https://orcid.org/0000-0003-3272-2115
Abstract
To test in a double-blinded, randomized trial whether the combination of electrically guided left ventricular (LV) lead placement and post-implant interventricular pacing delay (VVd) optimization results in superior increase in LV ejection fraction (LVEF) in cardiac resynchronization therapy (CRT) recipients.
Stratified according to presence of ischaemic heart disease, 122 patients were randomized 1:1 to LV lead placement targeted towards the latest electrically activated segment identified by systematic mapping of the coronary sinus tributaries during CRT implantation combined with post-implant VVd optimization (intervention group) or imaging-guided LV lead implantation by cardiac computed tomography venography, 82Rubidium myocardial perfusion imaging and speckle tracking echocardiography targeting the LV lead towards the latest mechanically activated non-scarred myocardial segment (control group). Follow-up was 6 months. Primary endpoint was absolute increase in LVEF. Additional outcome measures were changes in New York Heart Association class, 6-minute walk test, and quality of life, LV reverse remodelling, and device related complications. Analysis was intention-to-treat. A larger increase in LVEF was observed in the intervention group (11 ± 10 vs. 7 ± 11%; 95% confidence interval 0.4–7.9%, P = 0.03); when adjusting for pre-specified baseline covariates this difference did not maintain statistical significance (P = 0.09). Clinical response, LV reverse remodelling, and complication rates did not differ between treatment groups.
Electrically guided CRT implantation appeared non-inferior to an imaging-guided strategy considering the outcomes of change in LVEF, LV reverse remodelling and clinical response. Larger long-term studies are warranted to investigate the effect of an electrically guided CRT strategy.
This was the first randomized controlled trial investigating the effect of an electrically guided CRT strategy positioning the left ventricular lead according to latest electrically activated myocardial segment as identified by mapping of the coronary sinus tributaries combined with post-implant interventricular pacing delay optimization to achieve the shortest biventricular paced QRS duration. This strategy was compared with an imaging-guided approach.
The procedural electrical activation mapping was safe and implementable in the hands of experienced electrophysiologists.
Considering echocardiographic and clinical findings, electrically guided CRT implantation appeared non-inferior to an imaging-guided strategy.
Introduction
Cardiac resynchronization therapy (CRT) improves left ventricular (LV) function and reduces morbidity and mortality in heart failure patients with prolonged QRS duration (QRSd).1,2 However, 30–40% of patients are non-responders.3 According to guidelines, contemporary CRT includes simultaneous biventricular pacing (SIM) with the LV lead positioned in a non-apical, postero-lateral viable myocardial region with late activation.4
Recent randomized controlled trials demonstrated improved CRT response when using pre-implant imaging to guide LV lead position towards the latest mechanically activated non-scarred myocardial segment.5–7 However, imaging-guided strategies are time-consuming, costly, expose the patients to excess contrast and radiation, and may be inapplicable in many centres. Thus, alternative individualized strategies without the need for pre-implant imaging may be applied to optimize CRT. Observational studies indicate a strong association between LV pacing in myocardial regions with late electrical activation and improved outcome.8–10 Furthermore, individual optimization of the interventricular pacing delay (VVd) to achieve the shortest QRSd has been suggested to improve CRT response.11,12 Therefore, in a prospective, randomized double-blinded study, we tested the hypothesis that an electrically guided CRT strategy positioning the LV lead according to latest electrically activated myocardial segment as identified by mapping of the coronary sinus (CS) tributaries combined with post-implant VVd optimization to achieve the shortest biventricular paced QRSd was superior in improving LV function when compared with an imaging-guided strategy.
Methods
Study design and patient population
The ElectroCRT study (electrically vs. imaging-guided implant of the LV lead in cardiac resynchronization therapy) was a single-centre, patient- and assessor-blinded, randomized, controlled clinical trial, conducted at Aarhus University Hospital, Denmark. The study design was previously described.13 In short, patients were included after written informed consent. Inclusion criteria were New York Heart Association (NYHA) Class II–IV despite optimal medical treatment, left bundle branch block (LBBB),14 or chronic right ventricular (RV) paced QRSd ≥180 ms, LV ejection fraction (LVEF) ≤35%, and age ≥40 years. Patients were randomized 1:1 to either (i) electrically guided LV lead positioning targeting latest electrically activated region combined with post-implant VVd optimization to achieve shortest possible QRSd (intervention group) or (ii) imaging-guided LV lead positioning towards latest mechanically activated non-scarred myocardial segment and SIM (control group). Randomization was stratified according to presence of ischaemic heart disease.
Blinding of patients was maintained throughout the study period as all patients underwent identical study procedures including imaging acquisition. Physicians responsible for imaging acquisition and analyses, follow-up, and endpoint adjudication were blinded to group allocation.13 Only the implanting physician and pacemaker technician performing VVd optimization were aware of group allocation. Information on optimal LV lead position according to pre-implant imaging was revealed to the implanting physician only when patients were assigned to the control group. The trial was approved by the local ethics committee and conforms to the declaration of Helsinki. The trial is registered on www.clinicaltrials.gov (NCT02346097).
Clinical evaluation and imaging assessment
All patients underwent pre-implant clinical evaluation and imaging including 6-minute walk test (6MWT), quality of life (QoL) assessment using Minnesota Living with Heart Failure Questionnaire, echocardiography, cardiac computed tomography (CT), and 82Rubidium positron emission tomography (Rb-PET).13
To assess the CS anatomy, pre-implant contrast-enhanced high-pitch electrocardiogram (ECG)-gated cardiac CT was performed using a double-flash protocol. The CS anatomy was visualized using multi-planar and three-dimensional reconstructions. Cardiac CT was omitted in patients with estimated glomerular filtration rate ≤35 mL/min/1.73 m2 or contrast media allergy. To determine final LV lead position, contrast-enhanced high-pitch ECG-gated CT was performed 1 day after implantation.
