEDd pre-CRT (cm) 6.4 (5.9–7.5) 6.9 (6.2–7.9) 0.39LVEDd post-CRT (cm) 5.9 (5.4–6.5) 6.1 (5.5–6.4) 0.63LVESd pre-CRT (cm) 5.6 (4.9–6.6) 6.3 (4.8–7.3) 0.48LVESd post-CRT (cm) 5.0 (4.5–5.7) 5.4 (4.6–5.9) 0.38

CRT = cardiac resynchronization therapy; IHD = ischemic heart disease; EF = ejection fraction; LVEDd = left ventricular enddiastolic diameter.∗Not significant using Bonferroni correction criteria.

study demonstrated,19 it requires a high levelof training and experience, before reproducibleresults can be obtained. Our experience with TDIis extensive and in this study the intra- and inter-operator variability correlation coefficients were0.82 and 0.76, respectively (P < 0.001 for both).With experienced operators the use of TDI to as-sess velocity timings has been validated in phan-toms,20 animal models,21,22 and human studies,23

making it a reliable tool for the assessment of re-gional LV myocardial dyssynchrony.

TABLE IV

Mulitvariate Analysis of Variables Related to Prediction ofResponse to CRT

MultivariateVariable Analysis (β) P Value

Age −0.89 0.14Presence of dyssynchrony pre-CRT −2.2 0.01Location of dyssynchrony pre-CRT −0.30 0.54EF post-CRT 1.51 0.12e:e’ post-CRT 0.51 0.43

CRT = cardiac resynchronization therapy; EF = ejection frac-tion.

Response to CRT:In the total subject cohort, there was a signifi-cant increase in EF, a decrease in LV end diastolicdimension, and a trend toward improvement inthe diastolic function, as reflected by a decreasein the e:e’ (Fig. 2). We also found that respondersto CRT had a significant decrease in their LV fillingpressures (as reflected by e:e’), an increase in theirEF, and a decrease in LV end diastolic dimension.Our responder rate is consistent with previouslypublished studies4–7 and it is, therefore, likely tobe representative of the population as a whole.

Assessment of Dyssynchrony:Our cutoff value of 105 msec between twoor more opposing wall segments for abnormaldyssynchrony, was defined as greater than 95%confidence interval of the mean from a nor-mal population.18 Other studies using TDI to as-sess dyssynchrony have used various cutoff val-ues ranging from 60 msec,12 65 msec,13 and upto 110 msec.14 These cutoff values were deter-mined from their CHF subject cohort retrospec-tively, according to the response to CRT. Applyingthese cutoff values in our subject cohort resultedin all our subjects having baseline dyssynchrony.In our experience, we have observed levels ofdelay between opposing walls of greater than

965

Perry, et al.

Figure 2. Graph demonstrating the change inejection fraction and e:e’ up to a 12-month pe-riod.

70–100 msec in the normal population, especiallywhen using only a six-segment model or a max-imal global delay. Defining significant dyssyn-chrony by first using a 12 segment model andby ensuring that at least two or more opposingwalls meet the dyssynchrony criteria, limits aber-rant results and may reduce confounding factors.Our cutoff value whilst being different seems tobe more robust in this subject cohort for predic-tion of response.Responders versus Nonresponders:Our definition of response was based on both,improvements in EF and symptoms, rather thanon only one of these, thereby giving a more accu-rate assessment of responder rate. To be classifiedas a responder, the patient had to improve by atleast one NYHA class and have an increase in theirEF of greater than 10%. Many studies have usedsymptoms or EF, rather than both, to determineresponse. If only one of these criteria were used,the apparent responder rate would be higher, butwe believe that the use of two criteria, one sub-jective and one objective, is a more precise def-inition of a responder and helps limit potentialbias from the study. Indeed, there were a smallnumber (three in total) of patients, within the co-hort, that showed a symptomatic improvementbut not a significant change in their EF, leavingthem to be labeled as “nonresponders” for thisstudy due to our strict criterion for response. In-terestingly, these patients were found to have animprovement in their filling pressures that mayaccount for their improvement symptomatically.

We found that TDI-derived mechanicaldyssynchrony was able to predict response toCRT (93% of responders had baseline levels ofdyssynchrony), although 22% of nonrespondersalso had baseline dyssynchrony. It is possible thatthe LV lead placement may not have been in themost delayed region, but without lead placementdata this is impossible to assess.

