Background
Heart failure (HF) is a debilitating and highly prevalent condition. Recent estimates of the prevalence of heart failure in Asia range from almost 1 % to greater than 6.5 %, [
1,
2], while it ranges from 2 % to 3 % in the United States [
3], and in two of the representative European populations [
4,
5]. The World Health Organization’s most recent report on disease burden indicated that almost 6 million people were diagnosed with HF in a single year [
6]. Many of these were elderly, and multiple studies have demonstrated that the prevalence of HF is greater in older individuals [
3,
7]. The ageing of the world’s population translates to an even greater burden in the coming decades [
1,
8‐
10]. Moreover, the increasing prevalence of diabetes, a risk factor and co-morbidity of HF, further worsens the clinical outcomes [
11‐
22]. The management of HF may need to focus not only on prevention, but also on increasing the treatment success: one recent study of approximately 5 million people in Scotland found that from 1986 to 2003, the median survival after first hospitalization only increased to 1.79 years for women and 2.34 years for men (up from 1.32 and 1.33 years, respectively) [
23].
Among those with HF, arrhythmias, and dyssynchrony are common abnormalities. One treatment option for these patients is cardiac resynchronization therapy (CRT). Since its implementation over 20 years ago, CRT has been proven to improve the outcomes of those with HF [
24‐
28]. However, almost one third of patients who receive CRT do not benefit, giving rise to the concept of non-response [
29,
30]. Although non-response is an ongoing and controversial issue in CRT [
24,
31], studies conducted over 10 years ago revealed that optimizing the pacing site can improve the outcomes and response [
32‐
34].
The use of echocardiography (ECHO) is one method by which lead placement may be optimized. Speckle tracking ECHO (STE), which traces patterns of acoustic signals (i.e., speckles) over time, is a validated device to measure myocardial strain. By measuring strain and strain rate, STE can detect regions of scarred myocardium as well as crucial features of dyssynchrony [
35]. Vector velocity imaging ECHO (ECHO-VVI) is a variation on STE that also tracks the movement of the user-defined endocardial border [
36]. Several groups have used STE or ECHO-VVI not only to define the systolic volume and the latest site of activation, but also to show an association between these factors, concordant lead placement, and CRT response [
37‐
39]. Given the success of these studies, research is now focused on the use of ECHO techniques such as, STE and VVI to prospectively guide lead placement to maximize the effects of CRT. However, a quantitative synthesis of the available data has not yet been published. Accordingly, the objective of the current study was to determine the comparative efficacy and safety of image-guided and conventional CRT.
Methods
Search strategy
The methods used in this review and meta-analysis adhere to the current best practices for conducting systematic reviews of the literature. We searched the PubMed, Cochrane, Embase, and ISI Web of Knowledge databases through April 28, 2014 by using the following combinations of search terms: left ventricular lead placement, cardiac resynchronization therapy, imaging-guided, and echocardiographically guided. Our search strategy in PubMed included the following terms: left ventricular lead placement AND cardiac resynchronization therapy AND (imaging OR echocardiographically guided). Our search filters were: abstract available, English language, and human species.
After the removal of duplicate citations, we identified relevant studies using a 2-step search process. In the first step, we screened the title and abstracts of all citations identified in the search against the inclusion and exclusion criteria (see below). In the second step, we retrieved the full texts of the remaining articles. Studies meeting all of the inclusion criteria and none of the exclusion criteria were included in the analysis.
Selection of studies
The inclusion criteria for this meta-analysis were as follows: 1- The study was a randomized controlled trial (RCT) or a 2-arm prospective study; 2- The study enrolled patients with heart failure who received a CRT pacemaker (CRT-P) or a CRT implantable cardioverter defibrillator (CRT-D) device. A study was excluded for the following reasons: 1- It was a retrospective or cohort study, letter, comment, editorial, or case report; 2- The enrolled patients received other interventions besides those mentioned above.
The following data were extracted from the included studies: first author, year of publication, study design, trial name if applicable, type of intervention, follow-up time, number of patients enrolled, proportion of male patients, New York Heart Association (NYHA) grade of the enrolled patients, width of the QRS, baseline and post-treatment left ventricular ejection fraction (LVEF), baseline and post-treatment left ventricular end-systolic volume (LVESV), concordance of the LV lead to the latest site of activation, odds ratio (OR) for the CRT response, and the rates of CRT response, all-cause mortality, and HF-related hospitalization. Two independent reviewers extracted the data from the eligible studies, and any disagreements were resolved by consulting a third reviewer.
