OUP user menu

Baseline echocardiographic characteristics of heart failure patients enrolled in a large European multicentre trial (CArdiac REsynchronisation Heart Failure study)

Stefano Ghio, Nick Freemantle, Alessandra Serio, Giulia Magrini, Laura Scelsi, Michele Pasotti, John G.F. Cleland, Luigi Tavazzi
DOI: http://dx.doi.org/10.1016/j.euje.2005.10.006 373-378 First published online: 1 October 2006


Aims Information on the prevalence and clinical, electrocardiographic and echocardiographic inter-relationships of mechanical dyssynchrony among patients with heart failure (HF) and left ventricular systolic dysfunction derives mainly from relatively small studies. The CARE-HF trial provides the opportunity to address these issues in a large population of patients with advanced HF.

Methods and results The CARE-HF trial enrolled patients with New York Heart Association (NYHA) class III or IV HF, with a QRS duration ≥120ms, left ventricular (LV) ejection fraction (EF) ≤35% and LV end diastolic diameter ≥30mm/m (height in m). Patients underwent a thorough echocardiographic evaluation, which included assessment of LV structure, systolic function, mitral inflow pattern, right ventricular (RV) dimensions and function, and interventricular mechanical delay (IVMD) as an index of interventricular dyssynchrony. Echocardiographic measurements were made in a Core Laboratory to ensure consistent quantitative analysis.

Of the 813 patients enrolled, 735 had a baseline echocardiographic examination suitable for measurement. Overall patients had advanced LV dysfunction (mean EF 25.5%) but few had a restrictive mitral filling pattern (18%) and both the mean RV diameter and RV function were within normal limits. Interventricular dyssynchrony defined as IVMD >40ms was present in 455 patients (62%). Clinical, electrocardiographic and standard echocardiographic variables were only loosely associated with IVMD.

Conclusions Interventricular dyssynchrony appears to be an independent characteristic of patients with advanced HF, and is poorly related to clinical, electrocardiographic or standard echocardiographic variable.

  • Resynchronisation
  • Heart failure
  • Echocardiography


In patients with advanced heart failure (HF) and left bundle branch block, biventricular pacing improves cardiac function and clinical status.1,2 The rationale for biventricular stimulation is resynchronisation of mechanical dyssynchrony secondary to regional delays in electrical conduction.3 However, QRS duration is known to be a relatively crude marker of ventricular dyssynchrony. Information on the prevalence of mechanical dyssynchrony among HF patients with prolonged QRS duration is sparse and the clinical, electrocardiographic and echocardiographic correlates of ventricular dyssynchrony were derived mainly from relatively small studies.4–7 The CArdiac REsynchronisation Heart Failure (CARE-HF) trial provides the opportunity to address this issue in a large, representative population of patients with advanced HF.

This paper describes the baseline echocardiographic characteristics of patients enrolled in the CARE-HF trial. The extent to which interventricular dyssynchrony is associated to QRS duration, the degree of left ventricular (LV) dysfunction and the aetiology of the disease (ischaemic vs non-ischaemic) is evaluated.



The CARE-HF trial enrolled patients with New York Heart Association (NYHA) class III or IV HF (despite undergoing optimal medical treatment), with a QRS duration ≥120ms, LV ejection fraction (EF) ≤35%, LV end diastolic diameter ≥30mm/m (height in m)8 and in sinus rhythm. Patients with a QRS duration ≥150ms did not require echo confirmation of dyssynchrony. Patients with a QRS duration between 120 and 149ms required at least two of the following dyssynchrony criteria: (a) aortic pre-ejection delay >140ms; (b) interventricular mechanical delay >40ms; and (c) delayed activation of the posterolateral wall of the left ventricle as shown by any overlap between time from QRS to peak systolic movement of the posterior wall at M-mode and time from QRS to the beginning of the E-wave at pulsed Doppler transmitral flow.

Of the 813 patients enrolled in the trial, 735 had an analysable baseline echocardiographic examination and constitute the population described. The population included a subgroup of 92 patients with a QRS duration of between 120 and 149ms measured in a Core Electrocardiographic Laboratory. Most of these patients (44/62) were enrolled after demonstrating dyssynchrony criteria a and b (described above).

Echocardiographic qualification

To participate in the CARE-HF trial, each centre had to go through a qualification procedure regarding accuracy and reproducibility of echocardiographic recordings. Meetings were held before the procedure to train the echocardiographers and a manual was prepared describing the echo protocol in detail. The trials were conducted in accordance with the Declaration of Helsinki.