Both responders and nonresponders did nothave any significant change in the level and lo-cation of TDI-derived mechanical dyssynchrony

with CRT. This is in contrast to the study byBleeker et al.15 in which they found improvementin the level of TDI-derived dyssynchrony up to 6months post-CRT in responders to CRT only. Ourresponder cohort certainly had a trend toward areduction in their level of mechanical delay; how-ever, this did not reach statistical significance. Incontrast, the nonresponder cohort had a non-significant increase in their level of mechanicaldyssynchrony post-CRT.

The electrical dyssynchrony (as reflected byQRS duration) also did not change with CRT, in-dicating that electrical dyssynchrony has no cor-relation to improvement in symptoms and EF. Al-though, one might argue that the ability of QRSwidth to be a useful marker of dyssynchrony islost with ventricular pacing. These patients all un-derwent CRT optimization every 3 months to at-tempt to improve the Doppler-derived cardiacoutput before the TDI data was taken for thisstudy. Without this optimization, it is possible thatthe level of mechanical dyssynchrony may havebeen worse.

The TDI has good prediction power for re-sponse to CRT but does not seem to play a rolein the improvement of the responder cohort. Itis possible that the improvement in contractileefficiency may be from more synchronous con-traction in the radial planes (where most of theejection force comes from) rather than in the lon-gitudinal plane that is measured using the TDImethod. Theoretically, two-dimensional speckletracking may be the better modality to determinethis, although image quality is a major issue withthis novel technology.

Limitations:This is a small subject cohort; however, this studywas adequately powered to detect a 20% changein the average delay, if one was there. It has beenrecently demonstrated that other imaging modal-ities including three-dimensional echocardiogra-phy may be more suitable to the assessment of

966


dyssynchrony,24 however, this data was not avail-able in this subject cohort.

Some previous studies have used a reductionof LV end systolic volume (LVESV) of greater than15% as a criterion for response to CRT; however,in this study we have used EF. This is because inour hands EF, rather than LVESV, is a more reli-able and repeatable measurement that correlatesbetter with MRI. Statistically, all of our respondersdid have a LVESV reduction of at least 15%, how-ever, we have not added this to the manuscriptas this would potentially create more confusion inthe definition of response to CRT.

Our subject cohort had a high number of pa-tients with ischemic heart disease. This may haveconfounded the results to some degree; however,our response rate is high indicating that the pres-ence of scar tissue did not play a large role in thiscohort.

Conclusion:Our results certainly confirm the findings of pre-vious studies that TDI is useful for predicting CRTresponse. However, surprisingly the TDI methodof opposing wall contractile delay does not showa significant improvement in dyssynchrony post-CRT implantation. The reasons for this are notclear but suggest that it is too simplistic to as-sume that correcting mechanical dyssynchrony isthe sole mediator of benefit in CRT, and thereis a complex interplay of other factors that areinvolved in its beneficial effect.

References1. MacCarthy PA, Keamey MT, Nolan J, et al: Prognosis in

heart failure with preserved left ventricular systolic func-tion: Prospective cohort study. BMJ 2003;327:78–79.

2. Sun JP, Chinchoy E, Donal E, et al: Evaluation of ventric-ular synchrony using novel Doppler echocardiographicindices in patients with heart failure and receiving car-diac resynchronisation therapy. J Am Soc Echocardiogr2004;17:845–850.

3. Shamim W, Francis DP, Yousufuddin M, et al: Intraven-tricular conduction delay: A prognostic marker in chronicheart failure. Int J Cardiol 1999;70:171–198.

4. Abraham WT, Fisher WG, Smith AL, et al; for the MIRACLEstudy group: Cardiac resynchronisation in chronic heartfailure. N Engl J Med 2002;346:1845–1853.

5. Cazeau S, Leclercq C, Lavergne T, et al; for the MUS-TIC study investigators: Effects of multisite biventricularpacing in patients with heart failure and intraventricularconduction delay. N Engl J Med 2001;344:873–880.

6. Bristow MR, Saxon LA, Boehmer J, et al; for theCOMPANION study investigators: Cardiac resynchroni-sation therapy with or without an implantable defibril-lator in advanced chronic heart failure. N Engl J Med2004;350:2140–2150.

7. Cleland JG, Daubert JC, Erdmann E, et al; for the CARE-HFstudy investigators: The effect of cardiac resynchronisa-tion on morbidity and mortality in heart failure. N Engl JMed 2005;352:1539–1549.

