- Email: [email protected]
Clinical outcomes with synchronized left ventricular pacing: Analysis of the adaptive CRT trial
Clinical outcomes with synchronized left ventricular pacing: Analysis of the adaptive CRT trial David Birnie, MD, MB, ChB,* Bernd Lemke, MD,† Kazutaka Aonuma, MD, PhD,‡ Henry Krum, MBBS, PhD,§ Kathy Lai-Fun Lee, MD,║ Maurizio Gasparini, MD,¶ Randall C. Starling, MD, MPH,# Goran Milasinovic, MD,** John Gorcsan III, MD,†† Mahmoud Houmsse, MD, FHRS,‡‡ Athula Abeyratne, PhD,§§ Alex Sambelashvili, PhD,§§ David O. Martin, MD, MPH# From the *University of Ottawa Heart Institute, Ottawa, Ontario, Canada, †Luedenscheid Clinic, Luedenscheid, Germany, ‡ Tsukuba University Hospital, Tsukuba, Japan, §Monash Centre of Cardiovascular Research & Education in Therapeutics, Melbourne, Australia, ║Queen Mary Hospital, The University of Hong Kong, Hong Kong, ¶IRCCS Istituto Clinico Humanitas, Rozzano, Italy, #The Cleveland Clinic, Cleveland, Ohio, **Clinical Center of Serbia, Belgrade, Serbia, ††University of Pittsburgh, Pittsburg, Pennsylvania, ‡‡Ohio State University, Columbus, Ohio, and §§Medtronic, Mounds View, Minnesota. BACKGROUND Acute studies have suggested that left ventricular pacing (LVP) may have benefits over biventricular pacing (BVP). The adaptive cardiac resynchronization therapy (aCRT) algorithm provides LVP synchronized to produce fusion with the intrinsic activation when the intrinsic atrioventricular (AV) interval is normal. The randomized double-blind adaptive cardiac resynchronization therapy trial demonstrated noninferiority of the aCRT algorithm compared to echocardiography-optimized BVP (control). OBJECTIVE To examine whether synchronized LVP (sLVP) resulted in better clinical outcomes. METHODS First, stratification by percent sLVP (%sLVP) and multivariate Cox proportional hazards model was used to assess the relationship between %sLVP and clinical outcomes. Second, outcomes were compared between patients in the aCRT arm (n ¼ 318) and control patients (n ¼ 160) stratified by intrinsic AV interval at randomization. RESULTS In the aCRT arm, %sLVP ≥50% (n ¼ 142) was independently associated with a decreased risk of death or heart failure hospitalization (hazard ratio 0.49; 95% confidence interval 0.28– 0.85; P ¼ .012) compared with %sLVP o50% (n ¼ 172). A greater proportion of patients with %sLVP ≥50% improved in Packer’s The adaptive cardiac resynchronization therapy trial (NCT00980057) was sponsored by Medtronic Inc, Mounds View, MN. Dr Birnie has received honoraria and research grants from Medtronic. Dr Lemke has received honoraria and speaker’s fees from Medtronic and St Jude Medical and speaker’s fees from Boston Scientific. Dr Krum has received honoraria from Medtronic. Dr Lee has received research grants from Medtronic. Dr Aonuma has received honoraria, speaker’s fees, and research grants from Medtronic. Dr Gasparini has received honoraria and served on advisory boards for Medtronic and Boston Scientific. Dr Starling has received honoraria from Novartis. Dr Milasinovic has received honoraria from Medtronic. Dr Gorcsan has received research grants from Biotronic, GE, Medtronic, St Jude Medical, and Toshiba Medical. Dr Abeyratne is a statistician employed by Medtronic. Dr Sambelashvili is a scientist employed by Medtronic. Dr Martin has received research grants from Medtronic. Address reprint requests and correspondence: Dr David Birnie, University of Ottawa Heart Institute, 40 Ruskin St, Ottawa, Ontario, Canada K1Y 4W7. E-mail address: [email protected].
1547-5271/$-see front matter B 2013 Heart Rhythm Society. All rights reserved.
clinical composite score at 6-month (82% vs 68%; P ¼ .002) and 12-month (80% vs 62%; P ¼ .0006) follow-ups compared to controls. In the subgroup with normal AV (n ¼ 241), there was a lower risk of death or heart failure hospitalization (hazard ratio 0.52; 95% confidence interval 0.27–0.98; P ¼ .044) with the aCRT algorithm. A greater proportion of patients in the aCRT arm improved in the clinical composite score at 6-month (81% vs 69%; P ¼ .041) and 12-month (77% vs 66%; P ¼ .076) follow-ups compared to controls. CONCLUSIONS Higher %sLVP was independently associated with superior clinical outcomes. In patients with normal AV conduction, the aCRT algorithm provided mostly sLVP and demonstrated better clinical outcomes compared to echocardiography-optimized BVP. KEYWORDS Cardiac resynchronization therapy; Synchronized left ventricular pacing; Optimization; AV delay ABBREVIATIONS aCRT ¼ adaptive cardiac resynchronization therapy; AV ¼ atrioventricular; BVP ¼ biventricular pacing; CCS ¼ clinical composite score; CI ¼ confidence interval; CRT ¼ cardiac resynchronization therapy; HF ¼ heart failure; HR ¼ hazard ratio; LBBB ¼ left bundle branch block; LV ¼ left ventricular; LVEF ¼ left ventricular ejection fraction; LVP ¼ left ventricular pacing; NYHA ¼ New York Heart Association; sLVP ¼ synchronized left ventricular pacing; %sLVP ¼ percent synchronized left ventricular pacing; RV ¼ right ventricular; VV ¼ interventricular (Heart Rhythm 2013;10:1368–1374) All rights reserved.
I
2013 Heart Rhythm Society.
Introduction Cardiac resynchronization therapy (CRT) improves symptoms, cardiac function, exercise capacity,1 and survival2 in patients with heart failure (HF) who have prolonged QRS duration. Although CRT is normally achieved through biventricular pacing (BVP), acute3,4 and chronic5–7 studies have shown that left ventricular (LV) only pacing (LVP) is in http://dx.doi.org/10.1016/j.hrthm.2013.07.007
Birnie et al
Outcomes with Synchronized LV Pacing
1369
general noninferior to BVP and can be superior in some patients.7 In these studies, LVP was not specifically timed to synchronize with intrinsic activation. However, small studies have examined patients with intact intrinsic atrioventricular (AV) conduction and normal or near-normal conduction via the right bundle branch. In these patients, LVP synchronized to produce fusion with the intrinsic activation via the right bundle branch (sLVP) results in superior LV8–11 and right ventricular (RV) electrical activation12 and function13 compared to BVP. The adaptive CRT (aCRT) algorithm is a novel algorithm that periodically measures intrinsic conduction and dynamically adjusts CRT pacing parameters.14 If the conduction interval from the right atrium to the right ventricle is normal (ie, the combined AV node and right bundle conduction time is normal), the algorithm provides sLVP. As intrinsic conduction can change over time, maintenance of an appropriate degree of fusion requires dynamic adjustment of the device AV delay. The AV delay is adjusted to produce optimal fusion with the intrinsic activation.15 Conversely, if intrinsic AV conduction interval is prolonged, the algorithm provides BVP. Safety and noninferiority of the aCRT algorithm to echocardiographically optimized BVP with respect to the primary end point of Packer’s clinical composite score (CCS) has been demonstrated in the prospective, multicenter, randomized, double-blind aCRT trial.16 The goal of this study was to evaluate the clinical benefit of sLVP provided by the aCRT algorithm.
improvement in the Packer’s CCS at 6-month follow-up after the randomization. The first secondary end point determined sLVP expressed in terms of the percentage (%sLVP) of total ventricular pacing over the follow-up, which was obtained from the device interrogations. Other end points included time to death or first heart failure (HF) hospitalization, whichever occurred first. Structural and functional end points included changes in left ventricular ejection fraction (LVEF), left ventricular end-systolic volume index, New York Heart Association (NYHA) class, 6-minute hall walk distance, and Minnesota Living with Heart Failure Quality of Life score. All patients were followed at 6- and 12 months and subsequently every 6 months until the study closure. Overall, the aCRT trial demonstrated noninferiority of the aCRT algorithm to echocardiographically optimized CRT with respect to the primary end point of CCS and secondary structural and functional end points listed above.16 In this study, we performed 2 analyses. First, we compared the treatment arm subjects with % sLVP ≥50% and those with %sLVP o50%. Second, we compared patients in the aCRT arm and control patients in the subgroup of patients with normal AV interval at the randomization visit. The end points for the two comparisons were as follows: (1) the time to death or first HF hospitalization over the 12-month follow-up and (2) improvement in CCS at 6- and 12-month follow-ups (primary end point of the trial). We also examined changes in structural and functional end points at 6- and 12-month follow-ups.
