A novel method combining mechanical contraction and electrical activation to recommend left ventricular lead placement to improve CRT response

Study Population

A total of 76 patients with systolic HF were retrospectively enrolled in this study at The First Affiliated Hospital of Nanjing Medical University from October 2011 to October 2021. All the patients were in sinus rhythm with intraventricular conduction delay (QRS duration ≥120 ms), LV ejection fraction (LVEF) ≤35%, New York Heart Association (NYHA) functional class II to IV symptoms, and optimal medical therapy at least 3 months before CRT.

All the patients underwent resting gated SPECT MPI, echocardiography, standard 12-lead electrocardiogram (ECG), and NYHA function classification at baseline. The actual LV lead position was identified according to pre-implantation coronary venous angiograms and post-implantation LV lead fluoroscopy images. Echocardiography and NYHA function classification were reevaluated 6 months after CRT. A clinical response to CRT was defined as a ≥1 NYHA class improvement at 6 months.¹⁶ The study was approved by the institutional review board, and written informed consent was obtained from all patients.

Evaluation of LV Function by Echocardiography

Transthoracic echocardiography was performed on all patients by an ultrasound specialist who was blinded to the study before and 6 months after CRT. LV end-systolic volume (LVESV), LV end-diastolic volume (LVEDV), and LVEF were recorded using the 2-dimensional modified biplane Simpson method.⁹ Decrease of LVESV ≥15% on follow-up echocardiography was considered as volumetric response to CRT.⁶

Evaluation of Segments with the Latest Mechanical Contraction from SPECT MPI

We performed resting ECG-gated SPECT MPI before CRT implantation, following a protocol similar to a previously employed method.⁹ 60-90 minutes after 25-30 mCi of Tc-99m sestamibi injection, SPECT was performed on a dual-headed camera (CardioMD, Philips Medical Systems) with a standard resting protocol. Image reconstruction and reorientation were completed by Emory Reconstruction Toolbox (ERToolbox; Atlanta, GA).

We reoriented the gated and ungated transaxial images into short-axis images, and submitted these results to the phase analysis to measure LVMD¹⁷ and segments of LMC¹⁸ and to assess regional myocardial perfusion. ¹⁹ Myocardial scar was defined as LV samples with <50% of the maximum uptake and a scarred segment was identified as the segment where over 50% of LV samples in the region was myocardial scar. The percentage of tracer uptake was displayed on polar map using a 17-segment model and mean phase angle of each segment was displayed on the 17-segment polar map, like described in our previous study.⁹ Excluding 5 apical segments, septal segments, and scarred segments, the LMC segments were defined as the top 4 phase angles which were within 10 degrees of the maximum phase angle. To simplify, the basal and mid anterolateral, lateral, posterolateral segments were grouped together as the lateral wall. The basal anterior and mid anterior segments were grouped together as the anterior wall. The basal inferior and mid inferior segments were grouped together as the inferior wall. Phase standard deviation (PSD, unit: degree) and 95% bandwidth of phase histogram (PHB, unit: degree) were used to assess global LVMD.

Evaluation of the LV Site with the Latest Electrical Activation by VCG

A standard 12-lead ECG was recorded at baseline. The digital PDF/XML ECG files with vector graphics were used to extract the original digital ECG-signal. We developed a tool implemented in Python that semi-automatically synthesizes digital ECG signals into a VCG. The Kors transformation matrix was used to transform the 12-lead ECG to VCG.²⁰ The direction of the maximal T vector was considered as the LEA site.¹⁴ Excluding LV apex and septal segments, one of the three candidate sites (anterior, lateral, and inferior) from VCG was recommended (Figure 1).

• Download figure
• Open in new tab

Figure 1.

Patient examples of how to recommend the left ventricular lead placement. (A) The LMC segment was in the anterior wall and overlapped with the LEA site, so the anterior wall was recommended. (B) LMC segment was in the lateral wall and LEA site in the anterior wall, so the lateral wall was recommended. (C) LMC segment was in the anterior wall and LEA site in the inferior wall, so the lateral wall was recommended.