8. Yu CM, Chau E, Sanderson JE, et al: Tissue Dopplerechocardiographic evidence of reverse remodelling andimproved synchronicity by simultaneously delaying re-

gional contraction after biventricular pacing therapy inheart failure. Circulation 2002;105:438–445.

9. Sogaard P, Egeblad H, Pedersen AK, et al: Sequential ver-sus simultaneous biventricular resynchronisation for se-vere heart failure: Evaluation by tissue Doppler imaging.Circulation 2002;106:2078–2084.

10. Breithardt OA, Stellbrink C, Kramer AP, et al: Echocar-diographic quantification of left ventricular asynchronypredicts an acute hemodynamic benefit of cardiac resyn-chronisation therapy. J Am Coll Cardiol 2002;40:536–545.

11. Cazeau S, Bordachar P, Jauvert G, et al: Echocardiographicmodelling of cardiac dyssynchrony before and duringmultisite stimulation: A prospective study. Pacing Clin Elec-trophysiol 2003;26:137–143.

12. Bax JJ, Marwick TH, Molhoek SG, et al: Left ventriculardyssynchrony predicts benefit of cardiac resynchronisa-tion therapy in patients with end-stage heart failure beforepacemaker implantation. Am J Cardiol 2003;92:1238–1240.

13. Gorcsan J, Kanzaki H, Bazaz R, et al: Usefulness of echocar-diographic tissue synchronisation imaging to predictacute response to cardiac resynchronisation therapy. AmJ Cardiol 2004;93:1178–1181.

14. Notabartolo D, Merlino JD, Smith AL, et al: Usefulness ofthe peak velocity difference by tissue Doppler imagingtechnique as an effective predictor of response to cardiacresynchronisation therapy. Am J Cardiol 2004;94:817–20.

15. Bleeker GB, Mollema SA, Holman ER, et al: Left ventric-ular resynchronization is mandatory for response to car-diac resynchronization therapy: Analysis in patients withechocardiographic evidence of left ventricular dyssyn-chrony at baseline. Circulation 2007;116:1440–1448.

16. Khouri SJ, Maly GT, Suh DD, et al: A practical approachto the echocardiographic evaluation of diastolic function.J Am Soc Echocardiogr 2004;17:290–297.

17. Gorcsan J, Abraham T, Agler DA, et al: Echocardiographyfor cardiac resynchronization therapy: Recommendationsfor performance and reporting–a report from the Amer-ican Society of Echocardiography Dyssynchrony WritingGroup Endorsed by the Heart Rhythm Society. J Am SocEchocardiogr 2008;21:191–213.

18. Perry R, De Pasquale CG, Chew DP, et al: QRS durationalone misses cardiac dyssynchrony in a substantial propor-tion of chronic heart failure patients. J Am Soc Echocardiogr2006;19:1257–1263.

19. Chung ES, Leon AR, Tavazzi L, et al: Results of the pre-dictors of response to CRT (PROSPECT) trial. Circulation2008;117(20): 2608–2616.

20. Fleming AD, McDicken WN, Sutherland GR, et al: As-sessment of colour Doppler tissue imaging using test-phantoms. Ultrasound Med Biol 1994;20: 937–951.

21. Gorcsan J, Strum DP, Mandarino WA, et al: Quantitativeassessment of alterations in regional left ventricular con-tractility with colour-coded tissue Doppler echocardiog-raphy. Comparison with sonomicrometry and pressure-volume relations. Circulation 1997;95:2423–2433.

22. Naito J, Masuyama T, Mano T, et al: Validation oftransthoracic myocardial ultrasonic tissue characteriza-tion: Comparison of transthoracic and open-chest mea-surements of integrated backscatter. Ultrasound Med Biol1995;21:33–40.

23. Rodriguez L, Garcia M, Ares M, et al: Assessment of mitralannular dynamics during diastole by Doppler tissue imag-ing: Comparison with mitral Doppler inflow in subjectswithout heart disease and in patients with left ventricularhypertrophy. Am Heart J 1996;131:982–987.

24. Kleijn SA, van Dijk J, de Cock CC, et al: Assessment ofintraventricular mechanical dyssynchrony and predictionof response to cardiacresynchronization therapy: Com-parison between tissue Doppler imaging and real-timethree-dimensional echocardiography. J Am Soc Echocar-diogr 2009;22:1047–1054.

967

tissue doppler derived mechanical dyssynchrony does not change after cardiac resynchronization...

Documents