Methods
Statistical analysis
Details of the aCRT trial design and main results have been published previously.14,16 Briefly, the trial enrolled patients who did not have permanent atrial tachyarrhythmias and were clinically indicated for the implantation of a de novo cardiac resynchronization therapy-defibrillator device. Within 2 weeks of the cardiac resynchronization therapydefibrillator implant, the patients were randomized to either receive the aCRT algorithm (treatment arm) or undergo echocardiographic optimization of the AV and interventricular (VV) delays (control arm) in 2:1 ratio. The protocol for echocardiographic optimization required initial VV delay adjustment to produce the greatest aortic velocity time integral. Then, at the optimal VV delay, AV optimization was performed by using the iterative method.17 The details of the aCRT algorithm have been published previously.14 Briefly, if the conduction interval from the right atrium to the right ventricle is normal (intrinsic AV ≤200 ms, if in sinus rhythm, or AV ≤250 ms, if receiving atrial pacing) and the heart rate does not exceed 100 beats/min, the algorithm provides sLVP. The AV delay is adjusted to produce optimal fusion with the intrinsic activation.15 Conversely, if intrinsic AV conduction interval is prolonged, the algorithm provides BVP. At 6-month follow-up, the control arm patients underwent another echocardiographic optimization. The first primary end point of the study was the proportion of patients with
Categorical variables, which are reported as number (percent) with condition, were compared by using the Fisher exact test, while continuous variables, which are reported as mean ⫾ SD, were compared by using the Student t test. Kaplan-Meier survival curves with the log-rank test were used to compare the time to death or HF hospitalization, whichever occurred first, over the 12-month follow-up. A multivariate Cox proportional hazard model was used to evaluate the interaction between the composite of mortality or HF hospitalization and variables potentially affecting the treatment effect of CRT. The following variables were considered: %sLVP (in the aCRT arm), baseline QRS duration (dichotomized at the median of 156 ms), presence of left bundle branch block (LBBB), age, sex, body mass index, ischemic etiology of the HF, LVEF, NYHA functional class, renal dysfunction, beta-blocker use, angiotensinconverting enzyme or angiotensin receptor blocker use, and AV interval at randomization. The covariates were entered into a stepwise regression model at a P value of o.3 and remained in the model if the P value was o.05. The hazard ratios (HRs) were adjusted for the covariates that were found to be statistically significant independent predictors by using the Cox proportional hazard model. All analyses were performed according to the intention-to-treat principle by using SAS version 9.2 software (SAS Institute Inc, Cary, NC). Statistical significance was set at P values o.05.
1370
Heart Rhythm, Vol 10, No 9, September 2013
Results Patients The analysis included all 318 patients in the aCRT arm and all 160 control patients of the trial. Four of the 318 patients in the aCRT arm did not have complete device data on ventricular pacing and were excluded from the calculation of %sLVP and its effect on the combined death or HF hospitalizations.
Outcomes stratified by %sLVP in the aCRT arm The distribution of %sLVP in patients in the aCRT arm was reported previously.16 There were 142 of 314 (45%) of patients with %sLVP ≥50% and 172 of 314 (55%) patients who received %sLVP o50% over the 12-month follow-up. The subgroups differed with respect to several patient characteristics (Table 1). Specifically, patients with %sLVP ≥50% predominantly had normal intrinsic AV interval at randomization, had higher prevalence of LBBB and nonischemic cardiomyopathy, were slightly younger, and had fewer men. Figure 1 shows that patients with %sLVP ≥50% experienced a lower rate of death or HF hospitalization compared to patients with %sLVP o50% (log-rank P ¼ .003). To adjust for the baseline confounders, the variables listed in Table 1 together with %sLVP, dichotomized at 50%, were entered into a multivariate Cox proportional hazards stepwise model. As shown in Table 2, %sLVP was found to be a significant independent predictor of the outcome along with renal dysfunction, baseline LVEF, and QRS duration. A greater proportion of patients with %sLVP ≥50% improved in CCS at 6-month (82% vs 68%; P ¼ .002) and 12-month (80% vs 62%; P ¼ .0006) follow-ups compared to controls (Figure 2). Patients with sLVP ≥50% had significantly Table 1
greater improvements in structural and functional changes at 12-month follow-up except Minnesota Living with Heart Failure Quality of Life score (Table 3).
Outcomes in the patients in the aCRT arm and control patients stratified by intrinsic AV interval Patients in the aCRT arm who had normal AV interval at randomization (n ¼ 150) received sLVP 73% ⫾ 28% of the time over 6-month follow-up as opposed to 18% ⫾ 28% in patients with prolonged AV at randomization (n ¼ 164) who received primarily BVP (P o .001). The total percent ventricular pacing was not different (95.5% ⫾ 4.6% vs 95.4% ⫾ 6.5% in the normal and prolonged AV subgroups, respectively; P ¼ .920). Over the 12-month follow-up, % sLVP in subjects with normal and prolonged AV interval was 68% ⫾ 28% and 17% ⫾ 28%, respectively (P o .001; Figure 3). Within the subgroup of patients with normal AV interval at randomization, the aCRT and control arms were well matched on all baseline variables. In this subgroup, there was a lower risk of death or HF hospitalization (HR 0.52; 95% confidence interval [CI] 0.27–0.98; P ¼ .044) with aCRT than with control (log-rank P ¼ .103; Figure 4). A greater proportion of patients in the aCRT arm improved in CCS at 6-month (81% vs 69%; P ¼ .041) and 12-month (77% vs 66%; P ¼ .076) follow-ups compared to controls (Figure 5). There were no significant differences between the two arms in structural or functional changes, although there were trends toward greater improvement in NYHA class (0.98 ⫾ 0.79 vs 0.83 ⫾ 0.81 for aCRT vs control; P ¼ .168) and 6-minute hall walk distance (49.7 ⫾ 96.0 m vs 24.1 ⫾ 126.8 m for aCRT vs control; P ¼ .116) at 6-month follow-up with aCRT. LBBB was present in 80% and 77% of the patients in
Patient characteristics for subgroups with percent synchronized left ventricular pacing ≥50% and o50%
Characteristic *
Normal device AV interval at randomization Intrinsic AV interval measured by the device (As-Vs if sensed rhythm; Ap-Vs ¼ 50 ms if paced rhythm) QRS duration (ms) LBBB RBBB Age (y) Sex: male BMI (kg/m2) Ischemic etiology LVEF (%) NYHA class III Renal dysfunction Beta-blocker use ACE/ARB use
%sLVP ≥50% (n ¼ 142)
%sLVP o50% (n ¼ 172)
P
117 (82%) 174 ⫾ 22
32 (19%) 242 ⫾ 49 (n ¼ 171)
o.001 o.001
154 ⫾ 19 114 (80%) 10 (7%) 64 ⫾ 11 87 (61%) 30 ⫾ 7 55 (41%) (n ¼ 133) 25 ⫾ 6 137 (96%) 25 (18%) 133 (94%) 124 (87%)
155 113 19 67 132 29 84 25 161 44 152 146
⫾ 23 (n ¼ 170) (66%) (11%) ⫾ 11 (77%) ⫾5 (55%) (n ¼ 154) ⫾7 (94%) (26%) (88%) (85%)
.699 .005 .246 .026 .003 .102 .033 .716 .308 .101 .120 .625
Values are expressed as n (%) or as mean ⫾ SD. ACE/ARB ¼ angiotensin-converting enzyme/angiotensin receptor blocker; AV ¼ atrioventricular; BMI ¼ body mass index; CRT-D ¼ cardiac resynchronization therapy-defibrillator; LBBB ¼ left bundle branch block; LVEF ¼ left ventricular ejection fraction; RBBB ¼ right bundle branch block; %sLVP ¼ synchronized leftventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up. * Normal AV as determined by the CRT-D device at randomization: the interval from the right atrial activation to the right ventricular intrinsic activation is ≤200 ms if in sinus rhythm or ≤250 ms if receiving atrial pacing.