Recommendation of the LV Lead placement segment

After excluding septal segments, LV apex, and segments with more than 50% scar, the segments for LV lead placement were recommended as follows: (1) the LMC viable segments that overlapped with the LEA site(Figure 1A);(2) the LMC viable segments adjacent to the LEA site(Figure 1B);(3) If no segment met either of the above, the LV lateral wall was recommended(Figure 1C).

Patients with LV lead located in the recommended site were assigned to the recommended group, and those located in the non-recommended site were assigned to the non-recommended group.

Cardiac Resynchronization Therapy implantation

All patients were implanted with a biventricular pacemaker according to standard procedures.²¹ The LV lead was put into the coronary sinus (CS) system and stably positioned into the lateral, posterior, anterolateral, or posterolateral coronary vein by experienced electrophysiologists. The atrial lead was implanted in the right atrial appendage, and the right ventricular (RV) lead was placed at the RV apex or septum.

Identification of the Actual LV Lead Position

A method similar to the method used in the Multicenter Automatic Defibrillator Implantation Trial–Cardiac Resynchronization Therapy (MADIT-CRT) study⁵ was used for identification of the actual LV lead position. A preimplantation coronary venous angiogram in left anterior oblique 20° to 45° was performed for every patient, and post-implantation fluoroscopic images or CT venography were stored. The left anterior oblique view was used to divide the LV wall into 4 equal parts: anterior, anterolateral, posterolateral, and posterior along the short axis of the heart. To be consistent with SPECT and VCG segmentation, the anterolateral and posterolateral segments were included in the lateral segments. The final LV lead position was categorized as anterior, lateral, or posterior.

Statistical Analysis

Continuous variables were expressed as mean ± standard deviation (SD). Categorical data are summarized as frequencies and percentages. Comparisons between groups were performed by student’s t-test for parametric variable, Chi-square or Fisher exact tests for categorical variables. The binary logistic regression (P <.05 in the univariate regression were included and variables with P >.05 in the multivariate regression were excluded) were performed to assess the independent predictors of volumetric response to CRT. Significance was defined as P <.05 using two-tailed analysis. Statistical analysis was performed with IBM SPSS Statistics (IBM, Chicago, IL, USA) version 25.0.

Patient Characteristics

The baseline characteristics of the 76 enrolled patients are summarized in Table 1. Patients had severely depressed LV function, with a mean LVEF of 27.0±5.0%. Mean LVEDV was 286.3±88.8 ml, and mean LVESV was 211.5±77.3 ml. Forty-one patients (53.9%) had a LV lead position concordant with the LMC segment from SPECT, and 13 patients (17.1%) had a LV lead position concordant with the LEA site by VCG. Fifty-four patients (71.1%) had a recommended LV lead position, whereas 22 (28.9%) had an LV lead positioned in the non-recommended segment. There were no significant differences in baseline data between the two groups, except that, in the recommended group, less patients had diabetes, more patients had a LV lead position concordant with the LMC segment, and more patients had a LV lead in the lateral wall.

Echocardiographic Measurements at Follow-Up

At 6-month follow-up, echocardiographic parameters of patients in the recommended group improved significantly; LVEDV decreased from 290.7±91.4 ml to 219.8±106.2 ml, LVESV from 215.3±78.4 ml to 142.3±91.9 ml, and LVEF increased from 26.7±4.8% to 39.5±12.6% (all p<.001). However, patients in the non-recommended group did not have a significant reduction in LV volumes (LVEDV from 275.5±83.1 ml to 245.0±101.8 ml and LVESV from 202.3±75.5 ml to 173.1±91.9 ml, both p = NS) and no improvement in LVEF was observed (from 27.7±5.6% to 33.2±11.6%, p = NS). The relative change in LVEDV, LVESV, and LVEF showed significant difference between the recommended group and non-recommended group (25.5±19.6% vs. 13.3±20.1%, p= .017; 36.6±24.7% vs. 18.0±27.5%, p= .005; 48.4±40.6% vs. 18.5±28.0%, p= .002) (Figure 2).