Two-dimensional echocardiography was performed at baseline and 6-month follow-up. We used Simpson’s biplane method to assess LV volumes and LVEF. Speckle tracking radial strain analysis was performed in the basal and mid-LV short-axis views to identify latest mechanically activated segment. Echocardiographic measurements were averaged over three cycles.
Pre-implant ECG-gated Rb-PET imaging was performed at rest and data were displayed using the 17-segment model. Segments with tracer uptake <50% was considered transmural myocardial scar.15 Scar burden was calculated by summation of points according to segmental tracer uptake (>75% = 0; 50–75% = 1; 25–49% = 2; <25% = 3 points).7
Device implantation
A transvenous approach was used to implant commercially available leads and devices. Preferably, a quadripolar LV lead was implanted. The lead was aimed towards a septal position. Fluoroscopic occlusion venography was performed in 45–60˚ left anterior oblique and 30–45˚ right anterior oblique to guide LV lead positioning. The local LV electrical delay (QLV) at the LV implantation site was measured for all patients, defined as time interval from earliest onset of QRS in the surface ECG to the maximum sensed LV lead voltage change over time recorded in the local electrogram.
Electrically guided cardiac resynchronization therapy strategy
Figure 1 illustrates the electrically guided CRT strategy. The cardiac CT venous angiogram was the only imaging modality available for the implanting physician. To identify and target the latest electrically activated segment, we used the LV lead or a specialized guidewire (Visionwire®, Biotronik, Berlin, Germany) for systematic electrical mapping measuring QLV in basal, mid, and apical segments of CS branches where the LV lead could potentially be implanted.
Electrical mapping to guide implantation of LV lead. Upper part: electrical mapping of QLV was performed in basal, mid, and apical position in all veins suitable for LV lead implantation using a specialized guidewire. QLV was defined as time interval from earliest QRS onset in the surface ECG to the maximum sensed LV lead voltage change over time recorded in the local electrogram. For simplicity, only mapping results of vein B—basal, D—mid, and E—apical are depicted. Longest QLV was measured in vein E—apical. Bottom left: coronary sinus fluoroscopic venography in RAO and LAO view visualizing veins A–E. Bottom right: mapping of QLV, using the specialized guidewire, visualized according to D mid and E apical. ECG, electrocardiogram; LAO, left anterior oblique; LV, left ventricular; QLV, local LV electrical delay; RAO, right anterior oblique.
After deployment of all three leads, individual ECG-guided VVd optimization was performed, choosing the VVd producing the shortest QRSd among five settings: RV pacing 20 ms prior to LV pacing (RV20), SIM, and LV pacing 20 (LV20)—40 (LV40)—60 (LV60) ms prior to RV pacing.16
Imaging-guided cardiac resynchronization therapy strategy
Optimal LV pacing site was defined as the CS branch position closest to the centre of the latest mechanically activated segment without transmural scar. All pre-implant imaging and a prioritized recommendation on optimal LV lead positions were available to the implanting physician. Left ventricular lead position was considered concordant when positioned exactly according to the recommended segment, adjacent if in a neighbouring segment, and remote if ≥2 segments away.
Follow-up study endpoints
The predefined primary endpoint was absolute increase in LVEF at 6-month follow-up.
A composite secondary endpoint of clinical response to CRT was evaluated at 6-month follow-up; this endpoint was reached if the patient was alive, had not been hospitalized for heart failure, and had improved NYHA class or/and increased 6MWT >10%.7 Tertiary endpoints included changes in NYHA class, 6MWT, QoL, LV end-systolic and end-diastolic volumes, and N-terminal prohormone of brain natriuretic peptide levels from baseline to 6-month follow-up. Mortality, hospitalization for heart failure, implantation procedure time, procedural radiation exposure, and device related complications were recorded.
For adjudication of a combined endpoint of all-cause mortality or heart failure hospitalization obtained from the patients’ medical records, patients were additionally followed until 6 months after enrolment of the last included patient.
Statistical analysis
To achieve statistical power of 80% to detect an increase in LVEF of 4% in the intervention group in a superiority analysis, a sample size of 98 patients was needed [given a standard deviation (SD) of 7% in both groups and a two-sided alpha value of 0.05].5 To achieve statistical power of >80% with a 20% margin of non-inferiority for the composite secondary endpoint of clinical response to CRT and assuming 75% response rate in the control group, a sample size of 116 patients was needed. Considering expected loss for follow-up in 5% of patients, 122 patients were included.
All analyses were intention-to-treat and performed before breaking the randomization code. Linear regression was used for continuous outcome measures and logistic regression for binary outcome measures, including presence of ischaemic heart disease as covariate, as randomization was stratified according to this parameter. Additionally, effect of the electrically guided CRT strategy on LVEF change was evaluated in a multivariate regression analysis adjusting for pre-specified baseline covariates: baseline LVEF, ischaemic heart disease, sex, and pre-implant QRSd.13 Continuous variables were presented as mean ± SD for normally distributed variables, otherwise median [interquartile range (IQR)]. Categorical variables were displayed as frequencies with percentages. Baseline and implantation data were compared with Student’s t-tests, Wilcoxon’s rank sum tests, or χ2 tests as appropriate. Intra- and inter-class correlation coefficients were calculated to assess intra- and inter-observer agreement for LVEF assessment. Risk difference was calculated for the composite secondary endpoint of clinical response to CRT at 6-month follow-up. Cox proportional hazards regression analysis stratified according to heart failure aetiology was applied for the combined endpoint of all-cause mortality or heart failure hospitalization during the additional follow-up. Two-sided P-value <0.05 was considered statistically significant. We used commercially available software (Stata 14.2; Stata-Corp, College Station, TX, USA) for statistical analysis.