Outcome measures
The primary outcome measures of this analysis were CRT response rate, change in LVEF, and change in LVESV. All 3 included studies defined the CRT response, differently. Khan et al. defined response as LV reverse remodeling (≥15 % decrease in LVESV) at 6 months, Saba et al. as LV reverse remodeling or a ≥5 % absolute increase in LVEF, both with no primary endpoint, and Bai et al. required 2 of 3 criteria: LV reverse remodeling, a ≥20 % relative increase in LVEF, or a 1-class decrease in NYHA functional status.
Secondary outcomes included the rates of all-cause mortality and HF-related hospitalization.
Quality assessment
The quality of the included studies was assessed with the Cochrane Risk of Bias tool, a component of Review Manager 5.1. The assessment is categorized into six domains—random sequence generation (selection bias), allocation concealment (selection bias), blinding of patients and personnel (detection bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), and selective reporting risk (reporting bias)—that are described in Chapter 8 of the Cochrane Handbook for Systematic Reviews Interventions [
40].
Statistical analysis
The OR and corresponding 95 % confidence intervals (CI) of the CRT response, difference in means of change in LVEF, and the difference in means of change in LVESV were used to evaluate the efficacy of image-guided CRT compared to standard CRT. For the OR of the CRT response, an OR > 1, indicates that the image-guided CRT group tended to have a higher response rate than the standard CRT group, while an OR < 1, indicates that the image-guided CRT group tended to have a lower response rate than the standard CRT group, and an OR = 1, indicates that the response rates were similar for both the groups. For the difference in means of LVEF, an OR > 0, indicates that image-guided CRT lead to a greater change (i.e., from pre- to post-treatment) in outcome than did standard CRT, an OR < 0, indicates that image-guided CRT lead to a lesser change in the outcome than the standard CRT, while an OR = 0, indicates that the two procedures had similar effects on the change in outcome. For the difference in means of LVESF, an OR > 0, indicates that image-guided CRT lead to a lesser change in the outcome than the standard CRT, an OR < 0, indicates that image-guided CRT lead to a greater change in outcome than did the standard CRT, and an OR = 0 indicates that the 2 procedures affected the change in outcome to a similar extent. Either a Cochran Q statistic of P < 0.1 or an I
2 statistic > 50 % was considered as obvious heterogeneity among studies. When heterogeneity was detected, a random-effects model was preferred; otherwise, a fixed-effects model was considered. Sensitivity analysis was performed by using the leave-one-out approach. A funnel plot and Egger’s test were not used because publication bias cannot be analyzed accurately with 5 studies or less [
41]. A two sided
P value < 0.05 was considered statistically significant. All statistical analyses were performed with Comprehensive Meta-Analysis statistical software, version 2.0 (Biostat, Englewood, NJ, USA).
Discussion
In this meta-analysis, we determined the comparative efficacy of image-guided and standard CRT in patients with HF. Three studies enrolling approximately 500 patients were included in the analysis. Although other meta-analyses have addressed the ability of STE and/or VVI to predict or identify the determinants of CRT response, we believe that ours is one of the only analyses published to date that examines the ability of these procedures to prospectively optimize lead placement. Accordingly, this work provides an important update to the field. The results indicate that image-guided CRT is significantly better than the standard CRT in improving the overall response, LVEF, and LVESV. A better CRT response is consistent with better improvement of LVEF and LVESV in the image-guided group, as compared to the standard group. A meta-analytic approach highlights the need to synthesize data from multiple studies, as some of the individual data lacked statistical significance. This analysis also revealed the variation in endpoints reported by RCTs or prospective studies of CRT and leads us to propose that greater standardization will increase the benefits of future research. This is especially true for HF-related hospitalization, as these data could provide crucial information on the financial burden of HF.
The effect estimates determined by the pooled analysis were largely concordant with each study’s individual result. The diverse definitions of the CRT response in the studies included, may have contributed to the wide range of response rates (control group: 22 %–57 %, image-guided group: 41 %–72 %). The data reported for the change in LVEF, while supporting image-guided CRT, did not reach statistical significance in one of the three studies. Our pooled estimate was able to take advantage of the increased patient number to determine whether the use of STE or VVI in guiding the LV lead placement has a significant effect on this clinical outcome. Similarly, when data for the change in LVESV were pooled, we found a significant benefit for patients in the image-guided group.
