Skip to main content
NCBI home page
Europace logoLink to Europace
. 2016 Feb 22;17(Suppl 5):v1–v2. doi: 10.1093/europace/euv324

Young Investigators Competition

PMCID: PMC4892103
Europace. 2016 Feb 22;17(Suppl 5):v1–v2.

1: Left ventricular lead position, mechanical activation and myocardial scar in relation to the clinical outcome of cardiac resynchronisation therapy: the role of feature-tracking and contrast-enhanced cardiovascular magnetic resonance

RJ Taylor 1, F Umar 2, B Stegemann 3, F Leyva 4

Abstract

Introduction: Late mechanical activation (LMA) and myocardial viability have been proposed as targets for left ventricular (LV) pacing during cardiac resynchronisation therapy (CRT). Feature-tracking (FT) cardiovascular magnetic resonance (CMR), the CMR equivalent of speckle-tracking echocardiography, has been validated against the gold-standard of CMR-tagging for the assessment of myocardial deformation. Late gadolinium enhancement (LGE) CMR is the gold-standard for myocardial scar imaging. We hypothesised that deployment of the LV lead over non-scarred segments with LMA, assessed using LGE-CMR and FT-CMR, portends a better outcome from CRT.

Methods: 89 patients who underwent successful CRT device implantation were recruited through a dedicated heart failure (HF) device clinic from a single centre. Patients required a pre-implantation CMR scan for inclusion. FT-CMR of the basal, mid and apical short axis slices was used to determine the LMA segment; that with the latest peak systolic circumferential strain within the imaged cardiac cycle. Only the 11 free wall segments were considered. The presence or absence of myocardial scar was recorded in each segment based on visual assessment of LGE-CMR short axis slices. The LV lead tip was mapped to a specific segment based on fluoroscopic acquisitions in the right anterior (RAO) and left anterior (LAO) projections at the time of implantation. Lead positions were classified into concordant (no scar and LMA) or non-concordant (scar and/or earlier-activated segments).

Results: 89 patients (age: 66.8 ± 10.8 yrs, LVEF: 23.1 ± 9.9%, QRS: 149.0 ± 25.7 ms) were followed up for a median of 4.2 years. LV reverse remodelling (LVRR) occurred in 68% and 24% of patients with concordant and non-concordant LV lead positions respectively (p < 0.001). LMA in the paced segment was the strongest predictor of LVRR (OR: 18.9; 95%C.I. 3.84-84.31). LV lead concordance predicted cardiac mortality (Log Rank [LR] χ2 = 7.32, p= 0.007; hazard ratio [HR]: 0.27, 95% C.I. 0.12–0.62) (Figure 1) and cardiac mortality or hospitalisations for HF (LR χ2: 8.31, p < 0.004; HR: 0.26, 95% C.I. 0.12–0.58). 'No scar' in the paced segment predicted cardiac mortality (HR: 0.24, 95% C.I. 0.11–0.52) and cardiac mortality or HF hospitalisations (HR: 0.24, 95% C.I. 0.12–0.49). LMA in the paced segments did not predict these endpoints.

Conclusion: A LV lead position over segments with LMA and no scar is associated with a better clinical outcome and LVRR after CRT. Clinical events were primarily related to scar, whereas LVRR was primarily related to LMA. These findings support the combined use of LGE-CMR and FT-CMR in guiding LV lead deployment.

graphic file with name euv32401.jpg

Figure 1 Cardiovascular Mortality after CRT according to LV lead position.

Europace. 2016 Feb 22;17(Suppl 5):v1–v2.

2: Does the haemodynamic improvement of biventricular pacing truly arise from cardiac resynchronisation? quantifying the contribution of av and vv adjustment

SMA Sohaib 1, I Wright 2, E Lim 2, M Koawing 2, PB Lim 2, P Moore 2, N Linton 1, DC Lefroy 2, DW Davies 2, NS Peters 1, P Kanagaratnam 2, DP Francis 1, ZI Whinnett 1

Abstract

Background: When cardiac resynchronisation therapy (CRT) is instituted, the atrio-ventricular (AV) delay must be shortened to enable ventricular capture. An important component of CRT benefit may thus come from optimisation of AV delay rather than ventricular resynchronisation. The implications of optimising the AV delay in heart failure without the confounder of ventricular pacing are unknown. His bundle pacing, however, may offer a model to allow us to understand this.