Birnie et al
Outcomes with Synchronized LV Pacing
1371
Figure 1 Time to all-cause death or first heart failure hospitalization over the 12-month follow-up stratified by percent synchronized left ventricular pacing in the adaptive CRT arm. CI ¼ confidence interval; CRT ¼ cardiac resynchronization therapy; HF ¼ heart failure; HR ¼ hazard ratio; %sLVP ¼ synchronized left ventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up.
the aCRT arm and control patients with normal intrinsic AV interval, respectively (P ¼ .630). There was no statistically significant difference in the risk of death or HF hospitalization between patients in the aCRT arm and control patients with LBBB (HR 0.56; 95% CI 0.25–1.25; P ¼ .155) or without LBBB (HR 0.82; 95% CI 0.28–2.43; P ¼ .718). In patients with LBBB, there was a trend for greater improvement in CCS at 6-month follow-up with aCRT than with control (82.5% vs 71.4%; P ¼ .099); there was no significant difference in patients without LBBB (76.6% vs 61.9%; P ¼ .351). The interaction between LBBB and CCS improvement was not significant (χ2 test P ¼ 0.240). In the subgroup with prolonged AV interval at baseline, the aCRT and control arms were well matched on all baseline variables, except body mass index and age. In patients with prolonged AV at randomization, patients in the aCRT arm received primarily electrogram-optimized BVP. In this subgroup, there were weak trends toward higher risk of death or HF hospitalizations (HR 1.46; 95% CI 0.76–2.78; P ¼ .256) and somewhat less improvement in CCS compared to the control arm at 6-month (67% vs 77%; P ¼ .161) and 12-month (62% vs 68%; P ¼ .457) followups.
Discussion The aCRT algorithm provides LVP synchronized to produce fusion with intrinsic activation when the intrinsic AV interval is normal. The current study presents two analyses that examined whether patients who received a greater quantity of sLVP had better clinical outcomes. Both analyses found that greater %sLVP was independently associated with a decreased risk of death or HF hospitalization. Also, % LVP was associated with a greater improvement in Packer’s CCS. These findings suggest that there is a subset of patients, that is, those with normal AV intervals, who may do better with sLVP than with conventional BVP. These observations require further prospective investigation. There are a number of potential pathophysiological explanations for our observations. First, sLVP may prevent RV pacing-induced dyssynchrony and allow for more simultaneous RV and LV electrical activation. Second, the continuous adjustment of AV delay may facilitate
Table 2 Multivariate predictors of all-cause death and HF hospitalization at 12-month follow-up Covariate
Value
Hazard ratio (95% CI)
P
Renal dysfunction LVEF (%) QRS duration (ms) %sLVP
Yes Per 1% increase ≤156 ≥50%
2.22 0.93 2.34 0.49
.004 o.001 .003 .012
(1.30–3.81) (0.90–0.97) (1.33–4.10) (0.28–0.85)
Hazard ratio 41.0 indicates that the covariate is associated with worsened outcome. HF ¼ heart failure; LVEF ¼ left ventricular ejection fraction; %sLVP ¼ synchronized left ventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up.
Figure 2 Proportion of patients improved in clinical composite score at 6and 12-month follow-ups stratified by percent synchronized left ventricular pacing in the adaptive CRT arm. CRT ¼ cardiac resynchronization therapy; %sLVP ¼ synchronized left ventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up.
1372
Heart Rhythm, Vol 10, No 9, September 2013
Table 3 Clinical composite score details and structural and functional end points for patients with percent synchronized left ventricular pacing ≥50% and o50% (changes from baseline to 12-month follow-up) Clinical composite score Primary end point Improved Moderately or markedly improved patient global assessment and improved NYHA class Improved NYHA class only Moderately or markedly improved patient global assessment only Unchanged Worsened Death Hospitalized due to or associated with worsening HF Crossover due to worsening HF Moderately or markedly worse patient global assessment and worsened NYHA class Worsened NYHA class Moderately or markedly worse patient global Assessment Secondary end point Change in LVESVi (mL/m2) Change in LVEF (%) Change in NYHA class Change in 6-min hall walk distance (m) Change in MLWHF QoL
%sLVP ≥50% (n ¼ 142)
%sLVP o50% (n ¼ 172)
P
113 (79.6%) 83 (58.5%)
106 (61.6%) 72 (41.9%)
o.001 .005
23 (16.2%) 7 (4.9%)
24 (14.0%) 10 (5.8%)
.635 .806
13 (9.2%) 16 (11.3%) 5 (3.5%) 10 (7.0%) 0 (0.0%) 0 (0.0%)
26 (15.1%) 40 (23.3%) 13 (7.7%) 26 (15.1%) 0 (0.0%) 0 (0.0%)
.124 .007 .148 .032 1.000 1.000
0 (0.0%) 1 (0.7%)
0 (0.0%) 1 (0.6%)
1.000 1.000
18.5 ⫾ 27.3 (n ¼ 115) 6.7 ⫾ 11.6 (n ¼ 115) 1.18 ⫾ 0.72 (n ¼ 132) 53.7 ⫾ 125.5 (n ¼ 131) 21.8 ⫾ 23.0 (n ¼ 119)
11.1 ⫾ 3.6 ⫾ 0.95 ⫾ 16.5 ⫾ 18.9 ⫾
22.9 (n ¼ 128) 10.3 (n ¼ 128) 0.76 (n ¼ 146) 128.5 (n ¼ 142) 21.7 (n ¼ 130)
.022 .030 .011 .016 .309
Values are expressed as n (%) or as mean ⫾ SD. HF ¼ heart failure; LVEF ¼ left ventricular ejection fraction; LVESVi ¼ left ventricular end-systolic volume index; MLWHF QoL ¼ Minnesota Living with Heart Failure Quality of Life; NYHA ¼ New York Heart Association; %sLVP ¼ synchronized left ventricular pacing as a percentage of total ventricular pacing over the 12month follow-up.
resynchronization during exercise. Early studies on CRT4,8 in patients with ventricular conduction delays showed that the maximum acute benefit was obtained with properly timed single-site LV stimulation. For instance, Kass et al8 reported improvement in LV contractility (LV dP/dtmax) of 23% ⫾ 19% and 12% ⫾ 9% for LV free wall pacing and BVP, respectively (P o .05 by paired t test).The authors8
Figure 3 Percent synchronized left ventricular pacing in patients with adaptive CRT who had normal vs prolonged intrinsic AV interval* at randomization. *Normal AV interval as determined by the device: the interval from the right atrial activation to the right ventricular intrinsic activation is ≤200 ms if in sinus rhythm or ≤250 ms if receiving atrial pacing. If the AV interval is not normal, it is classified as prolonged. AV ¼ atrioventricular; CRT ¼ cardiac resynchronization therapy; %sLVP ¼ synchronized left-ventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up.
hypothesized that multisite activation and resynchronization in this case was achieved due to fusion between the slow LV epicardial pacing and fast intrinsic conduction through the still preserved portions of the right bundle branch system. The idea of LV pacing with fusion received further support in an acute study by van Gelder et al.9 In 34 patients with CRT who had LBBB and normal PR interval, the authors compared LV dP/ dtmax with LVP and simultaneous BVP across a variety of AV delays and found superior hemodynamic improvement with LVP, but only when it produced fusion with intrinsic conduction on surface ECG. In a study of 22 patients with CRT, Kurzidim et al10 observed that LVP yielded the greatest improvement in LV dP/dtmax mainly in patients with a PR interval ≤200 ms. Avoidance of RV pacing with sLVP may result in superior RV function when conduction to the RV is sufficiently preserved. Lee et al13 measured RV contractility (RV dP/ dtmax) in 17 CRT recipients and demonstrated that, compared to BVP, LVP timed to precede intrinsic RV activation resulted in greater RV dP/dtmax, preserved RV cycle efficiency, and stroke volume. In an attempt to elucidate the mechanisms of better RV function with LVP, Varma et al12 used noninvasive electrocardiographic imaging to map epicardial electrical excitation during LVP, BVP, and RV pacing at optimized AV intervals in 14 patients with HF who did not have RV dysfunction. They showed that RV pacing and BVP prolonged RV activation time whereas LVP preserved natural RV activation and resulted in RV activation times that were no different from those for intrinsic conduction.12
Birnie et al
Outcomes with Synchronized LV Pacing
1373
Figure 4 Time to all-cause death or first HF hospitalization over the 12-month follow-up in the adaptive CRT (aCRT) and control patients with normal intrinsic AV interval* at randomization. *Normal AV interval as determined by the device: the interval from the right atrial activation to the right ventricular intrinsic activation is ≤200 ms if in sinus rhythm or ≤250 ms if receiving atrial pacing. AV ¼ atrioventricular; CI ¼ confidence interval; CRT ¼ cardiac resynchronization therapy; HF ¼ heart failure; HR ¼ hazard ratio.
