Author + information
- Received May 5, 2017
- Revision received November 13, 2017
- Accepted November 16, 2017
- Published online January 31, 2018.
- Emilce Trucco, MDa,b,c,
- José María Tolosana, MD, PhDa,b,d,
- Elena Arbelo, MD, PhDa,b,d,
- Ada Doltra, MD, PhDa,b,
- María Ángeles Castel, MD, PhDa,b,d,
- Eva Benito, MDa,b,
- Roger Borràs, BSca,b,
- Eduard Guasch, MD, PhDa,b,d,
- Silvia Vidorreta, RNa,b,
- Barbara Vidal, MD, PhDa,b,d,
- Silvia Montserrat, MD, PhDa,b,d,
- Marta Sitges, MD, PhDa,b,d,
- Antonio Berruezo, MD, PhDa,b,d,
- Josep Brugada, MD, PhDa,b,d and
- Lluís Mont, MD, PhDa,b,d,∗ ()
- aInstitut Clínic Cardio-Vascular (ICCV), Hospital Clínic, Universitat de Barcelona, Catalonia, Spain
- bInstitut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS). Barcelona, Catalonia, Spain
- cDepartment of Cardiology, Hospital Universitari Doctor Josep Trueta, Girona, Spain
- dCentro de Investigacíon Biomédica en Red Enfermedades Cardiovasculares (CIBERCV), Madrid, Spain
- ↵∗Address for correspondence:
Dr. Lluís Mont, Hospital Clínic de Barcelona, C/Villarroel 170, 6º, escala 3 08036 Barcelona, Spain.
Objectives The aim of this study was to compare patient response to cardiac resynchronization therapy (CRT) using fusion-optimized atrioventricular (AV) and interventricular (VV) intervals versus nominal settings.
Background The additional benefit obtained by AV- and VV-interval optimization in patients undergoing CRT remains controversial. Previous studies show short-term benefit in hemodynamic parameters; however, midterm randomized comparison between electrocardiogram optimization and nominal parameters is lacking.
Methods A group of 180 consecutive patients with left bundle branch block treated with CRT were randomized to fusion-optimized intervals (FOI) or nominal settings. In the FOI group, AV and VV intervals were optimized according to the narrowest QRS, using fusion with intrinsic conduction. Clinical response was defined as an increase >10% in the 6-min walk test or an increment of 1 step in New York Heart Association functional class. The left ventricular (LV) remodeling was defined as >15% decrease in left ventricular end-systolic volume (LVESV) at 12-month follow-up. Additionally, patients with LVESV reduction >30% relative to baseline were considered super-responders; by contrast, negative responders had increased LVESV relative to baseline.
Results Participant characteristics included a mean age of 65 ± 10 years, 68% male, 37% with ischemic cardiomyopathy, LV ejection fraction 26 ± 7%, and QRS 180 ± 22 ms. Baseline QRS was shortened significantly more by FOI, compared with nominal settings (−56.55 ± 17.65 ms vs. −37.81 ± 22.07 ms, respectively; p = 0.025). At 12 months, LV reverse remodeling was achieved in a larger proportion of the FOI group (74% vs. 53% [odds ratio: 2.02 (95% confidence interval: 1.08 to 3.76)], respectively; p = 0.026). No significant differences were observed in clinical response (61% vs. 53% [odds ratio: 1.43 (95% confidence interval: 0.79 to 2.59)], respectively; p = 0.24).
Conclusions Device optimization based on FOI achieves greater LV remodeling, compared with nominal settings. (ECG Optimization of CRT: Evaluation of Mid-Term Response [BEST]; NCT01439529)
Cardiac resynchronization therapy (CRT) reduces mortality and heart failure (HF) hospitalizations in selected patients with left ventricular (LV) systolic dysfunction and prolonged QRS duration. This therapy aims to resynchronize the electrical ventricular activation, improving cardiac function and functional status (1,2).
However, not all patients respond to CRT; up to 30% of CRT-implanted patients are currently considered clinical nonresponders (3,4), and up to 50% do not achieve LV reverse remodeling (5). Optimization of CRT pacing intervals may improve results (6–8), increasing the number of responders and the magnitude of the response. The short-term benefit obtained by echocardiographic optimization suggests that atrioventricular (AV) delay and interventricular (VV) interval optimization may further improve response, compared with nominal settings (9,10).