Results
Study population
After screening 331 patients scheduled for elective CRT, we included 122 patients from February 2015 to December 2017 (Figure 2). Baseline characteristics were balanced between groups (Table 1).
Consort diagram. Exclusion due to logistical problems was mainly driven by limited access to all three imaging modalities prior to implantation. CABG, coronary artery bypass graft; CRT, cardiac resynchronization therapy; LBBB, left bundle branch block; LVEF, left ventricular ejection fraction; MI, myocardial infarction.
Baseline characteristics
| Intervention group (n = 60) | Control group (n = 62) | P-value | |
|---|---|---|---|
| Age (years) | 72 ± 8 | 70 ± 10 | 0.25 |
| Female | 14 (23) | 17 (27) | 0.60 |
| Ischaemic heart disease | 32 (53) | 28 (47) | 0.37 |
| Atrial fibrillation, permanent | 7 (12) | 9 (15) | 0.64 |
| Diabetes | 17 (28) | 20 (32) | 0.64 |
| NYHA Class I/II/III/IV | 0 (0)/35 (58)/24 (40)/1 (2) | 0 (0)/41 (66)/19 (31)/2 (3) | 0.50 |
| 6MWT (m) | 379 ± 106 | 410 ± 105 | 0.12 |
| QoL, MLHFQ (points) | 32 ± 20 | 31 ± 20 | 0.65 |
| Medical therapy | |||
| Beta-blockers | 54 (90) | 59 (95) | 0.28 |
| ACEI/ARB-II | 58 (97) | 56 (90) | 0.16 |
| Spironolactone | 38 (63) | 40 (65) | 0.89 |
| Loop diuretics | 44 (73) | 47 (76) | 0.75 |
| QRS duration (ms)a | 170 ± 17 | 169 ± 23 | 0.63 |
| Antero-septal to posterior radial strain delay (ms) | 333 ± 184 | 314 ± 182 | 0.58 |
| Dyssynchrony by radial strainb | 52 (87) | 52 (84) | 0.66 |
| LVEDV (mL) | 196 ± 69 | 189 ± 66 | 0.60 |
| LVESV (mL) | 142 ± 56 | 132 ± 54 | 0.33 |
| LVEF (%) | 29 ± 8 | 31 ± 8 | 0.06 |
| eGRF (mL/min/1.73 m2) | 62 ± 19 | 60 ± 21 | 0.62 |
| NT-proBNP (ng/L) | 3263 ± 3840 | 2150 ± 2346 | 0.06 |
| Scar burden by Rb-PET (points) | 8 ± 3 | 9 ± 4 | 0.15 |
| Chronic RV pacing | 4 (7) | 10 (16) | 0.10 |
| Intervention group (n = 60) | Control group (n = 62) | P-value | |
|---|---|---|---|
| Age (years) | 72 ± 8 | 70 ± 10 | 0.25 |
| Female | 14 (23) | 17 (27) | 0.60 |
| Ischaemic heart disease | 32 (53) | 28 (47) | 0.37 |
| Atrial fibrillation, permanent | 7 (12) | 9 (15) | 0.64 |
| Diabetes | 17 (28) | 20 (32) | 0.64 |
| NYHA Class I/II/III/IV | 0 (0)/35 (58)/24 (40)/1 (2) | 0 (0)/41 (66)/19 (31)/2 (3) | 0.50 |
| 6MWT (m) | 379 ± 106 | 410 ± 105 | 0.12 |
| QoL, MLHFQ (points) | 32 ± 20 | 31 ± 20 | 0.65 |
| Medical therapy | |||
| Beta-blockers | 54 (90) | 59 (95) | 0.28 |
| ACEI/ARB-II | 58 (97) | 56 (90) | 0.16 |
| Spironolactone | 38 (63) | 40 (65) | 0.89 |
| Loop diuretics | 44 (73) | 47 (76) | 0.75 |
| QRS duration (ms)a | 170 ± 17 | 169 ± 23 | 0.63 |
| Antero-septal to posterior radial strain delay (ms) | 333 ± 184 | 314 ± 182 | 0.58 |
| Dyssynchrony by radial strainb | 52 (87) | 52 (84) | 0.66 |
| LVEDV (mL) | 196 ± 69 | 189 ± 66 | 0.60 |
| LVESV (mL) | 142 ± 56 | 132 ± 54 | 0.33 |
| LVEF (%) | 29 ± 8 | 31 ± 8 | 0.06 |
| eGRF (mL/min/1.73 m2) | 62 ± 19 | 60 ± 21 | 0.62 |
| NT-proBNP (ng/L) | 3263 ± 3840 | 2150 ± 2346 | 0.06 |
| Scar burden by Rb-PET (points) | 8 ± 3 | 9 ± 4 | 0.15 |
| Chronic RV pacing | 4 (7) | 10 (16) | 0.10 |
Values are presented as mean ± standard deviation or n (%).
ACEI, angiotensin-converting enzyme inhibitors; ARB-II, angiotensin 2 receptor blockers; eGFR, estimated glomerular filtration rate; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; MLHFQ, Minnesota Living with Heart Failure Questionnaire; 6MWT, 6-minute walk test; NT-proBNP, N-terminal prohormone of brain natriuretic peptide; NYHA, New York Heart Association; QoL, quality of life; Rb-PET, 82Rubidium positron emission tomography; RV, right ventricular.
QRS duration was determined as the longest QRS duration in any of the 12 leads in printed paper 12’ lead ECGs (speed 25 mm/s).