Several chronic trials have compared LVP to BVP and, in general, have shown similar clinical, functional, and structural improvement.5–7 Gasparini et al18 randomized 74 patients to either LVP or BVP and observed that the percentage of clinical responders (defined as either 45% increase in LVEF or 410% increase in 6-minute walk distance) at 1-year follow-up in both groups was comparable (75% in the LVP arm and 70% in the BVP arm), and there were no differences in the number of ventricular arrhythmias, hospitalizations, or deaths. The Assessment of Safety and Effectiveness in Heart Failure (DECREASE-HF) trial5 trial randomized 306 patients with CRT who had QRS duration ≥150 ms to LVP, sequential BVP, and simultaneous BVP. At 6-month follow-up, all groups had similar improvement in stroke volume and LVEF but there was a trend toward greater
Figure 5 Proportion of patients improved in clinical composite score at 6and 12-month follow-ups in the subgroup with normal intrinsic AV interval* at randomization. *Normal AV interval as determined by the device: the interval from the right atrial activation to the right ventricular intrinsic activation is ≤200 ms if in sinus rhythm or ≤250 ms if receiving atrial pacing. aCRT ¼ adaptive cardiac resynchronization therapy; AV ¼ atrioventricular.
improvement in LV dimensions and volumes with simultaneous BVP. However, subsequent stratification of the data by baseline PR interval19 showed that patients in the LVP arm who had PR interval o190 ms exhibited a similar amount of reverse remodeling as patients in the BVP arm. Another randomized, double-blind trial, the biventricular versus left univentricular pacing in heart failure patients (B-LEFT HF) trial,6 demonstrated a similar improvement in CCS at 6-month follow-up in 176 CRT recipients randomized to LVP or BVP. A more recent multicenter, double-blind, crossover study, GREATER-EARTH,7 showed similar improvements in exercise duration at 75% of peak VO2, LVEF, LV end-systolic volumes in 121 patients with CRT. Importantly, 20.5% of the clinical nonresponders (defined as ≤20% improvement in exercise performance) to BVP improved with LVP. As for reverse remodeling (defined as 415% reduction in LV endsystolic volume), the proportion of nonresponders to BVP who benefited from LVP was 17.1%. It should be noted that synchronization with intrinsic conduction was not an optimization goal in any of the clinical trials comparing LVP with BVP. For example, in the left ventricular versus simultaneous biventricular pacing in patients with heart failure (GREATER-EARTH study) study, the investigators programmed “…the longest AV delay that fully captured the LV (i.e. no fusion with intrinsic conduction).”7 In the left ventricular versus simultaneous biventricular pacing in patients with heart failure (GREATER-EARTH) study trial, the delays were determined by a proprietary electrogram-based algorithm; in the study by Gasparini et al,18 AV optimization was done through the echocardiographic evaluation of transmitral filling pattern; in the B-LEFT HF trial, the choice of the optimization method was left to the investigators, although echocardiographic Ritter’s and aortic outflow methods were recommended.
1374 Although in patients with normal AV intervals we observed superiority of aCRT over echocardiographic optimization, there was a trend in the opposite direction in the subgroup of patients with prolonged AV interval. In these patients, the algorithm delivers BVP with optimized AV and VV delays on the basis of far-field electrograms. Specifically, the AV delay is optimized on the basis of the device measurements of the Pwave duration and intrinsic AV interval, which have been previously shown to correlate with echo-optimal AV delays.20 Since these correlations were moderate, the electrogram-based adjustment represents an approximation of hemodynamically optimal values, and this could explain the trend toward better outcomes with echo-optimized BVP in the prolonged AV subgroup. It should be noted that the echo optimization protocol in the study was extremely detailed and repeated at 6 months. It is unknown how the algorithm would compare to nominal programming (ie, without echo optimization) in this subgroup. In clinical practice, the majority of patients with CRT do not undergo any optimization.21
Study limitations This study represents a post hoc analysis of the aCRT trial data; therefore, it is possible that the observed differences in outcomes were due to chance or due to imbalances in characteristics that we could not adjust for. These results should be further confirmed by prospective studies. The intrinsic AV interval, which is used by the algorithm to select between sLVP and BVP, is determined automatically by the device. Approximately 50% of the patients in the aCRT arm had normal AV interval and received predominantly sLVP. This is consistent with patient characteristics from other CRT trials, where about half of the patients had normal PR interval.22 However, in individual patients there may be a substantial difference between the device AV delay and PR interval on surface ECG and, therefore, normal PR interval alone may not be used exclusively to predict the impact of the therapy provided by the aCRT algorithm. The AV and VV delays in the control arm of the trial were optimized at randomization and 6-month follow-up using echocardiography. It is, therefore, not known how the algorithm would compare to BVP with nominal AV and VV programming.
Conclusions A higher percentage of synchronized LVP in the aCRT trial was associated with a decreased risk of mortality and HF hospitalizations. In the subgroup of patients with normal AV conduction, the aCRT algorithm provided mostly synchronized LVP and demonstrated better clinical outcomes compared to echocardiographically optimized BVP. These results add to findings from other studies,7 suggesting that there is a substantial subset of patients who may benefit more from LVP than from BVP.
References 1. Abraham WT, Fisher WG, Smith AL, et al. Cardiac resynchronization in chronic heart failure. N Engl J Med 2002;346:1845–1853.
Heart Rhythm, Vol 10, No 9, September 2013 2. Bristow MR, Saxon LA, Boehmer J, et al. Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med 2004;350:2140–2150. 3. Blanc JJ, Etienne Y, Gilard M, et al. Evaluation of different ventricular pacing sites in patients with severe heart failure: results of an acute hemodynamic study. Circulation 1997;96:3273–3277. 4. Auricchio A, Stellbrink C, Block M, et al. Effect of pacing chamber and atrioventricular delay on acute systolic function of paced patients with congestive heart failure. The Pacing Therapies for Congestive Heart Failure Study Group. The Guidant Congestive Heart Failure Research Group. Circulation 1999;99:2993–3001. 5. Rao RK, Kumar UN, Schafer J, Viloria E, De Lurgio D, Foster E. Reduced ventricular volumes and improved systolic function with cardiac resynchronization therapy: a randomized trial comparing simultaneous biventricular pacing, sequential biventricular pacing, and left ventricular pacing. Circulation 2007;115: 2136–2144. 6. Boriani G, Kranig W, Donal E, et al. A randomized double-blind comparison of biventricular versus left ventricular stimulation for cardiac resynchronization therapy: the Biventricular versus Left Univentricular Pacing with ICD Back-up in Heart Failure Patients (B-LEFT HF) trial. Am Heart J 2010;159:1052–1058. 7. Thibault B, Ducharme A, Harel F, et al. Left ventricular versus simultaneous biventricular pacing in patients with heart failure and a QRS complex ≥120 milliseconds. Circulation 2011;124:2874–2881. 8. Kass DA, Chen CH, Curry C, et al. Improved left ventricular mechanics from acute VDD pacing in patients with dilated cardiomyopathy and ventricular conduction delay. Circulation 1999;99:1567–1573. 9. van Gelder BM, Bracke FA, Meijer A, Pijls NH. The hemodynamic effect of intrinsic conduction during left ventricular pacing as compared to biventricular pacing. J Am Coll Cardiol 2005;46:2305–2310. 10. Kurzidim K, Reinke H, Sperzel J, et al. Invasive optimization of cardiac resynchronization therapy: role of sequential biventricular and left ventricular pacing. Pacing Clin Electrophysiol 2005;28:754–761. 11. Verbeek XA, Auricchio A, Yu Y, et al. Tailoring cardiac resynchronization therapy using interventricular asynchrony. Validation of a simple model. Am J Physiol Heart Circ Physiol 2006;290:H968–H977. 12. Varma N, Jia P, Ramanathan C, Rudy Y. RV electrical activation in heart failure during right, left, and biventricular pacing. JACC Cardiovasc Imaging 2010;3: 567–575. 13. Lee KL, Burnes JE, Mullen TJ, Hettrick DA, Tse HF, Lau CP. Avoidance of right ventricular pacing in cardiac resynchronization therapy improves right ventricular hemodynamics in heart failure patients. J Cardiovasc Electrophysiol 2007;18: 497–504. 14. Krum H, Lemke B, Birnie D, et al. A novel algorithm for individualized cardiac resynchronization therapy: rationale and design of the adaptive cardiac resynchronization therapy trial. Am Heart J 2012;163:747–752. 15. Khaykin Y, Exner D, Birnie D, Sapp J, Aggarwal S, Sambelashvili A. Adjusting the timing of left-ventricular pacing using electrocardiogram and device electrograms. Europace 2011;13:1464–1470. 16. Martin DO, Lemke B, Birnie D, et al. Investigation of a novel algorithm for synchronized left-ventricular pacing and ambulatory optimization of cardiac resynchronization therapy: results of the adaptive CRT trial. Heart Rhythm 2012;9:1807–1814. 17. Gorcsan J III, Abraham T, Agler DA, et al. Echocardiography for cardiac resynchronization therapy: recommendations for performance and reporting–a report from the American Society of Echocardiography Dyssynchrony Writing Group endorsed by the Heart Rhythm Society. J Am Soc Echocardiogr 2008;21: 191–213. 18. Gasparini M, Bocchiardo M, Lunati M, et al. Comparison of 1-year effects of left ventricular and biventricular pacing in patients with heart failure who have ventricular arrhythmias and left bundle-branch block: the Bi vs Left Ventricular Pacing: an International Pilot Evaluation on Heart Failure Patients with Ventricular Arrhythmias (BELIEVE) multicenter prospective randomized pilot study. Am Heart J 2006;152:155–157. 19. Husby MP, Holt TA, Elias C, De Lurigio DB. The Significance of PR Interval in LV Only Pacing for Cardiac Resynchronization. Journal of Cardiac Failure 2006;12:S62. (Abstract). 20. Jones RC, Svinarich T, Rubin A, et al. Optimal atrioventricular delay in CRT patients can be approximated using surface electrocardiography and device electrograms. J Cardiovasc Electrophysiol 2010;21:1226–1232. 21. Gras D, Gupta MS, Boulogne E, Guzzo L, Abraham WT. Optimization of AV and VV delays in the real-world CRT patient population: an international survey on current clinical practice. Pacing Clin Electrophysiol 2009;32:S236–S239. 22. Olshansky B, Day JD, Sullivan RM, Yong P, Galle E, Steinberg JS. Does cardiac resynchronization therapy provide unrecognized benefit in patients with prolonged PR intervals? The impact of restoring atrioventricular synchrony: an analysis from the COMPANION Trial. Heart Rhythm 2012;9:34–39.