Echocardiography, considered the reference method for AV and VV intervals optimization (11), is complex and time-consuming. Due to its limited feasibility and large interobserver and intraobserver variability, a minority of clinicians perform CRT optimization in routine clinical practice (12). Moreover, despite several studies showing short-term benefits of AV and VV optimization, only limited data exist to suggest that systematic interval optimization results in long-term improvement (5,13,14), and previous randomized studies failed to show superiority of echocardiographic optimization over default parameters during follow-up. QRS-based optimization may be a simpler and more effective way to optimize CRT. A previous randomized study obtained the best short-term hemodynamic response by selecting a VV interval guided by narrowest QRS (15). These results were confirmed by a higher response rate at 6-month follow-up, compared with echocardiographic optimization (16). Furthermore, adding fusion with intrinsic conduction may increase CRT benefit, compared with only LV and right ventricular (RV) pacing (17).
A previous study described a simple new method to optimize AV and VV intervals in CRT, based on obtaining the narrowest QRS using fusion with intrinsic conduction (fusion-optimized intervals [FOI]) (18); short-term hemodynamic results were improved, compared with nominal settings (19). The aim of our study was to compare the clinical response and echocardiographic LV reverse remodeling of CRT using FOI versus the device’s nominal settings programming.
A cohort of 180 consecutive patients who received a CRT were included in the study. The inclusion criteria were: patients with HF, in sinus rhythm, with New York Heart Association (NYHA) functional class II to IV despite optimal medical treatment, LV ejection fraction (LVEF) ≤35%, and QRS width ≥120 ms with left bundle branch block (LBBB) and successful CRT implantation. Patients with treatable cardiomyopathies, life expectancy <1 year, conduction disturbances (AV interval ≥250 ms or complete AV block), or atrial fibrillation were excluded. Patients received a CRT device with or without a defibrillator, based on comorbidities and clinical indication.
Clinical response was defined as an increase >10% at the 6-min walking test (6MWT) or an increment of 1 step in NYHA functional class. The LV remodeling was defined as a decrease exceeding 15% in left ventricular end-systolic volume (LVESV) at 12-month follow-up. Additionally, patients with LVESV reduction >30% relative to baseline were considered super-responders, and negative responders had increased LVESV relative to baseline (20).
All patients underwent a 12-lead electrocardiogram (ECG), echocardiogram, and clinical evaluation before implantation and at 6- and 12-month follow-up. Symptoms of HF, functional capacity, and quality of life were assessed with NYHA functional class, the 6MWT (21), and the Minnesota Living With Heart Failure test (22), respectively. All pharmacological treatment was recorded. Readers blinded to randomization assessed response. Deaths were categorized as cardiac, noncardiac, or unknown. Cardiac deaths were classified as sudden (not preceded by HF or ischemic symptoms) or due to HF per Epstein et al. (23). When the cause of death could not be determined, it was classified as unknown.
Electrocardiographic measurements and FOI
The day after CRT implantation, in the absence of complications and with confirmation of a stable position of the LV electrode with appropriate capture, QRS measurements were performed (E.A., E.T.) in 3 different configurations: during spontaneous sinus rhythm, while using the device’s nominal settings, and after optimization of the AV and VV intervals. Interobserver variability was negligible (0.97 (95% confidence interval [CI]: 0.92 to 0.99) at a screen speed of 300 mm/s and 0.96 (95% CI: 0.89 to 0.98) at 50 mm/s.
The 12 leads of the surface ECG were recorded simultaneously and displayed in vertical alignment on the screen. The FOI measurements were obtained using computerized recordings that were digitally stored (EP-TRACER, Schwarzer CardioTek, Maastricht, the Netherlands). Time measurements were performed at a screen velocity of 300 mm/s by 2 experienced observers unaware of patient randomization, using digital cursors with an accuracy of 1 ms; the mean of 3 consecutive cycles was calculated and recorded. The QRS onset was considered to be the start of the fast deflection and not the spike, as previously reported (15), in order to avoid the mean 45-ms transmural delay due to epicardial pacing of the LV lead. In the case of disagreement in selecting the optimal VV, the measurements were reviewed and the final value was established by consensus. Measurements were repeated at a screen velocity of 50 mm/s (accuracy of 6 ms) (18).