Approval of a centre for inclusion in the CARE-HF trial was based on the assessment of videotapes submitted to the Echocardiographic Core Laboratory in Pavia, Italy. These included assessments of two patients, with one assessment being repeated within 24h to assess reproducibility of results. In the event of an initial disqualification, the site was given at least one more opportunity to submit further recordings and/or readings for approval; 98% of the centres qualified at the first attempt.

Recording and analysis of echocardiographic examinations

Echocardiographic recordings were obtained with the use of commercially available instruments. In each patient, multiple views were recorded on Super VHS videotapes. All echocardiograms were sent to the Core Echo Laboratory to ensure consistent measurement methodology. Videotaped echocardiograms were analysed with specialized software for image processing. This enabled the image to be captured, digitalized, analysed and stored with the numerical data. Each variable was measured three times, and the average value calculated. All readings were made by qualified technicians and subsequently reviewed by a senior echocardiographer.

The recorded views and the variables measured or calculated are outlined in Table 1. The end-diastolic (EDV) and end-systolic (ESV) volumes of the LV reported are those obtained using the Single Plane Area-Length method; EF was calculated as follows: (EDV−ESV)/EDV×100. The degree of mitral regurgitation (MR) was assessed as the area of the colour flow Doppler regurgitant jet divided by the area of the left atrium in systole. Pulsed Doppler velocity signals of transmitral flow were recorded at 100mm/s with the sample volume at the tips of the mitral valve leaflets. Peak velocities were measured during rapid LV filling (E-wave) and atrial contraction (A-wave), and the velocity ratio (E/A) was calculated. The deceleration time (DT) of the E-wave and the LV filling time were also measured; a DT <115ms was considered indicative of a restrictive LV filling pattern. The trans-tricuspidal pressure gradient was obtained whenever feasible. The tricuspid annular plane systolic excursion (TAPSE) was used as an indicator of right ventricular (RV) function. Stroke volume was estimated as the product of the LV outflow tract velocity time integral and the cross-sectional area of the LV outflow tract. Pulsed-wave Doppler flow velocity signals were recorded from the RV and LV outflow tracts, and the aortic and pulmonary pre-ejection times were measured as the time intervals between the Q-wave on the surface ECG and the onset of flow. The interventricular mechanical delay (IVMD) was calculated as the time difference between the onset of forward flow in the LV (APET) and RV (PPET) outflow tracts: IVMD=APET−PPET. An IVMD >40ms was considered indicative of interventricular dyssynchrony.

View this table:
Table 1

Views recorded and variables to be measured or calculated

ViewsVariables measured
M-mode of the left ventricleLVEDD, LVESD, RVEDD, EDPWT
Parasternal long axis of the left ventricleLVOTD
Parasternal short axis at mitral valve level
Parasternal short axis at mid-papillary level
Apical 4-chamber viewLVEDV, LVESV, LVEF according to Area-Length method
Apical 2-chamber viewLVEDV, LVESV, LVEF according to Simpson rule
Apical 4-chamber view modified to improve left atrium visualizationLAEDA, LAESA, LAFAC
Apical 4-chamber view with colour Doppler visualization of mitral regurgitationMR jet area
Mitral flow velocity curve at pulsed DopplerE/A ratio, DT, LVFT
Mitral flow velocity curve at continuous DopplerPresence or absence of EDMR, MRT
Aortic flow velocity curve at pulsed Doppler in 5-chamber viewAPET, TVI of aortic flow
M-mode recording of the lateral tricuspid annulusTAPSE
Apical 4-chamber view with colour Doppler visualization of the tricuspid regurgitationTR jet area
Tricuspid flow velocity curve at continuous DopplerPresence or absence of EDTR, TRT, TTPG
Pulmonary flow velocity curve at pulsed DopplerPPET
  • Abbreviations: APET=aortic pre-ejection time; DT=E-wave deceleration time; EDMR=end diastolic mitral regurgitation; EDPWT=end diastolic posterior wall thickness; EDTR=end diastolic tricuspid regurgitation; LAEDA=left atrial end-diastolic area; LAESA=left atrial end-systolic area; LAFAC=left atrial fractional area change; LVEDD=left ventricle end-diastolic diameter; LVESD=left ventricle end-systolic diameter; LVFT=left ventricular filling time; LVOTD=left ventricular outflow tract diameter; MR=mitral regurgitation; MRT=mitral regurgitation time; PPET=pulmonary pre-ejection time; RVEDD=right ventricular end-diastolic diameter; TAPSE=tricuspidal annular plane systolic excursion; TR=tricuspid regurgitation; TRT=tricuspid regurgitation time; TVI=time velocity integral; TTPG=trans-tricuspid pressure gradient.