Although we did not conduct a meta-analysis of data for the concordance of LV lead position to the site of latest activation, our summarized results support the hypothesis that the superior outcome in patients in the image-guided CRT group, is in part due to the improved concordance of lead placement. Both the TARGET and STARTER trials found that patients whose LV leads were exactly concordant or adjacent had significantly better outcomes than patients with remote leads. These results were confirmed by 3 recent studies: Kydd et al., analyzed a subset of patients from the derivation and randomized groups of the TARGET trial, [
45]. They have demonstrated a superior CRT response and improved survival by optimal LV lead position at the site of latest mechanical activation, and by avoiding low strain amplitude (scar). Similar results were reported by Marek et al. in a subgroup of the STARTER population of patients without left bundle branch block (LBBB) or a QRS duration between 120 and 149 milliseconds [
46], and by Adelstein et al., where they observed defibrillator therapy-free survival at 3 years among the patients in STARTER trial [
47]. Though, our present data is not sufficient enough to perform a meta-analysis of concordant/adjacent vs. remote LV lead, the aforementioned studies strongly suggest the favorable impact of image-guided CRT on long term survival.
Image-guided CRT may also provide superior results by virtue of its ability to target LV leads to sites that are remote from regions with low radial strain, a hallmark of scarred myocardium. The TARGET trial assessed the relationship between sites of scar and lead placement, and discovered that the lead placement in regions remote from a scar yields improved outcomes. Sade et al. addressed the issue of scars in a follow-up to the STARTER trial [
48]. This study evaluated both the effects of lead placement within the scarred segments and the interaction between scars and activation. In ischemic patients, LV lead placement within or adjacent to the scarred segments had a significantly negative effect on survival when compared to leads that were remote to the scar. Moreover, the clinical outcomes were better in patients when the lead was placed at the latest site of activation and remote to a scar, than in those with a lead placed both within a scar and at the latest-activation site. As Vernooy et al. recently noted [
49], the current body of research strongly suggests that the presence of scars near a lead will hinder the response to CRT. The results from the STARTER and TARGET trials, as well as those from a recent study by Kydd et al. [
50], support this conclusion. However, Vernooy and colleagues noted the lack of evidence to recommend targeting LV leads during CRT. The work of Sade et al. demonstrated the additive value of this aspect of optimization, as survival of patients with a lead being both concordant and remote to the scar was better than a non-concordant lead being remote to a scar [
48]. Taken together, the available data from the 3 patient populations included in our analysis indicate that the use of image-guided CRT will improve outcomes, as it provides superior ability not only to detect and avoid scar, but also to place leads in a location best suited for resynchronization.
The studies included in our meta-analysis showed variability in both the primary and secondary endpoints. Bai et al. and Khan et al. had a primary endpoint of response, but the primary endpoint of Saba et al.—all-cause mortality or HF-related hospitalization—was a secondary outcome only for Khan et al. Among secondary endpoints, just Khan et al. and Bai et al. assessed post-procedural NYHA functional class. The included studies also varied in the reporting of regression analyses to determine factors with predictive value: 2 used regression analyses to identify predictors of outcome; however, the outcomes differed: Bai et al. reported predictors of CRT response while Khan et al. identified predictors of LV reverse remodeling. These differences limit the strength of the present meta-analysis and prompt us to suggest an increased uniformity in endpoint classification across trials. Likewise, whether the efficacy and effectiveness of image-guided CRT response observed in this study will hold true irrespective of the underlying heart failure etiology is not known, as the studies did not compare the CRT response in patients with non-ischemic and ischemic heart failure. Nevertheless, studies elsewhere suggest that, though the improvement of the left ventricular function and remodeling is greater in non-ischemic cardiomyopathy patients, there was no difference in CRT response in terms of reduced mortality or heart failure hospitalization rate between ischemic and non-ischemic heart failure patients [
51]. An increase in LV outflow-tract velocity-time-integral and LV ejection fraction after simultaneous CRT was greater in non-ischemic heart failure patients. However, interventricular pacing interval optimization yielded a similar CRT response in both ischemic and non-ischemic heart failure patients [
52].
This study had several limitations. First, the relatively small sample size of the three studies that met the eligibility criteria, along with the differences in clinical outcomes analyzed, is a major limitation of the current meta-analysis. Although the pooled estimates were robust in our analysis, similar results from a larger cohort of patients would further strengthen our conclusions. However, reports indicate that systematic reviews with small numbers of included studies will be accurate in the final point estimates in the long run [
53]. Second, we could not conduct a meta-analysis of clinical outcomes such as HF-related hospitalization and mortality because of insufficient data, as mentioned above. Third, we did not conduct a subgroup analysis of the comparative performance of image-guided and standard CRT to place leads away from regions of scar.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
JY carried out the literature research, performed the data acquisition and statistical analysis, participated in the design of study, manuscript preparation and drafted the manuscript. ZQ carried out study concepts and participated the manuscript review. MJL carried out the definition of intellectual content and help to draft the manuscript. HB is the guarantor of integrity of the entire study. All authors read and approved the final manuscript.