As previously described, optimization is initially performed using only 2 activation waves, intrinsic through the RV and pacing in the LV, looking for a “fusion interval.” Afterward, a third wave is added by pacing the RV apex. If a further shortening of the QRS occurs, this is evidence of a third wave of activation (19). Briefly, to find the fusion band during atrial sensing (AS), the AV interval was progressively shortened with LV pacing only, starting with the longest AV interval that allowed LV capture, and followed by 20-ms decrements until the AV interval produced only LV capture. The AV interval that provided the narrowest QRS was selected and considered as the FO AV interval. This procedure was repeated during atrial pacing (AP) at 10 beats/min above the spontaneous sinus rhythm. Once the FO AV interval during AS and AP was selected, the VV interval was adjusted during AS, comparing QRS duration in different configurations: simultaneous RV and LV pacing (VV 0 ms), LV pre-excitation of 30 ms, and RV pre-excitation of 30 ms. The VV value that obtained the narrowest QRS was considered the FO VV interval (Figure 1) (15).
Standard Doppler echocardiography using a commercially available system (Vivid 7, GE-Vingmed, Milwaukee, Wisconsin) was performed at 24 to 72 h after device implantation and at 6- and 12-month follow-up, during intrinsic rhythm and biventricular pacing in AS mode. Echocardiography was performed by an independent observer (A.D.) unaware of any device programming, and was not used for optimization.
LV volumes and LVEF were calculated by the Simpson rule from the two- and four-chamber apical views. Color Doppler echocardiography was performed in all views after optimizing gain and Nyquist limit.
Following successful CRT implantation, patients were allocated to study groups according to a 1:1 blocked randomization list: FOI programming, based on obtaining the narrowest QRS having fusion with intrinsic conduction, and nominal device settings. Based on previous studies (1–3), the sample size was designed to provide 80% power to show the superiority of increased clinical response to CRT by FOI compared with nominal device settings in a test comparing the primary outcome, with a 2-sided 0.05 alpha level, assuming probabilities of 0.7 (QRS-based optimization by FOI) and 0.5 (nominal settings). The planned sample size was 180 patients. The analysis was based on intention-to-treat. Continuous data are presented as the mean ± SD. To compare means of 2 variables, we used the Student t test. Qualitative variables are expressed as the number of cases and proportions, and compared between groups using the chi-squared test.
Effect sizes are reported with means and 95% CIs for the continuous outcomes, and the odds ratio with 95% CI for categorical (dichotomous) variables, computed with logistic regression modeling. A 2-sided type I error of 5% was used for all tests. Statistical analysis was performed using R software for Windows version 3.3.1 (R project for Statistical Computing, Vienna, Austria).
From a cohort of 180 consecutive patients included in the study and treated with CRT, no patient was lost to follow-up (Figure 2). Baseline characteristics of the patients included in this study did not differ between the 2 groups (Table 1). The mean age was 65 ± 10 years, 68% of the population were men, 37% had ischemic cardiomyopathy, baseline LVEF was 26 ± 7%, and baseline QRS was 180 ± 22 ms. There were no differences in LV lead placement between groups, lateral was the most frequent localization (45% in both groups), followed by posterolateral (42% in FOI vs. 43% in nominal settings). The remaining placements were anterolateral (9% in FOI vs. 6% in nominal settings) and posterior (4% in FOI vs. 6% in nominal settings). Three patients required an epicardial approach due to a lack of an appropriate position. At 12-month follow-up, the percentage of biventricular pacing was 97% in the FOI group and 98% in the nominal settings group.
At 12 months, QRS was significantly narrower in both groups compared with baseline. All patients had the same setting programmed for 12 months after the implantation. However, the FOI group showed greater shortening of QRS, compared with the nominal settings group (181 ± 22 ms at baseline vs. 122 ± 14 ms at 12 months, and 180 ± 21 ms at baseline and 140 ± 17 ms at 12 months, respectively; p = 0.025).