Dyssynchrony by radial strain was defined as ≥130 ms delay between antero-septal and posterior time to peak from QRS onset.
Baseline characteristics
| Intervention group (n = 60) | Control group (n = 62) | P-value | |
|---|---|---|---|
| Age (years) | 72 ± 8 | 70 ± 10 | 0.25 |
| Female | 14 (23) | 17 (27) | 0.60 |
| Ischaemic heart disease | 32 (53) | 28 (47) | 0.37 |
| Atrial fibrillation, permanent | 7 (12) | 9 (15) | 0.64 |
| Diabetes | 17 (28) | 20 (32) | 0.64 |
| NYHA Class I/II/III/IV | 0 (0)/35 (58)/24 (40)/1 (2) | 0 (0)/41 (66)/19 (31)/2 (3) | 0.50 |
| 6MWT (m) | 379 ± 106 | 410 ± 105 | 0.12 |
| QoL, MLHFQ (points) | 32 ± 20 | 31 ± 20 | 0.65 |
| Medical therapy | |||
| Beta-blockers | 54 (90) | 59 (95) | 0.28 |
| ACEI/ARB-II | 58 (97) | 56 (90) | 0.16 |
| Spironolactone | 38 (63) | 40 (65) | 0.89 |
| Loop diuretics | 44 (73) | 47 (76) | 0.75 |
| QRS duration (ms)a | 170 ± 17 | 169 ± 23 | 0.63 |
| Antero-septal to posterior radial strain delay (ms) | 333 ± 184 | 314 ± 182 | 0.58 |
| Dyssynchrony by radial strainb | 52 (87) | 52 (84) | 0.66 |
| LVEDV (mL) | 196 ± 69 | 189 ± 66 | 0.60 |
| LVESV (mL) | 142 ± 56 | 132 ± 54 | 0.33 |
| LVEF (%) | 29 ± 8 | 31 ± 8 | 0.06 |
| eGRF (mL/min/1.73 m2) | 62 ± 19 | 60 ± 21 | 0.62 |
| NT-proBNP (ng/L) | 3263 ± 3840 | 2150 ± 2346 | 0.06 |
| Scar burden by Rb-PET (points) | 8 ± 3 | 9 ± 4 | 0.15 |
| Chronic RV pacing | 4 (7) | 10 (16) | 0.10 |
| Intervention group (n = 60) | Control group (n = 62) | P-value | |
|---|---|---|---|
| Age (years) | 72 ± 8 | 70 ± 10 | 0.25 |
| Female | 14 (23) | 17 (27) | 0.60 |
| Ischaemic heart disease | 32 (53) | 28 (47) | 0.37 |
| Atrial fibrillation, permanent | 7 (12) | 9 (15) | 0.64 |
| Diabetes | 17 (28) | 20 (32) | 0.64 |
| NYHA Class I/II/III/IV | 0 (0)/35 (58)/24 (40)/1 (2) | 0 (0)/41 (66)/19 (31)/2 (3) | 0.50 |
| 6MWT (m) | 379 ± 106 | 410 ± 105 | 0.12 |
| QoL, MLHFQ (points) | 32 ± 20 | 31 ± 20 | 0.65 |
| Medical therapy | |||
| Beta-blockers | 54 (90) | 59 (95) | 0.28 |
| ACEI/ARB-II | 58 (97) | 56 (90) | 0.16 |
| Spironolactone | 38 (63) | 40 (65) | 0.89 |
| Loop diuretics | 44 (73) | 47 (76) | 0.75 |
| QRS duration (ms)a | 170 ± 17 | 169 ± 23 | 0.63 |
| Antero-septal to posterior radial strain delay (ms) | 333 ± 184 | 314 ± 182 | 0.58 |
| Dyssynchrony by radial strainb | 52 (87) | 52 (84) | 0.66 |
| LVEDV (mL) | 196 ± 69 | 189 ± 66 | 0.60 |
| LVESV (mL) | 142 ± 56 | 132 ± 54 | 0.33 |
| LVEF (%) | 29 ± 8 | 31 ± 8 | 0.06 |
| eGRF (mL/min/1.73 m2) | 62 ± 19 | 60 ± 21 | 0.62 |
| NT-proBNP (ng/L) | 3263 ± 3840 | 2150 ± 2346 | 0.06 |
| Scar burden by Rb-PET (points) | 8 ± 3 | 9 ± 4 | 0.15 |
| Chronic RV pacing | 4 (7) | 10 (16) | 0.10 |
Values are presented as mean ± standard deviation or n (%).
ACEI, angiotensin-converting enzyme inhibitors; ARB-II, angiotensin 2 receptor blockers; eGFR, estimated glomerular filtration rate; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; MLHFQ, Minnesota Living with Heart Failure Questionnaire; 6MWT, 6-minute walk test; NT-proBNP, N-terminal prohormone of brain natriuretic peptide; NYHA, New York Heart Association; QoL, quality of life; Rb-PET, 82Rubidium positron emission tomography; RV, right ventricular.
QRS duration was determined as the longest QRS duration in any of the 12 leads in printed paper 12’ lead ECGs (speed 25 mm/s).
Dyssynchrony by radial strain was defined as ≥130 ms delay between antero-septal and posterior time to peak from QRS onset.
All echocardiographies were analysable for LVEF. At baseline, at least one echocardiographic short-axis plane was analysable for radial strain in all patients.
Similar proportions received CRT defibrillator in intervention vs. control groups; 64 vs. 56% (P = 0.46). One hundred and fourteen patients received a quadripolar LV lead, seven patients received a bipolar LV lead. No device specific algorithms were used.