Measurement reproducibility in the Core Laboratory

Key echocardiographic variables were measured twice (at intervals of >1 week) in 25 patients with HF by two echo technicians, and adequate echocardiographic images were used to assess inter- and intra-observer variability. Agreement of measurements was calculated according to Bland and Altman.9 As shown in Tables 2 and 3, reproducibility was optimal even for a derived parameter such as EF.

View this table:
Table 2

Intra-observer variability in the Core Laboratory

AverageSD95% limits of agreement
View this table:
Table 3

Inter-observer variability in the Core Laboratory

AverageSD95% limits of agreement

Electrocardiograms were recorded at 50mm/s in the majority of patients. Electrocardiographic intervals, including the QRS, were measured using hand-held calipers by two trained cardiology technicians.

Statistical analysis

Descriptive statistics were used to analyse the study population. A mixed model was developed in SAS statistical software (SAS ver. 9.1, SAS Institute Inc, Cary, NC) to account for the random effects associated with investigator sites. IVMD was used as the key index of dyssynchrony. APET is also an indicator of dyssynchrony, but is used to derive IVMD with which it is also co-linear. The linearity of potentially continuous predictors was examined by comparing the model fit in a stepwise manner with either the untransformed predictor, the log transformed predictor, or a restricted cubic spline with five knots. Proceeding to the next step only occurred once a statistically significant improvement in the model fit was achieved.10 The multivariate model to predict IVMD was run both in the total population and following exclusion of the subgroup of patients with QRS duration below 150ms (who were enrolled in the trial following echocardiographic demonstration of dyssynchrony [criteria specified earlier]). As the results of the analysis were similar, only data concerning the total population are presented.


Eight hundred and three baseline tapes of echocardiographic images were received by the Core Laboratory; 68 (9%) were excluded from further analysis because of poor quality. Therefore, the study population consisted of 735 patients with a mean age of 65 years (IQR 59–72); 73% were male, 46% had ischaemic aetiology and QRS duration was 168ms (IQR 155–180). There were 92 patients with a QRS duration <150ms. Demographic and clinical characteristics were similar to those of the whole CARE-HF population.11 Table 4 shows the baseline echocardiographic characteristics for the whole population. Of the 735 patients, 455 (62%) presented with interventricular dyssynchrony defined as IVMD >40ms; QRS duration was 171ms (IQR 160–180) in patients with IVMD >40ms and 164ms (IQR 152–174) in those with IVMD ≤40ms (p<0.0001). Table 4 shows the echocardiographic characteristics of the patients with (n=455) or without (n=280) interventricular dyssynchrony. Overall, the two groups were similar in terms of LV and RV function. The clinical, electrocardiographic and echocardiographic variables were then fitted in a multivariate model to identify factors independently associated with the degree of interventricular dyssynchrony (Table 5). Heart failure cause, (log) QRS and mitral valve E/A ratio were all strongly associated with interventricular dyssynchrony, and (log) EF moderately so. However, the model provided only poor predictive information at the individual patient level, with a corrected R value of 0.11. Fig. 1 shows the uncorrected relationship between IVMD and QRS duration.

Figure 1

Relationship between QRS duration and interventricular mechanical delay. The figure identifies the univariate relationship between increasing interventricular mechanical delay and QRS, which is plotted here on a log scale. The relationships between response variables including QRS and interventricular mechanical delay are described in Table 4.

View this table:
Table 4

Baseline echocardiographic characteristics by degree of interventricular dyssynchrony

Characteristic mean (IQR)All patients (n=735)IVMD ≤40ms (n=280)IVMD >40ms (n=455)p
LVEDD (cm)7.3 (6.5–8.0)7.2 (6.6–7.9)7.4 (6.4–8.2)0.4161
LVEDV (cm3)320 (244–381)309 (243–365)326.79 (245–388)0.0424
LVESV (cm3)242 (177–291)230 (176–270)249 (178–302)0.0121
LVEF (%)25.5 (21.3–28.9)26.2 (22.4–29.6)25.0 (20.8–28.6)0.0139
Detectable MR (n)532197335
MR index24 (11–34)24 (12–33)24 (11–34)0.8337
Mitral valve E/A ratio1.3 (0.6–1.5)1.5 (0.7–2.0)1.1 (0.6–1.4)<0.0001
Mitral valve192 (140–227)181(135–216)200 (146–234)0.0891
DT (ms)193 (142–227)180 (136–207)200 (147–135)0.0972
Restrictive LV filling pattern (n)5120311.00
LVFT (%)0.46 (0.39–0.52)0.47 (0.41–0.54)0.45 (0.38–0.51)0.0008
RVEDD (cm)2.5 (2.1–3.0)2.5 (2.2–2.9)2.5 (2.0–3.0)0.8769
TAPSE (cm)1.9 (1.6–2.2)1.9 (1.6–2.2)1.9 (1.6–2.2)0.3825
APET (ms)160 (134–187)139.2 (117.6–160)173.5 (150.0–198.5)<0.0001
Detectable TR (n)354126228
Detectable TTPG (n)328109219
TTPG (mmHg)31.8 (22.0–38.8)33.0 (22.9–38.9)31.3 (21.7–38.6)0.1045
PPET (ms)113 (90–133)120 (98–140)108 (86–129)<0.0001
Interventricular delay (ms)47.9 (29.3–66.4)
  • Abbreviations as in Table 1.