Similarly, when VV delay is adjusted, with the aim of improving ventricular synchrony, there are inevitably changes to the AV delay either between the right ventricle (A-RV) or left ventricle (A-LV). The consequences of using different conventions for keeping the AV delay constant during VV optimisation is unclear.

Methods: An algorithm measuring mean systolic blood pressure (SBP) changes from multiple alternations between reference and tested pacing settings allowed us to measure with high precision small changes in haemodynamics required to investigate these questions. The first protocol investigated patients with existing CRT with underlying broad QRS to separate the influence of AV delay changes during VV optimisation by using four conventions for keeping the AV delay constant (A-LV constant, A-RV constant, time to 1st-activated ventricle constant, and time to 2nd-activated ventricle constant). The second protocol investigated AV optimisation without the confounder of resynchronisation by measuring the haemodynamic benefit of AV optimised His bundle pacing in patients with heart failure, long PR, and narrow QRS (or RBBB). 40 patients were recruited (24 for VV optimisation, 16 for His pacing).

Results: During VV adjustment at AV 120ms, time to first ventricular lead (left or right) was the overwhelming determinant of haemodynamic response (13.8mmHg at ± 80ms from VV 0, p < 0.001) with no significant effect of time to second lead (0.5mmHg, p = 0.50), p < 0.001 for difference.

We successfully delivered temporary His bundle pacing in 14/16 patients (mean PR 254 ± 62ms, mean QRS 119 ± 17ms when narrow (n = 13), and 156 ± 18ms in RBBB (n = 3)). We performed AV optimisation using invasive SBP obtaining parabolic responses (mean R2 = 0.90). The mean increment in SBP compared to intrinsic ventricular conduction was 4.1 ± 3.8mmHg (p = 0.003). This was not significant with AV optimised RV pacing (0.9 mmHg, p = 0.49). The QRS widened with RV pacing (149 ± 29ms, p = 0.02) but not with His (123 ± 22ms, p = 0.77).

Conclusion: AV delay adjustment makes an important contribution to the CRT effect. Unconsidered changes in AV delay can explain most of the haemodynamic changes reported by some centres during VV optimisation. Following implant it may be wise to leave the VV at 0, and focus on AV optimisation. The ∼4mmHg increase in SBP seen here with AV optimised His pacing represents ∼60% of the benefit seen in studies of CRT in sinus rhythm and LBBB and may offer a novel approach to pacing for heart failure.

graphic file with name euv32402.jpg

Europace. 2016 Feb 22;17(Suppl 5):v1–v2.

3: Differential relationship of electrical delay with endocardial and epicardial left ventricular leads for cardiac resynchronisation therapy

JHP Gamble 1, R Khiani 2, N Herring 1, PWX Foley 1, MR Ginks 1, K Rajappan 1, Y Bashir 1, TR Betts 1

Abstract

Introduction: Choosing the position of the left ventricular lead based on electrical delay has been suggested to improve acute haemodynamic response (AHR) to CRT and clinical outcomes. The relationship of electrical delay, generally expressed as Q-LV time, to AHR may be different with endocardial and epicardial pacing.

Methods: Patients were recruited from two groups. The endocardial group were having endocardial left ventricular (LV) lead placement across the interventricular septum, due to failed or ineffective CRT via the coronary sinus (CS). The epicardial group had conventional LV lead placement into a branch of the CS.

AHR was recorded using a pressure wire as change in LV dp/dt max from a baseline of DDD-RV pacing to DDD-Biventricular (BiV) pacing over 30s, at multiple pacing sites and quadripolar lead cathodes in each patient.

We measured QLV at each site as time from surface ECG Q wave to LV activation on EGM during RV pacing, as well as surface ECG QRS width during pacing.

Multilevel modelling with random effects was used to compare the relationship of AHR and electrical delay.

Results: 9 endocardial and 16 epicardial patients were recruited, age 66 ± 15, 23 male (92%), 15 (60%) ischaemic, QRSD 153 ± 21ms, EF 28 ± 7%. AHR was assessed at a total of 195 sites, at a mean of 6 sites per patient.