Two patients withdrew consent during follow-up, one had no CRT implanted, and six patients died during follow-up; remaining 113 patients contributed to the primary outcome analysis (Figure 2).
Mean radiation doses from pre-implant cardiac CT and Rb-PET scans were 4.8 ± 3.1 and 0.9 ± 0.2 mSv.
Procedural data
Data on LV lead positions are shown in Table 2. Number of veins visualized by cardiac CT was comparable between groups, median 4 (range 0–6). In the intervention group, median 3 (range 1–7) veins and 7 (range 1–16) segments were electrically mapped. Mean maximum QLV during mapping was 146 ± 26 ms. In 47 (80%) patients, a stable LV lead position was achieved at the segment with latest electrical activation. Optimal VVd as assessed by RV20, SIM, LV20, LV40, and LV60 in the intervention group was 2 (3%), 24 (41%), 16 (27%), 15 (25%), and 2 (3%), respectively. Interventricular pacing delay optimization resulted in QRSd shortening by median (IQR) 8 (4–19) ms from simultaneous biventricular stimulation in the 35 patients with optimal VVd different from zero.
Procedural data and lead positioning
| Electrical mapping | Intervention group (n = 59) | Control group (n = 62) | P-value |
|---|---|---|---|
| QLV at implantation site (ms) | 137 ± 26 | 135 ± 27 | 0.69 |
| Imaging-guided recommended LV lead position made for all patients | |||
| LV short-axis anterior/lateral/posterior/inferior | 0 (0)/19 (32)/41 (68)/0 (0) | 3 (5)/19 (31)/38 (61)/2 (3) | 0.17 |
| LV long-axis basal/mid-LV/apical | 30 (50)/30 (50)/0 (0) | 31 (50)/31 (50)/0 (0) | 1.00 |
| Imaging-guided LV lead positiona | |||
| Concordant | 16 (30) | 30 (60) | 0.01 |
| Adjacent | 30 (57) | 16 (32) | |
| Remote | 7 (13) | 4 (8) | |
| Scar at pacing site | 4 (8) | 0 (0) | 0.04 |
| RV-lead positiona | |||
| RV short-axis septal/free-wall | 31 (61)/20 (39) | 25 (52)/23 (48) | 0.38 |
| RV long-axis basal/mid-RV/apical | 4 (8)/25 (48)/23 (44) | 0 (0)/27 (56)/21 (44) | 0.13 |
| Procedure time (min) | 104 ± 29 | 85 ± 33 | 0.001 |
| Procedural radiation (mSv) | 13.7 | 10.0 | 0.03 |
| Electrical mapping | Intervention group (n = 59) | Control group (n = 62) | P-value |
|---|---|---|---|
| QLV at implantation site (ms) | 137 ± 26 | 135 ± 27 | 0.69 |
| Imaging-guided recommended LV lead position made for all patients | |||
| LV short-axis anterior/lateral/posterior/inferior | 0 (0)/19 (32)/41 (68)/0 (0) | 3 (5)/19 (31)/38 (61)/2 (3) | 0.17 |
| LV long-axis basal/mid-LV/apical | 30 (50)/30 (50)/0 (0) | 31 (50)/31 (50)/0 (0) | 1.00 |
| Imaging-guided LV lead positiona | |||
| Concordant | 16 (30) | 30 (60) | 0.01 |
| Adjacent | 30 (57) | 16 (32) | |
| Remote | 7 (13) | 4 (8) | |
| Scar at pacing site | 4 (8) | 0 (0) | 0.04 |
| RV-lead positiona | |||
| RV short-axis septal/free-wall | 31 (61)/20 (39) | 25 (52)/23 (48) | 0.38 |
| RV long-axis basal/mid-RV/apical | 4 (8)/25 (48)/23 (44) | 0 (0)/27 (56)/21 (44) | 0.13 |
| Procedure time (min) | 104 ± 29 | 85 ± 33 | 0.001 |
| Procedural radiation (mSv) | 13.7 | 10.0 | 0.03 |
Values are presented as mean ± standard deviation or n (%).
Bold P values <0.05 are statistically significant.
LV, left ventricular; QLV, local LV electrical delay.
Verified by post-implant cardiac computed tomography in 53 patients (intervention group) and 50 patients (control group).
Procedural data and lead positioning
| Electrical mapping | Intervention group (n = 59) | Control group (n = 62) | P-value |
|---|---|---|---|
| QLV at implantation site (ms) | 137 ± 26 | 135 ± 27 | 0.69 |
| Imaging-guided recommended LV lead position made for all patients | |||
| LV short-axis anterior/lateral/posterior/inferior | 0 (0)/19 (32)/41 (68)/0 (0) | 3 (5)/19 (31)/38 (61)/2 (3) | 0.17 |
| LV long-axis basal/mid-LV/apical | 30 (50)/30 (50)/0 (0) | 31 (50)/31 (50)/0 (0) | 1.00 |
| Imaging-guided LV lead positiona | |||
| Concordant | 16 (30) | 30 (60) | 0.01 |
| Adjacent | 30 (57) | 16 (32) | |
| Remote | 7 (13) | 4 (8) | |
| Scar at pacing site | 4 (8) | 0 (0) | 0.04 |
| RV-lead positiona | |||
| RV short-axis septal/free-wall | 31 (61)/20 (39) | 25 (52)/23 (48) | 0.38 |
| RV long-axis basal/mid-RV/apical | 4 (8)/25 (48)/23 (44) | 0 (0)/27 (56)/21 (44) | 0.13 |
| Procedure time (min) | 104 ± 29 | 85 ± 33 | 0.001 |
| Procedural radiation (mSv) | 13.7 | 10.0 | 0.03 |
| Electrical mapping | Intervention group (n = 59) | Control group (n = 62) | P-value |
|---|---|---|---|
| QLV at implantation site (ms) | 137 ± 26 | 135 ± 27 | 0.69 |
| Imaging-guided recommended LV lead position made for all patients | |||
| LV short-axis anterior/lateral/posterior/inferior | 0 (0)/19 (32)/41 (68)/0 (0) | 3 (5)/19 (31)/38 (61)/2 (3) | 0.17 |
| LV long-axis basal/mid-LV/apical | 30 (50)/30 (50)/0 (0) | 31 (50)/31 (50)/0 (0) | 1.00 |
| Imaging-guided LV lead positiona | |||
| Concordant | 16 (30) | 30 (60) | 0.01 |
| Adjacent | 30 (57) | 16 (32) | |
| Remote | 7 (13) | 4 (8) | |
| Scar at pacing site | 4 (8) | 0 (0) | 0.04 |
| RV-lead positiona | |||
| RV short-axis septal/free-wall | 31 (61)/20 (39) | 25 (52)/23 (48) | 0.38 |
| RV long-axis basal/mid-RV/apical | 4 (8)/25 (48)/23 (44) | 0 (0)/27 (56)/21 (44) | 0.13 |
| Procedure time (min) | 104 ± 29 | 85 ± 33 | 0.001 |
| Procedural radiation (mSv) | 13.7 | 10.0 | 0.03 |
Values are presented as mean ± standard deviation or n (%).