1547-5271/$-see front matter B 2013 Heart Rhythm Society. All rights reserved.
clinical composite score at 6-month (82% vs 68%; P ¼ .002) and 12-month (80% vs 62%; P ¼ .0006) follow-ups compared to controls. In the subgroup with normal AV (n ¼ 241), there was a lower risk of death or heart failure hospitalization (hazard ratio 0.52; 95% confidence interval 0.27–0.98; P ¼ .044) with the aCRT algorithm. A greater proportion of patients in the aCRT arm improved in the clinical composite score at 6-month (81% vs 69%; P ¼ .041) and 12-month (77% vs 66%; P ¼ .076) follow-ups compared to controls. CONCLUSIONS Higher %sLVP was independently associated with superior clinical outcomes. In patients with normal AV conduction, the aCRT algorithm provided mostly sLVP and demonstrated better clinical outcomes compared to echocardiography-optimized BVP. KEYWORDS Cardiac resynchronization therapy; Synchronized left ventricular pacing; Optimization; AV delay ABBREVIATIONS aCRT ¼ adaptive cardiac resynchronization therapy; AV ¼ atrioventricular; BVP ¼ biventricular pacing; CCS ¼ clinical composite score; CI ¼ confidence interval; CRT ¼ cardiac resynchronization therapy; HF ¼ heart failure; HR ¼ hazard ratio; LBBB ¼ left bundle branch block; LV ¼ left ventricular; LVEF ¼ left ventricular ejection fraction; LVP ¼ left ventricular pacing; NYHA ¼ New York Heart Association; sLVP ¼ synchronized left ventricular pacing; %sLVP ¼ percent synchronized left ventricular pacing; RV ¼ right ventricular; VV ¼ interventricular (Heart Rhythm 2013;10:1368–1374) All rights reserved.
I
2013 Heart Rhythm Society.
Introduction Cardiac resynchronization therapy (CRT) improves symptoms, cardiac function, exercise capacity,1 and survival2 in patients with heart failure (HF) who have prolonged QRS duration. Although CRT is normally achieved through biventricular pacing (BVP), acute3,4 and chronic5–7 studies have shown that left ventricular (LV) only pacing (LVP) is in http://dx.doi.org/10.1016/j.hrthm.2013.07.007
Birnie et al
Outcomes with Synchronized LV Pacing
1369
general noninferior to BVP and can be superior in some patients.7 In these studies, LVP was not specifically timed to synchronize with intrinsic activation. However, small studies have examined patients with intact intrinsic atrioventricular (AV) conduction and normal or near-normal conduction via the right bundle branch. In these patients, LVP synchronized to produce fusion with the intrinsic activation via the right bundle branch (sLVP) results in superior LV8–11 and right ventricular (RV) electrical activation12 and function13 compared to BVP. The adaptive CRT (aCRT) algorithm is a novel algorithm that periodically measures intrinsic conduction and dynamically adjusts CRT pacing parameters.14 If the conduction interval from the right atrium to the right ventricle is normal (ie, the combined AV node and right bundle conduction time is normal), the algorithm provides sLVP. As intrinsic conduction can change over time, maintenance of an appropriate degree of fusion requires dynamic adjustment of the device AV delay. The AV delay is adjusted to produce optimal fusion with the intrinsic activation.15 Conversely, if intrinsic AV conduction interval is prolonged, the algorithm provides BVP. Safety and noninferiority of the aCRT algorithm to echocardiographically optimized BVP with respect to the primary end point of Packer’s clinical composite score (CCS) has been demonstrated in the prospective, multicenter, randomized, double-blind aCRT trial.16 The goal of this study was to evaluate the clinical benefit of sLVP provided by the aCRT algorithm.
improvement in the Packer’s CCS at 6-month follow-up after the randomization. The first secondary end point determined sLVP expressed in terms of the percentage (%sLVP) of total ventricular pacing over the follow-up, which was obtained from the device interrogations. Other end points included time to death or first heart failure (HF) hospitalization, whichever occurred first. Structural and functional end points included changes in left ventricular ejection fraction (LVEF), left ventricular end-systolic volume index, New York Heart Association (NYHA) class, 6-minute hall walk distance, and Minnesota Living with Heart Failure Quality of Life score. All patients were followed at 6- and 12 months and subsequently every 6 months until the study closure. Overall, the aCRT trial demonstrated noninferiority of the aCRT algorithm to echocardiographically optimized CRT with respect to the primary end point of CCS and secondary structural and functional end points listed above.16 In this study, we performed 2 analyses. First, we compared the treatment arm subjects with % sLVP ≥50% and those with %sLVP o50%. Second, we compared patients in the aCRT arm and control patients in the subgroup of patients with normal AV interval at the randomization visit. The end points for the two comparisons were as follows: (1) the time to death or first HF hospitalization over the 12-month follow-up and (2) improvement in CCS at 6- and 12-month follow-ups (primary end point of the trial). We also examined changes in structural and functional end points at 6- and 12-month follow-ups.
Methods
Statistical analysis
Details of the aCRT trial design and main results have been published previously.14,16 Briefly, the trial enrolled patients who did not have permanent atrial tachyarrhythmias and were clinically indicated for the implantation of a de novo cardiac resynchronization therapy-defibrillator device. Within 2 weeks of the cardiac resynchronization therapydefibrillator implant, the patients were randomized to either receive the aCRT algorithm (treatment arm) or undergo echocardiographic optimization of the AV and interventricular (VV) delays (control arm) in 2:1 ratio. The protocol for echocardiographic optimization required initial VV delay adjustment to produce the greatest aortic velocity time integral. Then, at the optimal VV delay, AV optimization was performed by using the iterative method.17 The details of the aCRT algorithm have been published previously.14 Briefly, if the conduction interval from the right atrium to the right ventricle is normal (intrinsic AV ≤200 ms, if in sinus rhythm, or AV ≤250 ms, if receiving atrial pacing) and the heart rate does not exceed 100 beats/min, the algorithm provides sLVP. The AV delay is adjusted to produce optimal fusion with the intrinsic activation.15 Conversely, if intrinsic AV conduction interval is prolonged, the algorithm provides BVP. At 6-month follow-up, the control arm patients underwent another echocardiographic optimization. The first primary end point of the study was the proportion of patients with
Categorical variables, which are reported as number (percent) with condition, were compared by using the Fisher exact test, while continuous variables, which are reported as mean ⫾ SD, were compared by using the Student t test. Kaplan-Meier survival curves with the log-rank test were used to compare the time to death or HF hospitalization, whichever occurred first, over the 12-month follow-up. A multivariate Cox proportional hazard model was used to evaluate the interaction between the composite of mortality or HF hospitalization and variables potentially affecting the treatment effect of CRT. The following variables were considered: %sLVP (in the aCRT arm), baseline QRS duration (dichotomized at the median of 156 ms), presence of left bundle branch block (LBBB), age, sex, body mass index, ischemic etiology of the HF, LVEF, NYHA functional class, renal dysfunction, beta-blocker use, angiotensinconverting enzyme or angiotensin receptor blocker use, and AV interval at randomization. The covariates were entered into a stepwise regression model at a P value of o.3 and remained in the model if the P value was o.05. The hazard ratios (HRs) were adjusted for the covariates that were found to be statistically significant independent predictors by using the Cox proportional hazard model. All analyses were performed according to the intention-to-treat principle by using SAS version 9.2 software (SAS Institute Inc, Cary, NC). Statistical significance was set at P values o.05.