QRS width was not significantly related to AHR in either group or overall (overall trend of 1.6% decrease in AHR per 20ms increase in QRSW, 95% CI -2 to 0.6%, p = 0.77).

QLV was significantly related to AHR in the endocardial pacing group (4.35% increase in AHR per 20ms increase in QLV, 95%CI 2.3 to 6.3%, p < 0.001), but in the epicardial pacing group there was not a significant relationship across the range of QLV values (p = 0.77 of no difference).

A cut-off value of QLV > 100ms was predictive of response in both groups, associated with an 8% increase in AHR in the epicardial group (p = 0.02) and a 1.8% increase in AHR in the epicardial group relative to QLV < 100ms (p = 0.04).

graphic file with name euv32403.jpg

Figure Mean AHR by QLV quartiles and pacing aspect. Error bars are 1 standard error. Quartiles of QLV were similar in the two groups, with a median QLV of 97ms.

Conclusions: There was a clear effect of QLV on AHR to BiV pacing in the patients paced endocardially, but this relationship was less robust in epicardial pacing via the coronary sinus. This may be due to faster conduction through the endocardial layers, resulting in a greater differential in response to pacing at longer electrical delays.

These results also suggest that there may be limited benefit from further increases in QLV over ∼100ms for pacing via the coronary sinus, and that choosing the site with the longest electrical delay is a reasonable strategy for endocardial pacing site selection.

Europace. 2016 Feb 22;17(Suppl 5):v1–v2.

4: Characterisation of the persistent af substrate through the assessment of electrophysiologic parameters in the organised vs. disorganised rhythm

NA Qureshi 1, SJ Kim 2, C Cantwell 1, R Ali 1, C Roney 1, M Shun-Shin 1, FS Ng 1, I Wright 1, E Lim 1, D Lefroy 1, Z Whinnett 1, N Linton 1, P Kanagaratnam 1, NS Peters 1, PB Lim 3

Abstract

Background: The complexities of mapping spatiotemporal variations in AF limit the characterisation of the AF substrate, and hinder the understanding of its functional electrophysiological (EP) determinants. Electrogram (EGM) fractionation and voltage are thought to be representative of the underlying AF substrate. Fractionated atrial EGMs and more recently, low voltage areas in AF, have been targeted as part of substrate modification. However, current 3D mapping technology only allows for AF voltage maps to be created from single-point EGM acquisitions. We systematically characterized voltages in AF over varying sampling windows, and assessedthe regional EP relationship between AF and the paced-rhythm (PR) in patients with persistent AF.

Methods: Pts undergoing AF ablation underwent electroanatomical mapping (EAM) using a 20-pole spiral catheter. Voltage and fractionation mapping were performed in AF over various left atrial (LA) sites over 8s, and following DCCV, voltage and activation mapping during LA pacing (300/600ms) were performed over the same sites. EGM deflections and scores were calculated using an custom EGM analysis tool.

Results: EAM data were obtained from 15 pts (age 65 ± 11y, LA 41 ± 7mm, CHADS2VASC2 2.8 (0-6)).160,200 AF cycles (AFCY) from 3560 electrode positions were utilised for voltage analysis. The cycle-to-cycle AF voltage (single peak-to-peak measurement) variability was large, with a coefficient of variation of 70%, improving with increasing number of AFCY sampled [10 AFCY (2s) – 18%, 20 AFCY (4s) – 12% and 40 AFCY (8s) −9%]. The mean AF peak-to-peak (AF-V) begins to approximate to the 8s index mean AF-V after 4s (Fig1). The % of sub-sampled single-point EGMs where the mean AF-V deviated from the 8s index mean AF-V by >10% increases as the EGM sampling duration shortens (Fig2&3). At each of the 35 LA sites mapped, the mean AF-V over 4s correlated with the mean paced voltages (P-V) (300 and 600ms) (R2 = 0.5114 p < 0.0001 and R2 = 0.4742 p < 0.0001 respectively)(Fig4). Higher AF-V and P-V (300/600ms) correlated to increasing % of fractionated EGMs (%EGMs with NavX CFE mean score <80ms) at each site [R2(AF) = 0.4310 p < 0.0001, R2(300ms) = 0.4310 p < 0.001 and R2(600ms) = 0.498; p < 0.001] (Fig. 5). There were poor correlations between conduction velocity (CV) at both 300 and 600ms LA pacing, with mean AF-V, P-V and % fractionated EGMs.