Bold P values <0.05 are statistically significant.
LV, left ventricular; QLV, local LV electrical delay.
Verified by post-implant cardiac computed tomography in 53 patients (intervention group) and 50 patients (control group).
In both groups, two patients with permanent atrial fibrillation underwent atrioventricular nodal ablation and one patient with atrial flutter underwent cavotricuspid isthmus block due to insufficient biventricular pacing (<90%) during follow-up. Median (IQR) biventricular pacing percentages at 6-month follow-up was 98 (95–99)% in the intervention group and 99 (97–99)% in the control group, P = 0.08.
Primary outcome
Mean absolute increase in LVEF was higher in the intervention group than in the control group, 11 ± 10 vs. 7 ± 11%, P = 0.03 (Figure 3). In pre-specified multivariate regression analysis, mean difference in LVEF increase between groups was no longer statistically significant 3 [95% confidence interval (CI) −0.4 to 6.4]%, P = 0.09. Intra-class correlation coefficients for intra- and inter-observer agreement on LVEF were 0.88 and 0.90, respectively and coefficient of variability was 29%. Post hoc on-treatment analysis is presented in Supplementary material online.
Change in left ventricular ejection fraction. Absolute percentage change in left ventricular ejection fraction in the intervention and control group from baseline to 6-month follow-up.
Secondary outcome measures
For the composite secondary endpoint of clinical response to CRT, groups did not differ [risk difference −6 (95% CI −24 to 12)% for the intervention group]. No differences between groups were observed for the tertiary outcome measures (Table 3); including LV reverse remodelling (Figure 4) and complication rate. During a median (IQR) follow-up of 16 (9–27) months, no significant difference between groups were observed for the combined endpoint of all-cause mortality or heart failure hospitalization [hazard ratio 0.52 (95% CI 0.23–1.16) for the control group, P = 0.11].
Left ventricular remodelling. Left ventricular remodelling in the intervention and control group from baseline to 6-month follow-up. LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume.
Secondary outcome measures at 6-month follow-up
| Intervention group (n = 54) | Control group (n = 59) | P-value | |
|---|---|---|---|
| All-cause mortality | 5 (8) | 1 (2) | 0.15 |
| Cardiovascular mortality | 2 (3) | 1 (2) | 0.55 |
| Heart failure hospitalization | 5 (8) | 2 (3) | 0.28 |
| NYHA Class I/II/III/IV | 18 (33)/26 (48)/10 (19)/0 (0) | 23 (39)/31 (53)/5 (9)/0 (0) | 0.31 |
| NYHA class improvement | 27 (50) | 36 (61) | 0.32 |
| 6MWT ≥10% increase | 20 (43) | 13 (27) | 0.06 |
| 6MWT absolute increase (m) | 40 ± 70 | 24 ± 64 | 0.06 |
| Clinical response to CRTa | 31 (53) | 35 (58) | 0.69 |
| QoL, MLHFQ change (points) | −12 ± 21 | −13 ± 18 | 0.84 |
| QRS duration (ms) | 138 ± 18 | 138 ± 19 | 0.80 |
| Antero-septal to posterior radial strain delay, change (ms) | −182 ± 322 | −118 ± 331 | 0.30 |
| NT-proBNP change (ng/L) | −1711 ± 3273 | −1000 ± 2195 | 0.18 |
| LVEDV relative change (%) | −16 ± 22 | −18 ± 25 | 0.83 |
| LVESV relative change (%) | −22 ± 33 | −25 ± 36 | 0.71 |
| Intervention group (n = 54) | Control group (n = 59) | P-value | |
|---|---|---|---|
| All-cause mortality | 5 (8) | 1 (2) | 0.15 |
| Cardiovascular mortality | 2 (3) | 1 (2) | 0.55 |
| Heart failure hospitalization | 5 (8) | 2 (3) | 0.28 |
| NYHA Class I/II/III/IV | 18 (33)/26 (48)/10 (19)/0 (0) | 23 (39)/31 (53)/5 (9)/0 (0) | 0.31 |
| NYHA class improvement | 27 (50) | 36 (61) | 0.32 |
| 6MWT ≥10% increase | 20 (43) | 13 (27) | 0.06 |
| 6MWT absolute increase (m) | 40 ± 70 | 24 ± 64 | 0.06 |
| Clinical response to CRTa | 31 (53) | 35 (58) | 0.69 |
| QoL, MLHFQ change (points) | −12 ± 21 | −13 ± 18 | 0.84 |
| QRS duration (ms) | 138 ± 18 | 138 ± 19 | 0.80 |
| Antero-septal to posterior radial strain delay, change (ms) | −182 ± 322 | −118 ± 331 | 0.30 |
| NT-proBNP change (ng/L) | −1711 ± 3273 | −1000 ± 2195 | 0.18 |
| LVEDV relative change (%) | −16 ± 22 | −18 ± 25 | 0.83 |
| LVESV relative change (%) | −22 ± 33 | −25 ± 36 | 0.71 |
Values are presented as mean ± standard deviation or n (%).