1370
Heart Rhythm, Vol 10, No 9, September 2013
Results Patients The analysis included all 318 patients in the aCRT arm and all 160 control patients of the trial. Four of the 318 patients in the aCRT arm did not have complete device data on ventricular pacing and were excluded from the calculation of %sLVP and its effect on the combined death or HF hospitalizations.
Outcomes stratified by %sLVP in the aCRT arm The distribution of %sLVP in patients in the aCRT arm was reported previously.16 There were 142 of 314 (45%) of patients with %sLVP ≥50% and 172 of 314 (55%) patients who received %sLVP o50% over the 12-month follow-up. The subgroups differed with respect to several patient characteristics (Table 1). Specifically, patients with %sLVP ≥50% predominantly had normal intrinsic AV interval at randomization, had higher prevalence of LBBB and nonischemic cardiomyopathy, were slightly younger, and had fewer men. Figure 1 shows that patients with %sLVP ≥50% experienced a lower rate of death or HF hospitalization compared to patients with %sLVP o50% (log-rank P ¼ .003). To adjust for the baseline confounders, the variables listed in Table 1 together with %sLVP, dichotomized at 50%, were entered into a multivariate Cox proportional hazards stepwise model. As shown in Table 2, %sLVP was found to be a significant independent predictor of the outcome along with renal dysfunction, baseline LVEF, and QRS duration. A greater proportion of patients with %sLVP ≥50% improved in CCS at 6-month (82% vs 68%; P ¼ .002) and 12-month (80% vs 62%; P ¼ .0006) follow-ups compared to controls (Figure 2). Patients with sLVP ≥50% had significantly Table 1
greater improvements in structural and functional changes at 12-month follow-up except Minnesota Living with Heart Failure Quality of Life score (Table 3).
Outcomes in the patients in the aCRT arm and control patients stratified by intrinsic AV interval Patients in the aCRT arm who had normal AV interval at randomization (n ¼ 150) received sLVP 73% ⫾ 28% of the time over 6-month follow-up as opposed to 18% ⫾ 28% in patients with prolonged AV at randomization (n ¼ 164) who received primarily BVP (P o .001). The total percent ventricular pacing was not different (95.5% ⫾ 4.6% vs 95.4% ⫾ 6.5% in the normal and prolonged AV subgroups, respectively; P ¼ .920). Over the 12-month follow-up, % sLVP in subjects with normal and prolonged AV interval was 68% ⫾ 28% and 17% ⫾ 28%, respectively (P o .001; Figure 3). Within the subgroup of patients with normal AV interval at randomization, the aCRT and control arms were well matched on all baseline variables. In this subgroup, there was a lower risk of death or HF hospitalization (HR 0.52; 95% confidence interval [CI] 0.27–0.98; P ¼ .044) with aCRT than with control (log-rank P ¼ .103; Figure 4). A greater proportion of patients in the aCRT arm improved in CCS at 6-month (81% vs 69%; P ¼ .041) and 12-month (77% vs 66%; P ¼ .076) follow-ups compared to controls (Figure 5). There were no significant differences between the two arms in structural or functional changes, although there were trends toward greater improvement in NYHA class (0.98 ⫾ 0.79 vs 0.83 ⫾ 0.81 for aCRT vs control; P ¼ .168) and 6-minute hall walk distance (49.7 ⫾ 96.0 m vs 24.1 ⫾ 126.8 m for aCRT vs control; P ¼ .116) at 6-month follow-up with aCRT. LBBB was present in 80% and 77% of the patients in
Patient characteristics for subgroups with percent synchronized left ventricular pacing ≥50% and o50%
Characteristic *
Normal device AV interval at randomization Intrinsic AV interval measured by the device (As-Vs if sensed rhythm; Ap-Vs ¼ 50 ms if paced rhythm) QRS duration (ms) LBBB RBBB Age (y) Sex: male BMI (kg/m2) Ischemic etiology LVEF (%) NYHA class III Renal dysfunction Beta-blocker use ACE/ARB use
%sLVP ≥50% (n ¼ 142)
%sLVP o50% (n ¼ 172)
P
117 (82%) 174 ⫾ 22
32 (19%) 242 ⫾ 49 (n ¼ 171)
o.001 o.001
154 ⫾ 19 114 (80%) 10 (7%) 64 ⫾ 11 87 (61%) 30 ⫾ 7 55 (41%) (n ¼ 133) 25 ⫾ 6 137 (96%) 25 (18%) 133 (94%) 124 (87%)
155 113 19 67 132 29 84 25 161 44 152 146
⫾ 23 (n ¼ 170) (66%) (11%) ⫾ 11 (77%) ⫾5 (55%) (n ¼ 154) ⫾7 (94%) (26%) (88%) (85%)
.699 .005 .246 .026 .003 .102 .033 .716 .308 .101 .120 .625
Values are expressed as n (%) or as mean ⫾ SD. ACE/ARB ¼ angiotensin-converting enzyme/angiotensin receptor blocker; AV ¼ atrioventricular; BMI ¼ body mass index; CRT-D ¼ cardiac resynchronization therapy-defibrillator; LBBB ¼ left bundle branch block; LVEF ¼ left ventricular ejection fraction; RBBB ¼ right bundle branch block; %sLVP ¼ synchronized leftventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up. * Normal AV as determined by the CRT-D device at randomization: the interval from the right atrial activation to the right ventricular intrinsic activation is ≤200 ms if in sinus rhythm or ≤250 ms if receiving atrial pacing.
Birnie et al
Outcomes with Synchronized LV Pacing
1371
Figure 1 Time to all-cause death or first heart failure hospitalization over the 12-month follow-up stratified by percent synchronized left ventricular pacing in the adaptive CRT arm. CI ¼ confidence interval; CRT ¼ cardiac resynchronization therapy; HF ¼ heart failure; HR ¼ hazard ratio; %sLVP ¼ synchronized left ventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up.
the aCRT arm and control patients with normal intrinsic AV interval, respectively (P ¼ .630). There was no statistically significant difference in the risk of death or HF hospitalization between patients in the aCRT arm and control patients with LBBB (HR 0.56; 95% CI 0.25–1.25; P ¼ .155) or without LBBB (HR 0.82; 95% CI 0.28–2.43; P ¼ .718). In patients with LBBB, there was a trend for greater improvement in CCS at 6-month follow-up with aCRT than with control (82.5% vs 71.4%; P ¼ .099); there was no significant difference in patients without LBBB (76.6% vs 61.9%; P ¼ .351). The interaction between LBBB and CCS improvement was not significant (χ2 test P ¼ 0.240). In the subgroup with prolonged AV interval at baseline, the aCRT and control arms were well matched on all baseline variables, except body mass index and age. In patients with prolonged AV at randomization, patients in the aCRT arm received primarily electrogram-optimized BVP. In this subgroup, there were weak trends toward higher risk of death or HF hospitalizations (HR 1.46; 95% CI 0.76–2.78; P ¼ .256) and somewhat less improvement in CCS compared to the control arm at 6-month (67% vs 77%; P ¼ .161) and 12-month (62% vs 68%; P ¼ .457) followups.
Discussion The aCRT algorithm provides LVP synchronized to produce fusion with intrinsic activation when the intrinsic AV interval is normal. The current study presents two analyses that examined whether patients who received a greater quantity of sLVP had better clinical outcomes. Both analyses found that greater %sLVP was independently associated with a decreased risk of death or HF hospitalization. Also, % LVP was associated with a greater improvement in Packer’s CCS. These findings suggest that there is a subset of patients, that is, those with normal AV intervals, who may do better with sLVP than with conventional BVP. These observations require further prospective investigation. There are a number of potential pathophysiological explanations for our observations. First, sLVP may prevent RV pacing-induced dyssynchrony and allow for more simultaneous RV and LV electrical activation. Second, the continuous adjustment of AV delay may facilitate
Table 2 Multivariate predictors of all-cause death and HF hospitalization at 12-month follow-up Covariate
Value
Hazard ratio (95% CI)
P
Renal dysfunction LVEF (%) QRS duration (ms) %sLVP
Yes Per 1% increase ≤156 ≥50%
2.22 0.93 2.34 0.49
.004 o.001 .003 .012
(1.30–3.81) (0.90–0.97) (1.33–4.10) (0.28–0.85)
Hazard ratio 41.0 indicates that the covariate is associated with worsened outcome. HF ¼ heart failure; LVEF ¼ left ventricular ejection fraction; %sLVP ¼ synchronized left ventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up.