There were no significant differences in percentage of AS and AP between the 2 groups. During the 12-month follow-up, the distributions of AS and AP were 95% and 5% in FOI and 96% and 4% in the nominal setting group, respectively. The mean AV interval was 141 ± 32 ms during AS and 160 ± 19 ms during AP; VV 0 ms was 41%, LV pre-excitation 30 ms was 35%, and RV pre-excitation 30 ms was 18%. Following FOI, 6% of the patients had LV-only pacing.
In the intention-to-treat analysis, 56 (61%) of patients in the FOI group showed a clinical response at 12 months, versus 47 (53%) in the nominal settings group (odds ratio [OR]: 1.43 [95% CI: 0.79 to 2.59]) (p = 0.24).
Improvement in the 6MWT was similar for both groups (Table 2). The quality of life as determined by Minnesota Living With Heart Failure scores also improved similarly for both groups 40 ± 22 to 20 ± 20 points and from 42 ± 19 to 21 ± 20 points, respectively (p = 0.73).
At the end of follow-up, 14 patients (15%) from the FOI group and 22 patients (25%) from the nominal settings group remained in advanced HF (NYHA functional class III to IV), and 77 (85%) and 69 (75%) patients, respectively, were in NYHA functional class I to II.
LV reverse remodeling
In the FOI group, 67 (74%) patients had LV remodeling, compared with 47 (53%) in the nominal settings group (p = 0.026) (Figure 3); LVESV decreased from 168 ± 85 ml to 122 ± 72 ml, and from 161 ± 71 ml to 132 ± 74 ml, respectively, at 12 months (p = 0.013). The LVEF improved from 25 ± 7 to 38 ± 11 in the FOI group and from 26 ± 7 to 32 ± 10 in the nominal settings group (p = 0.002) (Table 2). There was a correlation (r = 0.23; p = 0.01) between response and the degree of QRS narrowing: patients with the most QRS narrowing experienced the greatest benefit, as measured by echocardiographic remodeling. After 12 months of therapy, more patients in the FOI group, compared with the nominal settings group, were classified as super-responders (42 [46%] vs. 31 [35%]; p = 0.06) and fewer patients were negative responders (10 [11%] vs. 21 [24%]; p = 0.041) (Figure 4).
At 12 months, 16 patients (9%) required hospitalization due to worsened clinical condition, 8 in each group. Overall mortality was similar between the groups (p = 0.97). There were 5 deaths in the FOI group, 1 of end-stage HF, 1 of sudden cardiac death, and 3 of noncardiac causes. In the nominal settings group, there were also 5 deaths, all from cardiac events (4 end-stage HF and 1 sudden cardiac death). One patient in the nominal settings group had an infection of the device that required explanting the entire system.
It has been shown that optimization with FOI resulted in short-term improvement in LV dp/dt max in CRT recipients (19), but follow-up data were lacking. To the best of our knowledge, this study is the first prospective, randomized study to compare QRS-based optimization with nominal settings programming during follow-up of CRT patients. After 12 months, patients randomized to FOI had greater reduction in LVESV and improvement in LVEF than patients who received CRT using the parameters of nominal settings. FOI is a simple, highly reproducible method that easily allows optimization in daily clinical practice.
Optimization procedure is not always used in clinical practice, the European Society of Cardiology guidelines have not recommended it as part of routine clinical care (14).
Most trials of CRT in patients with HF have used some method of AV and VV delay optimization, despite the lack of randomized studies showing better CRT response during follow-up when optimization was used, compared with nominal settings. A recent meta-analysis of 13 controlled studies found a small, but significant, improvement in LVEF when CRT was optimized by echocardiogram in patients with HF (13). On the other hand, echocardiography optimization has limitations because it is time consuming and has limited reproducibility.
Three previous randomized studies (24–26) failed to show superiority over nominal settings when intrinsic electrogram (IEGM)-based algorithms were used. The SMART-AV (SmartDelay Determined AV Optimization: A Comparison to Other AV Delay Methods Used in Cardiac Resynchronization Therapy) trial (24) found similar effects on LVESV at 6 months after implantation in the IEGM-based optimization group, the echocardiography-based optimization group, and the group with no optimization, with an empirical atrioventricular setting of 120 ms.