CRT, cardiac resynchronization therapy; LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume; MLHFQ, Minnesota Living with Heart Failure Questionnaire; 6MWT, 6-minute walk test; NT-proBNP, N-terminal prohormone of brain natriuretic peptide; NYHA, New York Heart Association; QoL, quality of life.
The composite secondary endpoint of clinical response to CRT was reached if the patient was alive, had not been hospitalized for heart failure, and had improved NYHA class or/and increased 6MWT >10% at 6-month follow-up.
Secondary outcome measures at 6-month follow-up
| Intervention group (n = 54) | Control group (n = 59) | P-value | |
|---|---|---|---|
| All-cause mortality | 5 (8) | 1 (2) | 0.15 |
| Cardiovascular mortality | 2 (3) | 1 (2) | 0.55 |
| Heart failure hospitalization | 5 (8) | 2 (3) | 0.28 |
| NYHA Class I/II/III/IV | 18 (33)/26 (48)/10 (19)/0 (0) | 23 (39)/31 (53)/5 (9)/0 (0) | 0.31 |
| NYHA class improvement | 27 (50) | 36 (61) | 0.32 |
| 6MWT ≥10% increase | 20 (43) | 13 (27) | 0.06 |
| 6MWT absolute increase (m) | 40 ± 70 | 24 ± 64 | 0.06 |
| Clinical response to CRTa | 31 (53) | 35 (58) | 0.69 |
| QoL, MLHFQ change (points) | −12 ± 21 | −13 ± 18 | 0.84 |
| QRS duration (ms) | 138 ± 18 | 138 ± 19 | 0.80 |
| Antero-septal to posterior radial strain delay, change (ms) | −182 ± 322 | −118 ± 331 | 0.30 |
| NT-proBNP change (ng/L) | −1711 ± 3273 | −1000 ± 2195 | 0.18 |
| LVEDV relative change (%) | −16 ± 22 | −18 ± 25 | 0.83 |
| LVESV relative change (%) | −22 ± 33 | −25 ± 36 | 0.71 |
| Intervention group (n = 54) | Control group (n = 59) | P-value | |
|---|---|---|---|
| All-cause mortality | 5 (8) | 1 (2) | 0.15 |
| Cardiovascular mortality | 2 (3) | 1 (2) | 0.55 |
| Heart failure hospitalization | 5 (8) | 2 (3) | 0.28 |
| NYHA Class I/II/III/IV | 18 (33)/26 (48)/10 (19)/0 (0) | 23 (39)/31 (53)/5 (9)/0 (0) | 0.31 |
| NYHA class improvement | 27 (50) | 36 (61) | 0.32 |
| 6MWT ≥10% increase | 20 (43) | 13 (27) | 0.06 |
| 6MWT absolute increase (m) | 40 ± 70 | 24 ± 64 | 0.06 |
| Clinical response to CRTa | 31 (53) | 35 (58) | 0.69 |
| QoL, MLHFQ change (points) | −12 ± 21 | −13 ± 18 | 0.84 |
| QRS duration (ms) | 138 ± 18 | 138 ± 19 | 0.80 |
| Antero-septal to posterior radial strain delay, change (ms) | −182 ± 322 | −118 ± 331 | 0.30 |
| NT-proBNP change (ng/L) | −1711 ± 3273 | −1000 ± 2195 | 0.18 |
| LVEDV relative change (%) | −16 ± 22 | −18 ± 25 | 0.83 |
| LVESV relative change (%) | −22 ± 33 | −25 ± 36 | 0.71 |
Values are presented as mean ± standard deviation or n (%).
CRT, cardiac resynchronization therapy; LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume; MLHFQ, Minnesota Living with Heart Failure Questionnaire; 6MWT, 6-minute walk test; NT-proBNP, N-terminal prohormone of brain natriuretic peptide; NYHA, New York Heart Association; QoL, quality of life.
The composite secondary endpoint of clinical response to CRT was reached if the patient was alive, had not been hospitalized for heart failure, and had improved NYHA class or/and increased 6MWT >10% at 6-month follow-up.
Discussion
In this double-blinded, randomized controlled trial an electrically guided CRT strategy targeting LV lead towards the segment with latest electrical activation combined with post-implant VVd optimization seemed associated with superior improvement in LVEF after 6 months when compared with an imaging-guided strategy. However, the electrically guided strategy was not associated with additional LVEF increase after adjustment for baseline characteristics. Other clinical endpoints after CRT as well as device related complications did not differ significantly between groups.
