Characteristics of Responders to Cardiac Resynchronization Therapy: The Impact of Echocardiographic Left Ventricular Volume

Abstract

Background:

One‐third of patients who receive cardiac resynchronization therapy (CRT) are classified as nonresponders. Characteristics of responders to CRT have been studied in multiple clinical trials.

Hypothesis:

Independent predictors of CRT response may be identified by studying a series of patients in routine clinical practice.

Method:

One hundred twenty‐five patients were examined retrospectively from a multidisciplinary CRT clinic program. Echocardiographic CRT response was defined as a decrease in left ventricular (LV) end‐systolic volume of ≥15% and/or absolute increase of 5% in LV ejection fraction at the 6‐month visit.

Results:

There were 81 responders and 44 nonresponders. By univariate analyses, female sex, nonischemic cardiomyopathy etiology, baseline QRS duration, the presence of left bundle branch block (LBBB), and left ventricular end‐diastolic volume (LVEDV) index predicted CRT response. However, multivariate analysis demonstrated that only QRS duration, LBBB, and LVEDV index were independent predictors (QRS width, odds ratio [OR]: 1.027, 95% confidence interval [CI]: 1.004–1.050, P = 0.023; LBBB, OR: 3.568, 95% CI: 1.284–9.910, P = 0.015; LVEDV index, OR: 0.970, 95% CI: 0.953–0.987, P = 0.001). Although female sex and nonischemic etiology were associated with an improved CRT response on univariate analyses, after adjusting for LV volumes they were not independent predictors.

Conclusions:

QRS width, LBBB, and LVEDV index are independent predictors for echocardiographic CRT response. Previously reported differences in CRT response for sex and cardiomyopathy etiology are associated with differences in baseline LV volumes in our clinical practice.

Dr. Heist has received research grants (modest) from Biotronik, Boston Scientific, and St. Jude Medical; honoraria (modest) from Biotronik, Boston Scientific, Medtronic, Sorin, and St. Jude Medical; and consultant/advisory board positions (modest) from Boston Scientific, Sorin and St. Jude Medical. Dr. Singh has received research grants (significant) from Biotronik, Boston Scientific, Medtronic, and St. Jude Medical; and consultant/advisory board positions (modest) from Biosense Webster, Biotronik, Boston Scientific, CardioInsight, Medtronic, Sorin, St. Jude Medical, and Thoratec Inc.

The statistical analysis was conducted with support from Harvard Catalyst. The Harvard Clinical and Translational Science Center (National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health Award UL1 RR 025758, and financial contributions from Harvard University and its affiliated academic healthcare centers).

The authors have no other funding, financial relationships, or conflicts of interest to disclose.

Introduction

Despite efforts to identify predictors for cardiac resynchronization therapy (CRT) and thus optimize the selection of candidates for treatment, the current selection criteria for CRT have not changed.1 Although many echocardiographic parameters including mechanical dyssynchrony have been studied, only left ventricular (LV) ejection fraction (EF) remains in the current criteria.1, 2, 3, 4, 5 Although LV volumes are relatively simple measurements that are generally used for assessing reverse remodeling after CRT, the impact of baseline LV volumes on CRT response has shown varying results in previous trials.4, 5, 6, 7, 8, 9 We examined the characteristics of echocardiographic CRT responders based on the information obtained from a single‐center multidisciplinary CRT clinic.

Methods

Study Population

We identified 125 patients who completed 2 follow‐up echocardiograms in the multidisciplinary CRT clinic program at Massachusetts General Hospital from 2007 to 2010. All patients met the current criteria for CRT and were followed at regular intervals by a team of subspecialists from the electrophysiology, heart failure (HF), and echocardiography services during the first year after CRT (at 1, 3, and 6 months).10, 11 The protocol was approved by the hospital's institutional review board.

Echocardiography

Standard transthoracic 2‐dimensional echocardiograms were performed on all patients prior to CRT. Late follow‐up echocardiograms were performed at the 6‐month visit after CRT implantation (212 ± 48 d). Left ventricular end‐diastolic volume (LVEDV) and LV end‐systolic volume (LVESV) were measured by biplane method of discs, and LVEF and LV volume indices were calculated.

Response to CRT was defined as a decrease in LVESV of ≥15% and/or an absolute increase of >5% in LVEF on 6‐month visit compared with the baseline echocardiogram.

Statistical Analysis

Data were analyzed using PASW Statistics 18, version 18.0.0 (SPSS Inc., Chicago, IL). Continuous variables were expressed as mean ± SD. The Student t test and Mann‐Whitney U test were used for the comparison of continuous variables between groups. The paired t test and Wilcoxon signed‐rank test were used for the comparison of continuous variables between echocardiographic follow‐ups. The χ ² test and Fisher exact test were used for noncontinuous variables. Logistic multiple regression analysis was used for identifying predictors of CRT response. A 2‐sided P value <0.05 was considered statistically significant.

