Electrocardiography

Before CRT implantation, standard 12-lead ECGs were recorded at 25 mm/s paper speed and calibrated at 1.0 mV/cm before recording of the echocardiogram. The method used for ECG analysis has been reported . Heart rate, P wave duration, PR interval, _e PR interval (end of P wave to onset of QRS), QRS duration, QT interval, JT interval (end of QRS to onset of T wave) and RR cycle minus QT interval (as a measure of electrical diastole) were measured, and the QRS morphology and axis were analysed. The frontal plane QRS axis was considered normal when between −30° and +90°, left deviated when beyond −30°, and right deviated when beyond +90°. LBBB was defined as a QRS duration ≥ 120 ms, with a broad R wave in leads I, aV _L , V ₅ and V ₆ and an R peak time > 60 ms in leads V ₅ and V ₆ , according to the practice guidelines issued by major professional societies . Other intraventricular conduction disturbances were classified as right bundle branch block or non-specific intraventricular conduction delays. The intra- and interobserver reproducibility of the measurements were ascertained by comparing the analysis of 20 randomly selected ECGs by two experts unaware of each other’s interpretation.

Transthoracic echocardiography

All patients underwent resting two-dimensional Doppler and speckle-tracking transthoracic echocardiography before CRT device implantation, using Vivid 7 or Vivid E9 ultrasound instrumentation (General Electric Medical Systems, Horten, Norway), according to a standardized protocol for image acquisition. LV end-systolic and end-diastolic diameters were measured in the parasternal long-axis view with M-mode; LV volumes indexed to body surface area and LVEF were measured in apical four- and two-chamber views using the biplane Simpson’s method .

The septal and lateral mitral annular peak systolic velocities were measured from the apical four-chamber view using tissue Doppler imaging. LV strain was analysed by speckle-tracking echocardiography using the four-chamber and mid-LV short-axis views. Images were acquired at end-expiration and analysed off line, using the dedicated automated imaging function of the EchoPAC BT12 ^® software package (GE Healthcare, Chalfont St Giles, UK). We used the echocardiographic indices of mechanical dyssynchrony previously published by Gorcsan et al. , and measured according to the recommendations of the American Society of Echocardiography and the Heart Rhythm Society.

Definitions

The diastolic filling time (DFT)/RR interval ratio was used to characterize atrioventricular (AV) dyssynchrony in the left heart. AV dyssynchrony was defined as DFT/RR < 40% .

Interventricular mechanical dyssynchrony (IVMD), calculated as the time difference between right ventricular (RV) and LV ejection at the onset of pulsed Doppler flow velocities in the LV and RV outflow tracts, respectively, was used to characterize interventricular dyssynchrony. Interventricular dyssynchrony was defined as IVMD > 40 ms .

LV pre-ejection interval (LVPEI), the delay between the onset of QRS and the beginning of LV ejection flow by Doppler imaging, was used to characterize left intraventricular dyssynchrony. Intraventricular dyssynchrony was defined as LVPEI > 140 ms .

Intraventricular longitudinal dyssynchrony was defined as the maximum delay between opposing septum-to-posterior wall, in colour-coded tissue Doppler imaging in the apical long-axis view, or by the maximum delay between opposing septum-to-posterior wall in speckle-tracking longitudinal strain imaging, in the apical long-axis view .

Intraventricular radial dyssynchrony was defined as the delay between opposing anteroseptal-to-posterior wall in the mid-LV short-axis view in speckle-tracking radial strain imaging .

The intrinsic intra- and interobserver reproducibilities of intraventricular dyssynchrony measured by longitudinal strain were ascertained from corresponding repeated measurements, using intraclass correlations. The intra- and interobserver reproducibilities were evaluated by a second measurement of 20 randomly selected transthoracic echocardiograms. The intra- and interobserver reproducibilities of other echocardiography measurements are already known .

Statistical analysis

A descriptive analysis of pertinent patient characteristics is expressed as means ± standard deviations or counts and percentages, unless specified otherwise. Correlations between indices of mechanical dyssynchrony and electrocardiographic measurements used Spearman’s correlation coefficients (and 95% confidence intervals [CIs] through Fisher’s Z transformation) and linear regression estimates. ‘Good’ correlation was considered when the r coefficient was ≥ 0.50. Categorical variables were compared between groups using Fisher’s exact test, while continuous variables were compared using the non-parametric Kruskall–Wallis test. A multivariable regression analysis included covariates emerging at a P ≤ 0.05 statistical level in the univariate analysis; we estimated a squared semi-partial coefficient, which represents the proportion of variance in y that is explained by x ₁ only. Functional forms of continuous covariates were assessed graphically and statistically, using non-parametric regression and the PROC GAM smoothing technique (SAS Institute, Cary, NC, USA). A two-sided P value < 0.05 was considered statistically significant. The data were analysed with the SAS ^® software package, version 9.3 (SAS Institute, Cary, NC, USA).