The ElectroCRT study was designed to show superiority of an electrically guided CRT strategy over a time-consuming and costly imaging-guided strategy, which earlier was found superior to conventional CRT.5–7 Although a larger LVEF increase was observed in the intervention group in unadjusted analysis, this difference was not significant in multivariate analysis, and secondary clinical outcome measures did not differ between groups. Therefore, the electrically guided strategy appear non-inferior to the imaging-guided strategy. Potential explanations may include that the majority of patients in both groups had pre-implant characteristics consistent with a favourable CRT response such as wide LBBB configuration and were medically well-treated with most patients in NYHA class II.17 Furthermore, both individualized strategies for LV lead placement primarily resulted in LV lead positions in non-apical postero-lateral segments, associated with favourable outcome in CRT.3,18 At baseline, patients in the intervention group had slightly larger LV chambers and lower LVEF. This observation occurred by chance despite randomization, but may help explain the trend observed for larger risk of reaching the combined endpoint of all-cause mortality or heart failure hospitalization in the intervention group. However, the ElectroCRT study was not powered for this endpoint, and larger studies are needed to investigate clinical outcomes with this strategy.
Neither the electrically guided nor the imaging-guided approach for LV lead placement was completely successful in achieving optimal LV lead position. A potential reason for this includes limitations in CS anatomy to reach a stable LV lead position at a priori defined optimal pacing sites. Nevertheless, the majority of patients had the LV lead placed concordant with latest mechanically activated segment or concordant with latest electrically activated segment according to allocated group. Furthermore, both groups achieved comparable long QLV at LV implantation site with the majority of LV leads concordant with or adjacent to the imaging-guided optimal segment. These observations also support the comparable outcomes between groups as previous studies have shown increased CRT response rates in patients with long QLV and concordant or adjacent lead positions.6,19
In the intervention group, VVd optimization narrowed QRS only marginally, and we observed no difference in QRS narrowing between groups at 6-month follow-up. These findings may be explained by the similar LV lead placement in both treatment groups. Moreover, the non-negligible variation when measuring QRSd at different VVds may have influenced results.16
This trial documented that implementing procedural electrical mapping of CS branches for targeted LV lead placement is feasible in a large volume CRT centre. From a practical point of view, the electrically guided strategy holds the advantage of being immediately available during implant procedure. The electrically guided strategy was not associated with increased device related complication rate, which in both groups was overall comparable to rates reported in a recent population-based analysis.20 Implant procedure time was longer with the electrically guided strategy, but still within an acceptable timeframe.6 The imaging-guided strategy must be considered costlier and more time-consuming due to acquisition of pre-implant multimodality imaging. Difficulties correlating pre-implant images with fluoroscopy during device implantation may also apply. Furthermore, the imaging-guided strategy in most centres will be applicable only in patients admitted for planned implantation procedures, where access to scanners can be scheduled ahead. The increased procedural radiation dose with procedural electrical mapping equalled 1–2 years of background radiation and was less than mean pre-implant cardiac CT radiation dose. A potential disadvantage of the electrically guided strategy is the potential risk of implanting the LV lead in scarred myocardium. However, this risk is considered equal to the risk with routine CRT implantation, typically performed without pre-implant myocardial perfusion imaging.
Limitations
The trial design is complex, and a single-centre setting was chosen to ensure high study protocol adherence. Therefore, the sample size was moderate, and statistical power is inadequate to assess differences in endpoints such as mortality and heart failure hospitalization. However, strict protocol adherence is imperative in initial testing of new strategies, and we showed that a strategy of electrical mapping can be implemented successfully in an experienced centre.
Left ventricular ejection fraction was chosen as primary endpoint, because it is a universally accepted echocardiographic measure of LV systolic function and because early increase in LVEF predicts longer survival time in CRT patients.17
In the intervention group, we only mapped vein branches suitable for LV lead implantation.
Per protocol, VVd optimization was performed only in the intervention group. We cannot exclude that the effect of CRT had been different in the control group with VVd optimization performed. However, additional QRS shortening in intervention group was only marginal and no clinical long-term effect of routine VVd optimization has been shown.21
We used no independent core lab or event adjudicating committee. However, all imaging analyses and endpoint adjudication were performed before breaking the randomization code.
Not all patients underwent pre-implant cardiac CT. This limitation illustrates that imaging-guided LV lead placement is not feasible for all patients, favouring the electrically guided strategy.
Conclusion
In patients undergoing CRT, an electrically guided implantation strategy targeting LV lead towards the segment with latest electrical activation combined with post-implant VVd optimization seems associated with superior improvement in LVEF when compared with an imaging-guided LV lead implantation strategy, although not significant in adjusted analysis. As symptomatic improvement, LV reverse remodelling, and device related complication rates did not differ between treatment groups, the electrically guided strategy appear non-inferior to imaging-guided strategy. Larger multicentre trials are warranted to evaluate clinical benefit of targeting the LV lead towards latest local electrical activation in CRT.
Supplementary material
Supplementary material is available at Europace online.
Acknowledgements
The authors would like to thank Kirsten Andersen, Sonja Runge, and Rita Møhl for their expertise on pacemaker programming, Kamilla Bech Pedersen for technical assistance with cardiac CT, and Heidi Grønhøi and Henriette Holmberg for eminent study coordination.
Funding
This work was supported by the Danish Heart Foundation [14-R97-A5149-22865 and 15-R99-A5878-22937 to C.S.]; and the Health Research Fund of Central Denmark Region [A119 to C.S.].
Conflicts of interest: C.S.: conference attendance supported by Abbott. J.M.J.: speakers honorarium from Bracco Imaging. C.G.: lecturing fees from Novartis. Consulting fees from Abbott, Medtronic, and Boehringer Ingelheim. Proctoring fees from Biosense Webster. H.K.J.: supported by the Novo Nordisk Foundation (NNF18OC0031258). Lecturing fees from Abbott and Biosense Webster. J.C.O.: supported by the Novo Nordisk Foundation (NNF16OC0018658). Received an unrestricted institutional grant from Abbott. M.B.K., A.S., J.K., B.L.N., D.B.F., and K.B are declared no conflict of interest.