Results

Baseline Characteristics

There were 81 responders and 44 nonresponders. Table 1 shows baseline characteristics of the study population. There were no significant differences in age, heart rhythm, New York Heart Association class, HF medications, and pre‐CRT LVEF between responders and nonresponders. The distribution of the mitral regurgitation (MR) severity at baseline was also similar between responders and nonresponders. Female sex showed a trend toward an improved response. In contrast, there were significant differences in the etiology of cardiomyopathy, QRS duration, the presence of left bundle branch block (LBBB), and echocardiographic LV volumes and indices between groups. Responders showed more nonischemic etiology, more LBBB presence, longer QRS duration, and smaller baseline LV volumes and indices (Table 1).

Table 1.

Baseline Characteristics of the Study Population

Characteristics	Total (N = 125)	Responder (n = 81)	Nonresponder (n = 44)	P Valuea
Age, y	68 ± 12	67 ± 13	69 ± 10	0.337
Sex, M:F, n (%)	94:31 (75:25)	56:25 (69:31)	38:6 (86:14)	0.050
Etiology, Ncmp:Icmp, n (%)	53:72 (42:58)	42:39 (52:48)	11:33 (25:75)	0.004
Sinus rhythm, n (%)	101 (81)	65 (80)	36 (82)	1.000
QRS width, ms	162 ± 23	166 ± 22	154 ± 23	0.006
LBBB, n (%)	65 (52)	51 (63)	14 (32)	0.001
NYHA class II/III/IV, n (%)	7/100/12 (6/80/10)	3/67/8 (4/83/10)	4/33/4 (9/75/9)	0.451
Diuretics/Sp/digoxin, n (%)	104/39/29 (83/31/23)	66/30/15 (82/37/19)	38/9/14 (86/21/32)	0.619, 0.070, 0.121
BB/ACEI/ARB, n (%)	116/81/27 (93/65/22)	77/51/18 (95/63/22)	39/30/9 (89/68/21)	0.276, 0.695, 0.826
HTN/DM, n (%)	84/42 (67/34)	49/23 (61/28)	35/19 (80/43)	0.045, 0.114
CABG/PCI, n (%)	40/28 (32/22)	22/14 (27/17)	18/14 (41/32)	0.160, 0.075
LVEF, %	25 ± 6	25 ± 6	24 ± 6	0.330
MR ≥ mod, n (%)	40/112 (36)	26/75 (35)	14/37 (38)	0.835
LVESV (n), mL	160 ± 59 (111)	148 ± 50 (74)	183 ± 70 (37)	0.004
LVESVI, mL/m2	81 ± 29	74 ± 23	93 ± 35	0.001
LVEDV (n), mL	208 ± 67 (111)	195 ± 58 (74)	235 ± 76 (37)	0.003
LVEDVI, mL/m2	105 ± 32	98 ± 26	120 ± 38	0.001

Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin II receptor blocker; BB, β‐blocker; CABG, coronary artery bypass graft; DM, diabetes mellitus; F, female; HTN, hypertension; Icmp, ischemic cardiomyopathy; LBBB, left bundle branch block; LVEDV, left ventricular end‐diastolic volume; LVEDVI, left ventricular end‐diastolic volume index; LVEF, left ventricular ejection fraction; LVESV, left ventricular end‐systolic volume; LVESVI, left ventricular end‐systolic volume index; M, male; MR, mitral regurgitation; mod, moderate; N, number of cases; Ncmp, nonischemic cardiomyopathy; NS, not significant; NYHA, New York Heart Association; PCI, percutaneous coronary intervention; Sp, spironolactone.

^aResponder vs nonresponder.

At the 6‐month follow‐up visit after CRT, as was expected by the definition of CRT response, responders showed significant improvements in LVEF (mean increase, 13% ± 7%) and LV volumes (mean decrease of LVESV, 47 mL ± 32 mL; mean decrease of LVEDV, 36 mL ± 38 mL). Regarding MR, overall 26% of patients (29/112) achieved at least grade 1 improvement of MR severity at 6 months; however, there was no significant difference in the prevalence of MR improvement between responders and nonresponders (28% vs 22%, P = 0.617).

Predictors for CRT Response

Table 2 shows univariate and multivariate analyses to identify predictors to echocardiographic CRT response. By univariate analyses, sex, etiology of cardiomyopathy, QRS duration, presence of LBBB, and baseline LVEDV index were associated with predicting CRT response. However, when these variables were analyzed in a multiple logistic regression model, only QRS duration, LBBB, and indexed LVEDV were independent predictors for CRT response. According to the receiver operating characteristic analysis, indexed LVEDV ≤101 mL/m² predicted response with a sensitivity of 61%, specificity of 65%, and positive predictive value of 78% (area under the curve: 0.668, confidence interval [CI]: 0.554–0.781, P = 0.004). Baseline QRS width ≥159 ms predicted CRT responder with a sensitivity of 64%, specificity of 59%, and positive predictive value of 74% (area under the curve: 0.652, CI: 0.550–0.754, P = 0.005).

