Cardiac resynchronization therapy (CRT) has been mostly achieved by biventricular pacing (BVP) in patients with chronic heart failure (CHF), although it can also be provided by left ventricular pacing (LVP). The superiority of BVP over LVP remains uncertain. The present meta-analysis of randomized controlled trials was performed to compare the effects of LVP to BVP in patients with CHF. Outcomes analyzed included clinical status (6-minute walk distance, peak oxygen consumption, quality of life, New York Heart Association class), LV function (LV ejection fraction), and LV remodeling (LV end-systolic volume). Five trials fulfilled criteria for inclusion in analysis, which included 574 patients with CHF indicated for CRT. After a midterm follow-up, pooled analysis demonstrated that LVP resulted in similar improvements in 6-minute walk distance (weighted mean difference [WMD] 11.25, 95% confidence interval [CI] −12.39 to 34.90, p = 0.35), quality of life (WMD 0.34, 95% CI −3.72 to 4.39, p = 0.87), peak oxygen consumption (WMD 1.00, 95% CI −0.84 to 2.85, p = 0.29), and New York Heart Association class (WMD −0.19, 95% CI −0.79 to 0.42, p = 0.54). There was a trend toward a superiority of BVP over LVP for LV ejection fraction (WMD 1.28, 95% CI −0.11 to 2.68, p = 0.07) and LV end-systolic volume (WMD −5.73, 95% CI −11.86 to 0.39, p = 0.07). In conclusion, LVP achieves similar improvement in clinical status as BVP in patients with CHF, whereas there was a trend toward superiority of BVP over LVP for LV reverse modeling and systolic function.
Cardiac resynchronization therapy (CRT) has been widely accepted as adjuvant nonpharmacologic therapy for patients with chronic heart failure (CHF) that is mainly achieved by biventricular pacing (BVP). Although traditional BVP takes effect by minimizing regional left ventricular (LV) delay caused by prolonged ventricular conduction, isolated LV pacing (LVP) has been proposed as an alternative to provide a similar effect. There have been several studies focused on the short-term hemodynamic effect of LVP compared to BVP, in which LVP achieved equivalent or even greater benefits. Studies comparing the effect of LVP to BVP on functional recovery and ventricular reverse remodeling, however, have reached some equivocal conclusions. The superiority of BVP over LVP remains uncertain. This meta-analysis focused on the effect of LVP versus BVP on clinical status, LV systolic function, and LV remodeling in patients with CHF.
Methods
The public-domain databases MEDLINE (January 1948 through January 2011), EMBASE (1991 to last quarter of 2010), and the Cochrane Central Register of Controlled Trials (first quarter of 2011) were searched using terms for 3 themes: BVP, LVP, and CHF (see Online Appendix for complete search strategy). No language restriction was imposed and any non–English-language articles that were retrieved and regarded as relevant to our subject after examination of the title and abstract, which were provided in English in each article, would be translated into English with the assistance of language professionals. In addition, we performed a manual search of secondary sources including references of initially identified articles and a search of reviews, editorials, commentaries, and proceedings from international cardiology meetings.
Studies considered for inclusion met the following criteria: the design was a randomized controlled clinical trial, the study population consisted of patients with LV systolic dysfunction and wide QRS duration, and the main end points included improvement in clinical status (New York Heart Association class, 6-minute walk distance, peak oxygen consumption, and quality-of-life scores, etc.) and/or LV function (LV ejection fraction) and/or LV remodeling (LV end-systolic volume). Studies that did not meet these criteria or with <15 patients in total were excluded.
Two investigators (Y.L. and W.P.) independently performed data abstraction. The third author (Y.S.) checked the information and any discrepancies were resolved by consensus. Criteria for quality assessment included study design aspects such as randomized clinical trial, description of crossover, withdrawals and dropouts, completeness of follow-up, and objectivity of outcome assessment. Data on detailed inclusion criteria such as patient characteristics, duration of follow-up, rates of crossover, and improvements in end points described earlier were abstracted from each study.
For each study data on pooled effects were presented as weighted mean differences (WMDs) with 95% confidence intervals (CIs). Heterogeneity of results across trials was assessed with standard chi-square test with statistical significance set at a p value <0.10 and I 2 statistic with significance set at >50%. The fixed effect was preferentially used as the summary measurement if the assumption of homogeneity was identified; otherwise the random-effects model was used. In case of statistical heterogeneity sensitivity analysis was performed to assess the contribution of each study to the pooled estimate by excluding individual studies 1 at a time and the pooled estimate was recalculated for the remaining studies. When pooled analysis still resulted in significant heterogeneity, the random-effects model was used. All p values <0.05 were considered statistically significant. Data analysis was performed using Review Manager 5.0 (Nordic Cochrane Centre, Copenhagen, Denmark) and STATA 10.1 (STATA Corporation, College Station, Texas).
Results
The search initially retrieved 108 potential relevant articles and 3 other articles were obtained from searches of secondary resources. Ninety-one were excluded after examination of the title and abstract. Of the 20 articles retrieved for further full-text examination, 5 clinical trials comparing BVP to LVP in patients with CHF were included for analysis as outlined in Figure 1 .