Figure 2 Proportion of patients improved in clinical composite score at 6and 12-month follow-ups stratified by percent synchronized left ventricular pacing in the adaptive CRT arm. CRT ¼ cardiac resynchronization therapy; %sLVP ¼ synchronized left ventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up.
1372
Heart Rhythm, Vol 10, No 9, September 2013
Table 3 Clinical composite score details and structural and functional end points for patients with percent synchronized left ventricular pacing ≥50% and o50% (changes from baseline to 12-month follow-up) Clinical composite score Primary end point Improved Moderately or markedly improved patient global assessment and improved NYHA class Improved NYHA class only Moderately or markedly improved patient global assessment only Unchanged Worsened Death Hospitalized due to or associated with worsening HF Crossover due to worsening HF Moderately or markedly worse patient global assessment and worsened NYHA class Worsened NYHA class Moderately or markedly worse patient global Assessment Secondary end point Change in LVESVi (mL/m2) Change in LVEF (%) Change in NYHA class Change in 6-min hall walk distance (m) Change in MLWHF QoL
%sLVP ≥50% (n ¼ 142)
%sLVP o50% (n ¼ 172)
P
113 (79.6%) 83 (58.5%)
106 (61.6%) 72 (41.9%)
o.001 .005
23 (16.2%) 7 (4.9%)
24 (14.0%) 10 (5.8%)
.635 .806
13 (9.2%) 16 (11.3%) 5 (3.5%) 10 (7.0%) 0 (0.0%) 0 (0.0%)
26 (15.1%) 40 (23.3%) 13 (7.7%) 26 (15.1%) 0 (0.0%) 0 (0.0%)
.124 .007 .148 .032 1.000 1.000
0 (0.0%) 1 (0.7%)
0 (0.0%) 1 (0.6%)
1.000 1.000
18.5 ⫾ 27.3 (n ¼ 115) 6.7 ⫾ 11.6 (n ¼ 115) 1.18 ⫾ 0.72 (n ¼ 132) 53.7 ⫾ 125.5 (n ¼ 131) 21.8 ⫾ 23.0 (n ¼ 119)
11.1 ⫾ 3.6 ⫾ 0.95 ⫾ 16.5 ⫾ 18.9 ⫾
22.9 (n ¼ 128) 10.3 (n ¼ 128) 0.76 (n ¼ 146) 128.5 (n ¼ 142) 21.7 (n ¼ 130)
.022 .030 .011 .016 .309
Values are expressed as n (%) or as mean ⫾ SD. HF ¼ heart failure; LVEF ¼ left ventricular ejection fraction; LVESVi ¼ left ventricular end-systolic volume index; MLWHF QoL ¼ Minnesota Living with Heart Failure Quality of Life; NYHA ¼ New York Heart Association; %sLVP ¼ synchronized left ventricular pacing as a percentage of total ventricular pacing over the 12month follow-up.
resynchronization during exercise. Early studies on CRT4,8 in patients with ventricular conduction delays showed that the maximum acute benefit was obtained with properly timed single-site LV stimulation. For instance, Kass et al8 reported improvement in LV contractility (LV dP/dtmax) of 23% ⫾ 19% and 12% ⫾ 9% for LV free wall pacing and BVP, respectively (P o .05 by paired t test).The authors8
Figure 3 Percent synchronized left ventricular pacing in patients with adaptive CRT who had normal vs prolonged intrinsic AV interval* at randomization. *Normal AV interval as determined by the device: the interval from the right atrial activation to the right ventricular intrinsic activation is ≤200 ms if in sinus rhythm or ≤250 ms if receiving atrial pacing. If the AV interval is not normal, it is classified as prolonged. AV ¼ atrioventricular; CRT ¼ cardiac resynchronization therapy; %sLVP ¼ synchronized left-ventricular pacing as a percentage of total ventricular pacing over the 12-month follow-up.
hypothesized that multisite activation and resynchronization in this case was achieved due to fusion between the slow LV epicardial pacing and fast intrinsic conduction through the still preserved portions of the right bundle branch system. The idea of LV pacing with fusion received further support in an acute study by van Gelder et al.9 In 34 patients with CRT who had LBBB and normal PR interval, the authors compared LV dP/ dtmax with LVP and simultaneous BVP across a variety of AV delays and found superior hemodynamic improvement with LVP, but only when it produced fusion with intrinsic conduction on surface ECG. In a study of 22 patients with CRT, Kurzidim et al10 observed that LVP yielded the greatest improvement in LV dP/dtmax mainly in patients with a PR interval ≤200 ms. Avoidance of RV pacing with sLVP may result in superior RV function when conduction to the RV is sufficiently preserved. Lee et al13 measured RV contractility (RV dP/ dtmax) in 17 CRT recipients and demonstrated that, compared to BVP, LVP timed to precede intrinsic RV activation resulted in greater RV dP/dtmax, preserved RV cycle efficiency, and stroke volume. In an attempt to elucidate the mechanisms of better RV function with LVP, Varma et al12 used noninvasive electrocardiographic imaging to map epicardial electrical excitation during LVP, BVP, and RV pacing at optimized AV intervals in 14 patients with HF who did not have RV dysfunction. They showed that RV pacing and BVP prolonged RV activation time whereas LVP preserved natural RV activation and resulted in RV activation times that were no different from those for intrinsic conduction.12
Birnie et al
Outcomes with Synchronized LV Pacing
1373
Figure 4 Time to all-cause death or first HF hospitalization over the 12-month follow-up in the adaptive CRT (aCRT) and control patients with normal intrinsic AV interval* at randomization. *Normal AV interval as determined by the device: the interval from the right atrial activation to the right ventricular intrinsic activation is ≤200 ms if in sinus rhythm or ≤250 ms if receiving atrial pacing. AV ¼ atrioventricular; CI ¼ confidence interval; CRT ¼ cardiac resynchronization therapy; HF ¼ heart failure; HR ¼ hazard ratio.
Several chronic trials have compared LVP to BVP and, in general, have shown similar clinical, functional, and structural improvement.5–7 Gasparini et al18 randomized 74 patients to either LVP or BVP and observed that the percentage of clinical responders (defined as either 45% increase in LVEF or 410% increase in 6-minute walk distance) at 1-year follow-up in both groups was comparable (75% in the LVP arm and 70% in the BVP arm), and there were no differences in the number of ventricular arrhythmias, hospitalizations, or deaths. The Assessment of Safety and Effectiveness in Heart Failure (DECREASE-HF) trial5 trial randomized 306 patients with CRT who had QRS duration ≥150 ms to LVP, sequential BVP, and simultaneous BVP. At 6-month follow-up, all groups had similar improvement in stroke volume and LVEF but there was a trend toward greater
Figure 5 Proportion of patients improved in clinical composite score at 6and 12-month follow-ups in the subgroup with normal intrinsic AV interval* at randomization. *Normal AV interval as determined by the device: the interval from the right atrial activation to the right ventricular intrinsic activation is ≤200 ms if in sinus rhythm or ≤250 ms if receiving atrial pacing. aCRT ¼ adaptive cardiac resynchronization therapy; AV ¼ atrioventricular.
improvement in LV dimensions and volumes with simultaneous BVP. However, subsequent stratification of the data by baseline PR interval19 showed that patients in the LVP arm who had PR interval o190 ms exhibited a similar amount of reverse remodeling as patients in the BVP arm. Another randomized, double-blind trial, the biventricular versus left univentricular pacing in heart failure patients (B-LEFT HF) trial,6 demonstrated a similar improvement in CCS at 6-month follow-up in 176 CRT recipients randomized to LVP or BVP. A more recent multicenter, double-blind, crossover study, GREATER-EARTH,7 showed similar improvements in exercise duration at 75% of peak VO2, LVEF, LV end-systolic volumes in 121 patients with CRT. Importantly, 20.5% of the clinical nonresponders (defined as ≤20% improvement in exercise performance) to BVP improved with LVP. As for reverse remodeling (defined as 415% reduction in LV endsystolic volume), the proportion of nonresponders to BVP who benefited from LVP was 17.1%. It should be noted that synchronization with intrinsic conduction was not an optimization goal in any of the clinical trials comparing LVP with BVP. For example, in the left ventricular versus simultaneous biventricular pacing in patients with heart failure (GREATER-EARTH study) study, the investigators programmed “…the longest AV delay that fully captured the LV (i.e. no fusion with intrinsic conduction).”7 In the left ventricular versus simultaneous biventricular pacing in patients with heart failure (GREATER-EARTH) study trial, the delays were determined by a proprietary electrogram-based algorithm; in the study by Gasparini et al,18 AV optimization was done through the echocardiographic evaluation of transmitral filling pattern; in the B-LEFT HF trial, the choice of the optimization method was left to the investigators, although echocardiographic Ritter’s and aortic outflow methods were recommended.