Table 2.

Single‐Predictor and Multiple‐Predictor Logistic Regression Models of Predicting CRT Response

Variables	Univariate		Multivariate
	OR (95% CI)	P Value	OR (95% CI)	P Value
Age, y	0.985 (0.954–1.016)	0.334	0.981 (0.938–1027)	0.415
Female sex	2.827 (1.059–7.546)	0.038	1.555 (0.403–6.006)	0.522
Nonischemic etiology	3.231 (1.438–7.260)	0.005	1.533 (0.480–4.896)	0.471
QRS width, ms	1.025 (1.006–1.043)	0.008	1.027 (1.004–1.050)	0.023
LBBB	3.643 (1.673–7.934)	0.001	3.568 (1.284–9.910)	0.015
LVEDVI, mL/m2	0.979 (0.965–0.992)	0.002	0.970 (0.953–0.987)	0.001

Abbreviations: CI, confidence interval; CRT, cardiac resynchronization therapy; LBBB, left bundle branch block; LVEDVI, left ventricular end‐diastolic volume index; OR, odds ratio.

Although female sex was associated with moderately increased odds of response on univariate analysis, after adjusting for the other variables, including LVEDV index, it was not a significant independent predictor. Female sex showed smaller baseline LV volumes and indices (female vs male: LVESV index, 68 ± 24 mL/m² vs 84 ± 29 mL/m², P = 0.013; LVEDV index, 91 ± 25 mL/m² vs 110 ± 33 mL/m², P = 0.009) and more nonischemic etiology than males (female vs male: 71% vs 33%, P < 0.001). The presence of LBBB was not significantly different between sexes in our population (female vs male: 61% vs 49%, P = 0.301). Nonischemic etiology was also associated with increased odds of response on univariate analysis; however, after adjusting for indexed LVEDV, etiology was not an independent predictor. Nonischemic cardiomyopathy patients showed smaller baseline LV volumes and indices than did ischemic cardiomyopathy patients (nonischemic vs ischemic: LVESV index, 73 ± 26 mL/m² vs 86 ± 30 mL/m², P = 0.012; LVEDV index, 97 ± 30 mL/m² vs 111 ± 33 mL/m², P = 0.022).

Discussion

This study suggests that in addition to the previously described QRS duration and the presence of LBBB, LV volume prior to CRT implantation may be a determinant for echocardiographic‐defined CRT response. In previous studies, female sex and nonischemic etiology were shown as predictors to favorable CRT response.5, 6, 12, 13, 14, 15 In our study, although female sex and nonischemic etiology were univariate predictors of improved echocardiographic CRT response, when they were considered together with baseline LV volumes, they were no longer independent predictors.

The impact of baseline LV dimension and volume on CRT response has shown varying and sometimes conflicting results, depending on the study population.4, 5, 6, 7, 8, 9 Study populations from Delgado et al8 and Lim et al9 showed significant differences in baseline LV volumes between responders and nonresponders, whereas that from the Predictors of Response to Cardiac Resynchronization Therapy (PROSPECT) trial subanalysis5 presented no significant differences in baseline LV volumes. A recent study by Rickard et al4 demonstrated that the smaller LV dimension prior to CRT showed the most robust improvements in LVEF and in all‐cause mortality. In contrast, the Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronization Therapy (MADIT‐CRT),13 a study of patients with mild HF symptoms showed the opposite relationship between baseline LV volumes and CRT response. Because of the differences in the inclusion criteria for each study population and the use of different definitions for CRT response, direct comparisons of each result are difficult. Although female sex and nonischemic etiology have been reported as independent determinants of CRT response in some studies, their relationship to LV volume has not been explored in depth.6, 7, 12, 13, 15

By virtue of our study being a single‐center study, it has limitations: Data are retrospective, and the sample size of subgroups is relatively small. However, baseline characteristics of our study population, such as mean age, proportions of responder, gender, and cardiomyopathy etiology, QRS width, and LVEF are not different from those of other studies. Importantly, we utilized traditional CRT selection criteria and definitions for echocardiographic CRT response in our study.

Conclusion

In the examination of our multidisciplinary CRT practice, pre‐CRT QRS duration, LBBB, and LVEDV index were independent predictors for echocardiographic‐defined response. This study suggests that traditionally favorable relationships of female sex and nonischemic cardiomyopathy to CRT response may relate to smaller LV volumes of these patients.