• Download figure
• Open in new tab

Figure 2.

Comparison of echocardiographic measurements between recommended group and non-recommended group. (A-B-C): LVEDV, LVESV and LVEF change between baseline and follow up. (D-E-F): The relative change of LVEDV, LVESV and LVEF between the recommended and non-recommended group. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, no statistically significant.

Volumetric Response and Clinical Response

Volumetric Response

Six months after CRT implantation, 74.1% (40 of 54) of the patients in the recommended group were responders, while 36.4% (8 of 22) of the patients in the non-recommended group were responders (p = .002) (Figure 3A).

• Download figure
• Open in new tab

Figure 3.

CRT response rates in the recommended and non-recommended groups. (A) CRT volumetric response rates in patients with LV leads in the recommended segment (n = 54) and patients with LV leads in the non-recommended segment (n = 22) (74.1% vs. 36.4%, P = .002). (B) CRT clinical response rates in patients with LV leads in the recommended segment (n = 54) and patients whose LV leads were in the non-recommended segments (n = 22) (72.2% vs. 40.9%, P = .01).

We validated our recommendation method in subgroup analyses. The actual LV lead was located in the lateral wall in 62 patients and non-lateral in 14 patients, the response rate showed no significance difference between the two groups (67.7% vs. 42.9%, p=.081) (Figure 4A). Further analysis of patients with a LV lead located in the lateral wall found that the response rate was 77.6% (38 of 49) in patients with a LV lead located in the recommended segment, whereas 30.8% (4 of 13) in patients with a non-recommended LV lead position (p= .004) (Figure 4B). Similar results were found in patients with QRSd ≥150ms and in patients with LBBB morphology; the response rate was significantly higher in patients with a LV lead located in the recommended segment than in patients with a non-recommended LV lead position (83.0% vs. 47.1%, p =.011; 82.6% vs. 50.0%, p =.025) (Figure 4C and 4D).

• Download figure
• Open in new tab

Figure 4.

Subgroup analyses of CRT volumetric response. (A) CRT volumetric response rates in patients with LV leads in the lateral segment (n = 62) and patients with LV leads in the non-lateral segments (n = 14) (67.7% vs. 42.9%, P = .081). (B) CRT volumetric response rates in patients with LV leads in the lateral segment and meeting our recommended criteria (n=49) and patients with LV leads in the lateral segment and not meeting our recommended criteria (n=13) (77.6% vs. 30.8%, P = .004). (C) CRT volumetric response rates in patients whose QRSd≥150ms and meeting our recommended criteria (n=47) and patients whose QRSd≥150ms and not meeting our recommended criteria (n=17) (83.0% vs. 47.1%, P = .011). (D) CRT volumetric responses rate in patients with LBBB morphology and meeting our recommended criteria (n=46) and patients with LBBB morphology and not meeting our recommended criteria (n=16) (82.6% vs. 50.0%, P = .025).

Univariable and Multivariable Regression Analyses

Univariable and multivariable logistic regression analyses were used to determine which parameters predict the CRT volumetric response. In the univariate regression analysis, presence of diabetes, QRSd, LBBB morphology, scar burden, and recommended LV lead position were associated with the volumetric response. In multivariable regression analyses, LV lead placement in the recommended segments versus non-recommended position remained an independent predictor of the volumetric response (odds ratio 7.326; 95% confidence interval 1.527-35.144, p =.013) (Table 2).

Attempts for Improved CRT Response

LV lead position is known to be an important factor in determining response to CRT. Current practice for CRT is to place the LV lead at the lateral or posterolateral wall based on the benefit shown in early hemodynamic studies and the observation that the latest activated segment predominates at these sites. However, it has been controversial to direct the LV lead to the lateral or posterolateral wall. The MADIT-CRT study demonstrated that the benefit from CRT was similar for LV leads positioned along the anterior, lateral, or posterior wall.⁵ In the present study, we also found that the response rate showed no significant difference in patients with a LV lead located on the lateral wall compared with on the non-lateral wall (67.7% vs. 42.9%, p=.081). In addition, we found a volumetric response rate of only 30.8% in patients with a LV lead located on the lateral wall while the lateral wall was not our recommended position, further suggesting that the lateral wall may not be the optimal location for some patients.