FREEDOM (Frequent Optimization Study Using the QuickOpt Method) (25) reported no inferiority, compared with usual clinical practice (mainly echocardiographic optimization). However, there was no control group using nominal settings.
In aCRT (Adaptive Cardiac Resynchronization Therapy Study) (26), the authors evaluated the clinical effects of a novel IEGM-based algorithm to continuously adapt CRT delivery mode (LV-only pacing or biventricular pacing) over time, per the evolution of rhythm characteristics. This report demonstrated noninferiority of the algorithm in the overall population, compared with echocardiogram-guided optimization at 6 months of follow-up. It also lacks a control group of nominal parameters to ensure that echocardiographic optimization offers some additional benefit.
Another approach is contractility sensor-guided automatic optimization. In RESPOND-CRT (Clinical Trial of the SonRtip Lead and Automatic AV-VV Optimization Algorithm in the PARADYM RF SonR CRT-D) (27), patients were randomized to periodic automatic CRT optimizations with a SonR sensor versus an echocardiogram-guided optimization. The degree of LV remodeling was not evaluated, but SonR optimization resulted in a better clinical response.
In summary, device-based optimization methods have not demonstrated a better clinical or echocardiographic response (with the exception of the RESPONSE-CRT trial), compared with echocardiographic optimization. However, these methods have not been tested in comparison to nominal settings.
AV and VV delay may be optimized by adjusting them until precordial leads show fusion patterns between LV and RV activation wave fronts in the QRS complex (28). Fusion allows intrinsic excitation of the RV via the normal-conducting right bundle branch, which may result in improved RV contraction (29). It also results in a shortening of the QRS, a surrogate measurement for LV activation time, which is associated with a better response to CRT (30). In previous studies, QRS resulting from fusion has been shown to improve hemodynamics (31). Recent work has shown that the triple wave front fusion strategy, in which device settings are selected that allow fusion with intrinsic conduction, was associated with the best hemodynamic improvement by CRT in the majority of patients (32).
Differences in clinical and echocardiographic response
Our data did not show a significant functional improvement associated with FOI optimization. Previous studies have also failed to find a clinical benefit of CRT in less symptomatic patients with HF. Landolina et al. (33) concluded that CRT induced similar improvement in LV function but lesser improvement in functional status in patients in NYHA functional class I to II, compared with those in class III to IV. In the REVERSE (REsynchronization reVErses Remodeling in Systolic left vEntricular dysfunction) study (2), CRT did not improve the clinical composite score in HF but was associated with significant reverse remodeling. Another study (34) also found substantial improvements in LV function in contrast to minimal improvement in exercise capacity and clinical symptoms. Our cohort included a high proportion of baseline NYHA functional class II (30%), which may explain the lack of significant functional improvement. At the end of follow-up, only 36 patients (20%) remained in advanced HF (NYHA functional class III to IV). However, all the patients showed significant improvements in LVEF and LV dimensions, indicating that CRT promotes long-lasting reverse remodeling, even in patients with less symptomatic HF.
Echocardiographic measures of reverse remodeling have been used to assess CRT response (35). In our population, the significant differences in LV remodeling noted at 12 months suggest that the full effect of CRT optimization is achieved within this time range. The importance of LV remodeling derives from its value in predicting long-term survival after CRT. In a well-known substudy of the MADIT CRT (Multicenter Automatic Defibrillator Implantation With Cardiac Resynchronization Therapy) trial by Solomon et al. (36), the relationship between improvements in cardiac size and function, and subsequent outcomes was explored. LV remodeling was predictive of lower risk of HF or death. In a study by Yu et al. (37), LV dyssynchrony was predictive of long-term survival, but multivariable analysis identified reverse LV remodeling as the best predictor of long-term survival. The dramatic reduction in negative responders in our population highlights the importance of adequate programming of the CRT device, not only to achieve response, but also to not worsen LV function.
The main potential limitation of the FOI method of optimization is that it is not incorporated as an automatic algorithm. This may lead to pseudo-fusion or a change in the best interval due to variations in intrinsic AV time; however, the optimized group had better reverse remodeling, and their positive results were not affected by this potential limitation. Another limitation is that FOI is only suitable for those in sinus rhythm with LBBB and normal AV nodal conduction. ECG-based optimization may be useful in patients with AV block, atrial fibrillation, non-LBBB, or prolonged AV. However, intrinsic conduction cannot be included in the equation.