1374 Although in patients with normal AV intervals we observed superiority of aCRT over echocardiographic optimization, there was a trend in the opposite direction in the subgroup of patients with prolonged AV interval. In these patients, the algorithm delivers BVP with optimized AV and VV delays on the basis of far-field electrograms. Specifically, the AV delay is optimized on the basis of the device measurements of the Pwave duration and intrinsic AV interval, which have been previously shown to correlate with echo-optimal AV delays.20 Since these correlations were moderate, the electrogram-based adjustment represents an approximation of hemodynamically optimal values, and this could explain the trend toward better outcomes with echo-optimized BVP in the prolonged AV subgroup. It should be noted that the echo optimization protocol in the study was extremely detailed and repeated at 6 months. It is unknown how the algorithm would compare to nominal programming (ie, without echo optimization) in this subgroup. In clinical practice, the majority of patients with CRT do not undergo any optimization.21
Study limitations This study represents a post hoc analysis of the aCRT trial data; therefore, it is possible that the observed differences in outcomes were due to chance or due to imbalances in characteristics that we could not adjust for. These results should be further confirmed by prospective studies. The intrinsic AV interval, which is used by the algorithm to select between sLVP and BVP, is determined automatically by the device. Approximately 50% of the patients in the aCRT arm had normal AV interval and received predominantly sLVP. This is consistent with patient characteristics from other CRT trials, where about half of the patients had normal PR interval.22 However, in individual patients there may be a substantial difference between the device AV delay and PR interval on surface ECG and, therefore, normal PR interval alone may not be used exclusively to predict the impact of the therapy provided by the aCRT algorithm. The AV and VV delays in the control arm of the trial were optimized at randomization and 6-month follow-up using echocardiography. It is, therefore, not known how the algorithm would compare to BVP with nominal AV and VV programming.
Conclusions A higher percentage of synchronized LVP in the aCRT trial was associated with a decreased risk of mortality and HF hospitalizations. In the subgroup of patients with normal AV conduction, the aCRT algorithm provided mostly synchronized LVP and demonstrated better clinical outcomes compared to echocardiographically optimized BVP. These results add to findings from other studies,7 suggesting that there is a substantial subset of patients who may benefit more from LVP than from BVP.
References 1. Abraham WT, Fisher WG, Smith AL, et al. Cardiac resynchronization in chronic heart failure. N Engl J Med 2002;346:1845–1853.
Heart Rhythm, Vol 10, No 9, September 2013 2. Bristow MR, Saxon LA, Boehmer J, et al. Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med 2004;350:2140–2150. 3. Blanc JJ, Etienne Y, Gilard M, et al. Evaluation of different ventricular pacing sites in patients with severe heart failure: results of an acute hemodynamic study. Circulation 1997;96:3273–3277. 4. Auricchio A, Stellbrink C, Block M, et al. Effect of pacing chamber and atrioventricular delay on acute systolic function of paced patients with congestive heart failure. The Pacing Therapies for Congestive Heart Failure Study Group. The Guidant Congestive Heart Failure Research Group. Circulation 1999;99:2993–3001. 5. Rao RK, Kumar UN, Schafer J, Viloria E, De Lurgio D, Foster E. Reduced ventricular volumes and improved systolic function with cardiac resynchronization therapy: a randomized trial comparing simultaneous biventricular pacing, sequential biventricular pacing, and left ventricular pacing. Circulation 2007;115: 2136–2144. 6. Boriani G, Kranig W, Donal E, et al. A randomized double-blind comparison of biventricular versus left ventricular stimulation for cardiac resynchronization therapy: the Biventricular versus Left Univentricular Pacing with ICD Back-up in Heart Failure Patients (B-LEFT HF) trial. Am Heart J 2010;159:1052–1058. 7. Thibault B, Ducharme A, Harel F, et al. Left ventricular versus simultaneous biventricular pacing in patients with heart failure and a QRS complex ≥120 milliseconds. Circulation 2011;124:2874–2881. 8. Kass DA, Chen CH, Curry C, et al. Improved left ventricular mechanics from acute VDD pacing in patients with dilated cardiomyopathy and ventricular conduction delay. Circulation 1999;99:1567–1573. 9. van Gelder BM, Bracke FA, Meijer A, Pijls NH. The hemodynamic effect of intrinsic conduction during left ventricular pacing as compared to biventricular pacing. J Am Coll Cardiol 2005;46:2305–2310. 10. Kurzidim K, Reinke H, Sperzel J, et al. Invasive optimization of cardiac resynchronization therapy: role of sequential biventricular and left ventricular pacing. Pacing Clin Electrophysiol 2005;28:754–761. 11. Verbeek XA, Auricchio A, Yu Y, et al. Tailoring cardiac resynchronization therapy using interventricular asynchrony. Validation of a simple model. Am J Physiol Heart Circ Physiol 2006;290:H968–H977. 12. Varma N, Jia P, Ramanathan C, Rudy Y. RV electrical activation in heart failure during right, left, and biventricular pacing. JACC Cardiovasc Imaging 2010;3: 567–575. 13. Lee KL, Burnes JE, Mullen TJ, Hettrick DA, Tse HF, Lau CP. Avoidance of right ventricular pacing in cardiac resynchronization therapy improves right ventricular hemodynamics in heart failure patients. J Cardiovasc Electrophysiol 2007;18: 497–504. 14. Krum H, Lemke B, Birnie D, et al. A novel algorithm for individualized cardiac resynchronization therapy: rationale and design of the adaptive cardiac resynchronization therapy trial. Am Heart J 2012;163:747–752. 15. Khaykin Y, Exner D, Birnie D, Sapp J, Aggarwal S, Sambelashvili A. Adjusting the timing of left-ventricular pacing using electrocardiogram and device electrograms. Europace 2011;13:1464–1470. 16. Martin DO, Lemke B, Birnie D, et al. Investigation of a novel algorithm for synchronized left-ventricular pacing and ambulatory optimization of cardiac resynchronization therapy: results of the adaptive CRT trial. Heart Rhythm 2012;9:1807–1814. 17. Gorcsan J III, Abraham T, Agler DA, et al. Echocardiography for cardiac resynchronization therapy: recommendations for performance and reporting–a report from the American Society of Echocardiography Dyssynchrony Writing Group endorsed by the Heart Rhythm Society. J Am Soc Echocardiogr 2008;21: 191–213. 18. Gasparini M, Bocchiardo M, Lunati M, et al. Comparison of 1-year effects of left ventricular and biventricular pacing in patients with heart failure who have ventricular arrhythmias and left bundle-branch block: the Bi vs Left Ventricular Pacing: an International Pilot Evaluation on Heart Failure Patients with Ventricular Arrhythmias (BELIEVE) multicenter prospective randomized pilot study. Am Heart J 2006;152:155–157. 19. Husby MP, Holt TA, Elias C, De Lurigio DB. The Significance of PR Interval in LV Only Pacing for Cardiac Resynchronization. Journal of Cardiac Failure 2006;12:S62. (Abstract). 20. Jones RC, Svinarich T, Rubin A, et al. Optimal atrioventricular delay in CRT patients can be approximated using surface electrocardiography and device electrograms. J Cardiovasc Electrophysiol 2010;21:1226–1232. 21. Gras D, Gupta MS, Boulogne E, Guzzo L, Abraham WT. Optimization of AV and VV delays in the real-world CRT patient population: an international survey on current clinical practice. Pacing Clin Electrophysiol 2009;32:S236–S239. 22. Olshansky B, Day JD, Sullivan RM, Yong P, Galle E, Steinberg JS. Does cardiac resynchronization therapy provide unrecognized benefit in patients with prolonged PR intervals? The impact of restoring atrioventricular synchrony: an analysis from the COMPANION Trial. Heart Rhythm 2012;9:34–39.