QRS duration and morphology are used to select candidates for CRT. It has a class IA indication for symptomatic heart failure patients with sinus rhythm and a QRSd≥150 ms and LBBB morphology and with LVEF≤35% despite optimized medical treatments according to the 2021 ESC Guidelines on CRT.²² Several studies showed that patients with LBBB and prolonged QRS duration (QRSd≥150 ms) were more likely to be echocardiographic responders and were associated with favorable clinical outcomes.^23,24 In subgroup analysis of the current study, 62 patients had LBBB morphology, 64 patients had a QRSd ≥150ms, and 57 patients had LBBB morphology with QRSd ≥150ms. We found that in patients with LBBB morphology, the response rate was higher when the LV lead was placed in the recommended position than when it was placed in the non-recommended position (82.6% vs. 50.0%, p =.025). The same result was found in patients with QRSd≥150ms (83.0% vs. 47.1%, p =.011). Furthermore, in patients with LBBB morphology and QRSd ≥150ms, the response rate was 88.1% (37 of 42) when the LV lead was placed in the recommended position, and was significantly higher than that when placed in the non-recommended position (88.1% vs. 53.3%, p=.014). This seems that our recommended LV lead position could further improve CRT response after strict guideline-directed patient selection.

A previous study demonstrated that implanting LV lead in the scarred region had a detrimental impact on LV reverse remodeling and increased clinical events.⁹ In the present study, we excluded scarred segments and found overall myocardial scar burden quantified by SPECT MPI predicted the volumetric response to CRT, suggesting that reverse remodeling requires viable myocardium adjacent to the LV lead.

The latest mechanical contraction segment and electrical activation site

Several clinical studies have claimed that LV lead placement directs toward the segment of LMC improves the CRT response, including the TARGET,⁶ the STARTER,²⁵ and the Imaging CRT study.²⁶ These studies employed various imaging modalities, such as speckle tracking echocardiography, cardiac computed tomography (CT) and SPECT MPI, to guide the LV lead placement toward the LMC segment. Pacing remote from the LEA site may diminish the CRT response, as supported by studies utilizing non-invasive three-dimensional electrical activation mapping, a technique used to assess the distance of pacing site and the position of the LEA site. It was found that LV pacing close to the site of LEA provides improved volumetric response to CRT.¹¹ VCG offers a more comprehensive and accurate assessment of cardiac electrical activity by capturing both the direction and magnitude of the cardiac electrical vectors. VCG can be used to measure the LEA site of the heart. The segment of LMC does not always overlap with the site of LEA. In this study, we advocate a novel approach for LV lead placement, integrating information from the segment with the LMC and the site with the LEA. Our findings indicate that pacing in the recommended LV lead segments is significantly associated with an increased CRT volumetric response.

Although the recent studies of novel CRT by conduction system pacing showed better efficacy of cardiac function and clinical outcomes compared to the traditional biventricular pacing,^27,28,29 most of the enrolled patients were class one CRT recommendation with typical LBBB. To our best knowledge, CRT delivered by conduction system pacing was referred as at most IIa level recommendation at present.³⁰ This new integration of LMC and LEA may be very useful to improve traditional biventricular pacing CRT, especially for non-LBBB patients.

Limitations

This retrospective observational study had a limited sample size, and data collection occurred between 2011 and 2021. It’s worth noting that the guidelines for CRT have evolved, and optimal medical therapy has incorporated additional drug classes. Subgroup analyses were conducted, stratified by ECG morphology and QRS duration; however, caution is advised in interpreting the results due to the relatively low number of non-LBBB patients and those with QRS duration <150ms in this cohort. Larger prospective multicenter studies are warranted to provide more robust insights.