Predictive value of left atrial remodeling for response to cardiac resynchronization therapy

This was a prospective, non-randomized, observational, single-center study. The study was conducted following the Good Clinical Practice guidelines of the Declaration of Helsinki and was approved by the Regional Ethical Committee. All subjects gave written informed consent.

The study population consisted of patients with systolic heart failure who were referred for cardiac resynchronization therapy (CRT) to the Catharina Hospital (Eindhoven, the Netherlands) between January 2012 and December 2014⁽⁷⁾. All patients received optimized medical therapy before CRT implantation. Patients with insufficient image quality and/or atrial fibrillation were excluded. Functional status was assessed by the estimation of the New York Heart Association (NYHA) functional class.

Standard techniques were used to implant the CRT device. Three transvenous pacing leads were inserted: one in right atrium, another in the high interventricular septum or right ventricular apex, and the coronary sinus lead was positioned near the LV free wall through a coronary sinus tributary vein. The pacing leads were connected to a dual-chamber biventricular device with programmable interventricular delay. The device was programmed in the DDD mode and the atrio-ventricular (AV) delay was adjusted during simultaneous pacing.

Echocardiograms were obtained before and 3 months after CRT implantation and analyzed by two experienced readers. For the recording, an iE33 ultrasound scanner equipped with a S1-1 transducer (Philips Healthcare, Andover, MA, USA) was used. Standard 2D- and Doppler-echocardiographic measurements were performed following ASE/EACVI guidelines, including both LV end-systolic and end-diastolic volume (LVESV and LVEDV)⁽⁸⁾. LV ejection fraction (LVEF) was calculated using the modified biplane Simpson’s rule. Maximum LA volume indexed to body surface area (LAVI) was calculated by the biplane method of disks at end-systole. Commercially available software (QLAB 13, Philips Healthcare, Eindhoven, the Netherlands) was used to measure LA reservoir strain (LASr) on an apical four- and two-chamber views of the LA with a stable ECG recording and frame rated >60 frames per second^(9,10). The LA endocardial border was automatically drawn followed by manual adjustment if required. The zero strain reference was set at end-diastole in order to determine LASr⁽⁹⁾.

The objective of this study was to identify predictors of LV reverse remodeling 3 months after CRT delivery. LV reverse remodeling was defined as a relative reduction in LVESV of ≥15%, in concordance with previous trials^(11,12).

Both LA and LV baseline parameters were tested as potential predictors of LV reverse remodeling after CRT⁽¹³⁾, together with the following baseline parameters: gender, HF etiology, presence of typical LBBB pattern, and QRS duration⁽¹³⁾. For LA structural remodeling and LA functional remodeling, we selected baseline LA volume index (LAVI) and baseline LA reservoir strain (LASr), respectively. For LV remodeling baseline LVEDV and LVEF were used.

Data for continuous variables are presented as mean ± standard deviation (SD) or as median with interquartile range (IQR) in the case of skewed distribution. Categorical variables are presented as numbers (percentages). Mann-Whitney U test was used to compare medians between survivors and non-survivors. The paired t-test or Wilcoxon signed rank test was used to evaluate differences in continuous variables between baseline and 3-month follow-up. A two-sided p-value of less than 0.05 was considered to indicate statistical significance. The significance of predictors for the occurrence of CRT-induced LV response was evaluated using the univariate and multivariate logistic regression model. Backward stepwise selection based on the likelihood ratio was used in the multivariate analysis to reduce the number of predictors. Multicollinearity was detected using variance inflation factors. Odds ratios (ORs) with 95% confidence intervals (95 CI) were used to quantify the effect of the individual predictors. ORs not including 1 were considered significant. All analyses were performed using SPSS version 25 (IBM Corporation, Armonk, New York).

The study population consisted of 121 consecutive patients with systolic heart failure accepted for CRT. Nineteen patients were excluded because of atrial fibrillation (n = 15) or insufficient imaging quality (n = 4). Two patients died before CRT implantation and one patient died before the 3-month follow-up echocardiography. The final study population consisted of 99 patients in whom LV reverse remodeling could be determined (Fig. 1).

Fig. 1.

Table 1 shows patient characteristics of the study population. Heart failure was of ischemic origin in 51 (52%) patients. Patients were predominantly male (n = 63, 64%) with a median age of 70 (65–76) years. Main comorbidities were hypertension (n = 32, 32%) and diabetes (n = 23, 23%). The intraventricular conduction delay was most commonly based on a left bundle branch block (n = 72, 73%) with a mean QRS duration of 166 ± 22 ms. More than half of the patients (63%) were reported as NYHA class III at baseline. At baseline, median LVESV was 87 (71–106) mL/m², LVEF 28 ± 8%, and LAVI 40 (31–52) mL/m². The majority of patients were treated with beta blockers (86%), renin-angiotensin system antagonists (90%), and to a lesser extent with mineralocorticoid antagonist (36%) and diuretics (60%).

At 3 months, 63 patients (64%) were considered as responders to CRT with a relative reduction in LVESV of ≥15%. The LV volumetric responders showed also a significant improvement in LVEF (28 to 42%; p <0.01). No significant change in LVEDV and LVEF was observed in the group of LV volumetric non-responders (Δ LVESV <15%).

The prevalence of LBBB and QRS duration >150 milliseconds were significantly higher in LV responders compared to the non-LV responders (Tab. 1). LAVI, as a measure of baseline LA structural remodeling, was significantly higher in LV non-responders (45 versus 36 mL/m²; p <0.01). LASr at baseline was significantly lower (12 versus 18%; p <0.01). Baseline LVEF was comparable (p = 0.32).

Both LAVI and LAS_r were associated with LV reverse remodeling in unadjusted analysis. In multivariate analysis, LAVI was the only echocardiographic variable to remain significant (Tab. 2). Figure 2 shows the proportion of LV responders in the study population subdivided by tertiles of LAVI. Small LA size at baseline (LAVI <34 mL/m²) is associated with a high proportion of CRT responders of 81% in contrast to severely dilated LA at baseline (LAVI >48 mL/m²) with only 48% response rate. Figure 3 shows the correlation between relative change of LVESV and baseline LAVI (Panel A) and LASr (Panel B).

Fig. 2.

Fig. 3.

This study covered a relatively small number of patients in a single center. Therefore, the results need validation in larger external cohorts. Secondly, follow-up measurements of LVESV at 6 and 12 months were not performed. However, most LV reverse remodeling will occur within 3 months and will adequately identify responders and non-responders to CRT⁽¹⁷⁾. Thirdly, patients with atrial fibrillation were excluded from this study, and the predictive value of LAVI in this specific group of HF patients deserves further investigation. Intervendor variability of LASr quantification remains an issue and can be a possible explanation for the difference in the outcomes with previous studies.