Simultaneous electrocardiographic optimization of the AV and VV intervals to achieve fusion with intrinsic conduction significantly reduced QRS duration and increased LV remodeling, compared with default programming of CRT devices, at 12-month follow-up. This is the first prospective, randomized study comparing ECG optimization to the nominal setting on a CRT device. It provides strong evidence that the simple method of measuring QRS width in the ECG using FOI can easily be used to optimize the programmed AV delay and VV interval in daily clinical practice and allows better response to CRT.
COMPETENCY IN MEDICAL KNOWLEDGE: Simultaneous electrocardiographic optimization of the AV and VV intervals to achieve fusion with intrinsic conduction significantly reduced QRS duration and increased LV remodeling, compared with default programming of CRT devices, at 12-month follow-up.
TRANSLATIONAL OUTLOOK: ECG optimization to the nominal setting on a CRT device is a simple method of measuring QRS width in the ECG using FOI and can be used to optimize the programmed AV delay and VV interval in daily clinical practice, allowing better response to CRT.
The authors thank Elaine Lilly, PhD, for manuscript editing, and Neus Portella, RS, for administrative and editing support.
Dr. Mont has received institutional research grants; and lecture, consulting, and advisory board fees from Medtronic, Biotronik, Boston Scientific, Livanova, and Abbott. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- atrial pacing
- atrial sensing
- confidence interval
- cardiac resynchronization therapy
- fusion-optimized intervals
- heart failure
- intrinsic electrogram
- left bundle branch block
- left ventricular
- left ventricular ejection fraction
- left ventricular end-systolic volume
- New York Heart Association
- odds ratio
- right ventricular
- Received May 5, 2017.
- Revision received November 13, 2017.
- Accepted November 16, 2017.
- 2018 American College of Cardiology Foundation
- Linde C.,
- Abraham W.T.,
- Gold M.R.,
- St John Sutton M.,
- Ghio S.,
- Daubert C.,
- for the REVERSE (REsynchronization reVErses Remodeling in Systolic left vEntricular dysfunction) Study Group
- Tolosana J.M.,
- Mont L.
- Bleasdale R.A.,
- Turner M.S.,
- Mumford C.E.
- Saxon L.A.,
- De Marco T.,
- Schafer J.,
- et al.
- Cazeau S.,
- Gras D.,
- Lazarus A.,
- Ritter P.,
- Mugica J.
- Lunati M.,
- Magenta G.,
- Cattafi G.,
- et al.
- Ypenburg C.,
- van Bommel R.J.,
- Borleffs C.J.W.,
- et al.
- Guyatt G.H.,
- Sullivan M.J.,
- Thompson P.J.,
- et al.
- Epstein A.E.,
- Carlson M.D.,
- Fogoros R.N.,
- Higgins S.L.,
- Venditti F.J. Jr..
- Ellenbogen K.A.,
- Gold M.R.,
- Meyer T.E.,
- et al.
- Abraham W.T.,
- Gras D.,
- Yu C.M.,
- et al.
- Brugada J.,
- Delnoy P.P.,
- Brachmann J.,
- et al.
- Butter C.,
- Auricchio A.,
- Stellbrink C.,
- et al.
- Kobe J.,
- Dechering D.G.,
- Rath B.,
- et al.
- Ter Horst I.A.H.,
- Bogaard M.D.,
- Tuinenburg A.E.,
- et al.
- Landolina M.,
- Lunati M.,
- Gasparini M.,
- et al.,
- InSync/InSync ICD Italian Registry Investigators
- Bleeker G.B.,
- Schalij M.J.,
- Holman E.R.,
- Steendijk P.,
- van der Wall E.E.,
- Bax J.J.
- Barsheshet A.,
- Wang P.J.,
- Moss A.J.,
- et al.
- Solomon S.D.,
- Foster E.,
- Bourgoun M.,
- et al.
- Yu C.-M.,
- Bleeker G.B.,
- Fung J.W.-H.,
- et al.