Of the 5 included studies 4 had a parallel design and 1 had a cross-over design. Improvements in clinical status and LV function were evaluated in 3 studies and the other 2 studies evaluated these improvements separately. Improvements in LV remodeling were evaluated in 2 studies. Characteristics of included studies are listed in Table 1 . Of note, the Device Evaluation of CONTAK RENEWAL 2 and EASYTRAK 2: Assessment of Safety and Effectiveness in Heart Failure (DECREASE-HF) trial conducted by Rao et al was a prospective 3-arm study intended for a comparison of LVP, simultaneous BVP, and sequential BVP on ventricular volumes and systolic function, and we condensed the 2 BVP groups to create a single pairwise comparison with LVP. The cardiac resynchronization therapy: left or left-and-right for optimal symptomatic effect (LOLA ROSE) trial by Sirker et al was a cross-over study and data were analyzed as 2 separate parallel trials. Statistical tests had been applied for carryover for all comparisons for each clinical outcome measurement, some of which were included in our analysis, and resulted in no statistically significant evidence of carryover. In addition, because changes in New York Heart Association class were evaluated in dichotomous or continuous forms in different studies, data were pooled using the generic inverse-variance method. The Cochrane Risk of Bias Tool was used to assess risk of bias in the included studies, which was generally low as presented in Table 2 . In total our meta-analysis included data on 574 patients with CHF randomized to therapy with BVP or LVP. Baseline characteristics of participants are listed in Table 3 .
| Characteristics | Gasparini et al, 2006 | Rao et al, 2007 | Sirker et al, 2007 | Boriani et al, 2010 | Sedláček et al, 2010 |
|---|---|---|---|---|---|
| (n = 69) | (n = 306) | (n = 18) | (n = 148) | (n = 33) | |
| New York Heart Association class | II–IV | III–IV | III–IV | III–IV | III–IV |
| Left ventricular ejection fraction (%) | <35 | ≤35 | ≤35 | ≤35 | <35 |
| Left ventricular end-diastolic diameter (mm) | >55 | — | — | ≥55 | ≥55 |
| Sinus rhythm | + | + | + | + | + |
| QRS duration (ms) | >130 | ≥150 | ≥120 | ≥130 | ≥150 |
| Left branch bundle block | + | — | + | — | + |
| Biventricular pacing/left ventricular pacing (patients) | 33/36 | 101/205 | 18/18 | 78/70 | 16/17 |
| Follow-up (months) | 12 | 6 | 2 | 6 | 12 |
| End points ⁎ | NYHA class, 6MWD, LVEF, LVESV | LVEF, LVESV | NYHA class, 6MWD, QOL, pVO 2 | NYHA class, 6MWD, QOL, LVEF | NYHA class, 6MWD, QOL, pVO 2 , LVEF |
| Design | Random, parallel | Random, parallel | Random, crossover | Random, parallel | Random, parallel |
| Adequate Sequence Generation | Allocation Concealment | Blinding | Incomplete Outcome Data Addressed | Free of Selective Reporting | Free of Other Bias | |
|---|---|---|---|---|---|---|
| Gasparini et al, 2006 | + | + | Single blind | + | + | + |
| Rao et al, 2007 | − | − | Double blind | + | + | + |
| Sirker et al, 2007 | + | − | Single blind | + | + | + |
| Boriani et al, 2010 | + | + | Double blind | + | + | + |
| Sedláček et al, 2010 | − | − | Double blind | + | + | + |
| Gasparini et al, 2006 | Rao et al, 2007 | Sirker et al, 2007 ⁎ | Boriani et al, 2010 | Sedláček et al, 2010 | |||||
|---|---|---|---|---|---|---|---|---|---|
| BVP | LVP | BVP | LVP | BVP | LVP | BVP | LVP | ||
| Mean age (years) | 67 | 67 | 66 | 67 | 72 | 66 | 66 | 60 | 62 |
| Men | 88% | 94% | 68% | 65% | 88% | 76% | 73% | 56% | 65% |
| Nonischemic cause | 42% | 30% | 38% | 34% | 22% | 47% | 49% | 100% | 100% |
| New York Heart Association class III | — | — | 98% | 97% | 83% | 93% | 94% | 88% | 65% |
| Mean left ventricular ejection fraction (%) | 26 | 25 | 23 | 23 | — | 26 | 25 | 21 | 21 |
| Mean QRS duration (ms) | 176 | 169 | 167 | 165 | 160 | 160 | 162 | 148 | 160 |
| Left branch bundle block | 100% | 100% | 93% | 94% | 100% | 90% | 87% | 100% | 100% |
| Angiotensin-converting enzyme inhibitors/angiotensin receptor blockers | 84% | 81% | 87% | 87% | 94% | 100% | 99% | 100% | 100% |
| β Blockers | 51% | 49% | 81% | 83% | 39% | 91% | 92% | 100% | 94% |
⁎ Trial had a cross-over design and overall statistics were listed.
All studies included demonstrated improvement of clinical status and ventricular systolic function and remodeling from baseline in the BVP and LVP arms after a follow-up of 2 to 12 months. However, meta-analysis showed no superiority of BVP over LVP in 6-minute walk distance improvement (WMD 11.25, 95% CI −12.39 to 34.90, p = 0.35) and quality-of-life improvement (WMD 0.34, 95% CI −3.72 to 4.39, p = 0.87). Moreover, there was no significant difference in benefit from BVP or LVP in peak oxygen consumption (WMD 1.00, 95% CI −0.84 to 2.85, p = 0.29). For reports containing data on New York Heart Association classification BVP resulted in no difference in improvement of New York Heart Association classification compared to LVP (WMD −0.19, 95% CI −0.79 to 0.42, p = 0.54; Figure 2 ) . For LV function and remodeling pooled results showed a trend toward superiority of BVP over LVP for LV ejection fraction (WMD 1.28, 95% CI −0.11 to 2.68, p = 0.07) or LV end-systolic volume (WMD −5.73, 95% CI −11.86 to 0.39, p = 0.07: Figure 3 ) .