View this table:
Table 5

Multivariate model to predict interventricular dyssynchrony

VariableEffectLower 95% CIUpper 95% CIp value
Non-ischaemic aetiology9.995.3814.60<0.0001
Log QRS42.9223.8661.98<0.0001
Log ejection fraction−12.21−21.74−2.680.0122
Mitral valve E/A ratio−5.04−7.35−2.74<0.0001
  • Final multivariate model results examining predictive characteristics of interventricular dyssynchrony. ‘Log’ denotes log transformed variable. Investigational site (centre) included as random effects.


Baseline echocardiographic characteristics of the patients

The baseline echocardiographic data indicate that patients enrolled in the CARE-HF trial had advanced LV dysfunction similar to patients included in the MIRACLE trial.12 However, the low percentage of patients with a restrictive mitral filling pattern, the minimal increase in trans-tricuspid pressure gradient, the normal mean RV diameter and normal TAPSE all suggest that few patients had end-stage disease characterized by pulmonary hypertension and RV dysfunction. Indeed, few patients in the trial were classified as NYHA functional class IV.11

As has been observed in some previous smaller studies more than one-third of CARE-HF patients did not show interventricular dyssynchrony (280 out of 735 patients had an IVMD ≤40ms) despite marked QRS prolongation.6,7 In this respect, the CARE-HF trial is ideally placed to determine whether the echocardiographic evaluation of interventricular dyssynchrony predicts the response to CRT and whether it is superior to QRS duration. A limitation of the study is the lack of data on intraventricular dyssynchrony but a three dimensional tissue Doppler imaging substudy has been conducted which may offer a new approach to the problem of predicting response to CRT.

Electrocardiographic and echocardiographic correlates of interventricular dyssynchrony

There was a linear relationship between the predictor variables and the outcome of interest. Although the interventricular delay was significantly related to QRS duration, the correlation observed was too poor to be clinically helpful and was worse than in previously reported populations.6,7 A reason for this may be that the CARE-HF trial recruited mostly patients with a markedly prolonged QRS duration (>150ms) and those with intermediate QRS duration (of between 120 and 150ms) were selected on the basis of echocardiographic evidence of dyssynchrony. In contrast, previous studies have enrolled patients with a wide range of QRS durations, including some with normal intraventricular conduction. It was hypothesized that other baseline characteristics of the patients could be used to identify those with interventricular dyssynchrony in addition to QRS duration. Non-ischaemic aetiology, LVEF, the mitral valve E/A ratio as well as QRS duration were predictive variables included in the final model. However, these variables were of limited practical value in identifying patients with IVMD >40ms. Therefore, interventricular dyssynchrony in patients with HF may be considered an independent characteristic which is poorly related to any known clinical, electrocardiographic or echocardiographic variable. Indeed, classical left bundle branch block on the surface ECG may be due to diverse types of infranodal conduction abnormalities. The site of the delay in the conduction system (or ventricular myocardium) could be the most important factor in determining the degree of mechanical dyssynchrony.13


Patients enrolled in the CARE-HF trial were characterized by advanced LV dysfunction without echocardiographic evidence of pulmonary hypertension and RV dysfunction. In this population, the electrocardiographic, clinical and standard echocardiographic variables were of limited value in identifying those patients showing interventricular dyssynchrony (approximately two-thirds in total). It is concluded that, at present, a targeted echocardiographic evaluation is required to identify the presence of interventricular dyssynchrony in patients with advanced HF.


The authors would like to acknowledge the assistance and work of technicians C. Bassi, V. Pierota and E. Tellaroli at the Echo Core Laboratory, and support from Dr Berthold Stegemann and the Care-HF Study Management Team at Medtronic.


View Abstract