Conclusion: AF voltages maps using existing 3D mapping systems are inaccurate. Voltage in AF is highly variable but increasing its sampling window allows for meaningful spatial differentiation, and more precise characterisation of the AF substrate. There was an unexpected inverse relationship between AF-V and P-V; and EGM fractionation, suggesting a predominant functional component of EGM fractionation in AF. The EP substrate of AF can be interrogated in PR, and could provide crucial insights into the underlying mechanisms of AF.

graphic file with name euv32404.jpg

Europace. 2016 Feb 22;17(Suppl 5):v1–v2.

5: Targeting cyclical highest dominant frequency in the ablation of persistent atrial fibrillation

GS Chu 1, X Li 1, FJ Vanheusden 1, TP Almeida 1, JL Salinet 1, N Dastagir 1, SS Varanasi 1, SH Chin 1, S Siddiqui 1, SH Man 1, PJ Stafford 2, AJ Sandilands 2, FS Schlindwein 1, GA Ng 1

Abstract

Introduction: Atrial areas of highest dominant frequency (HDF) exhibit recurrent spatio-temporal patterns, suggesting a potential role in driving persistent AF (persAF). We aimed to explore this mechanism, in particular hypothesizing that ablating these regions may influence outcome.

Method: Ten patients underwent a novel global simultaneous HDF-targeting protocol for first time persAF ablation. Left and right atria (LA & RA) were mapped using 2 non-contact arrays (NCAs) concurrently (Fig. 1a; Ensite & Velocity, St Jude Medical). 5 minutes AF signals were recorded simultaneously using both NCAs at baseline (and also after HDF ablation).

A 30-second data segment and 2048 node geometry were exported into MATLAB (Mathworks) from the LA array, and target HDF areas were identified (Fig. 1b). LA AF cycle length (AFCL) was measured prior to the ablation of each HDF region, with a ±10 ms change considered significant. Finally, PVI was performed, followed by cardioversion if required. AF terminating to an organized LA rhythm was considered as the endpoint.

Results: Of 10 patients, 4 terminated with HDF-targeted ablation before PVI. Marked inter-patient variability was seen in LA HDF activity (median 2329, range 531–5274). Initial AFCL was correlated with the inverse LA mean DF (R2 = 0.77, p < 0.0001), and termination (p < 0.05, Fisher's exact test). Ablating a higher proportion of the HDFs increased termination likelihood (mean ± SEM: 27.5 ± 1.5% vs 21.6 ± 4.7%, p = 0.10). However, ablation extent was similar (1539 ± 332 vs 1310 ± 310 mm2, p = 0.83).

In the 6 patients remaining in AF after LA HDF ablation, only one had a positive LA-RA gradient (Fig. 1c). LA and RA mean DF ablation responses were not always concordant (Fig. 1d).

The 51 ablated LA HDF regions successfully targeted higher HDF activity (mean 1.81 ± 0.33 vs 1.22 ± 0.20 HDFs/node for ablated versus non-ablated LA, p < 0.01), but this did not predict AFCL response (ANOVA p = 0.52). Baseline regional mean DF and OI were significantly associated (r = −0.07 and 0.009 respectively, p < 0.0001 for both on testing for linear contrast). The analysis of 5 minute data showed similar outcomes.

At 4 months, 7 patients (including all who terminated during ablation) remain in sinus rhythm.

Conclusion: Targeting LA cyclical HDF regions alone can terminate persAF, but may paradoxically be masking more clinically significant (and more organized) lower DF areas, and may be inadequate if the RA is also active. Future refinement utilizing mean DF and OI may improve specificity, for which 30s of AF data may be sufficient.

graphic file with name euv32405.jpg

Figure 1 (A) Fluoroscopy with LA and RA NCAs (asterisks) and ablation catheter (arrow); (B) Example LA geometry from MATLAB in a patient who terminated during HDF ablation. Colour bar represents HDF activity, legend shows lesion effect on AFCL, triangles are the centres of HDF regions over 30s; (C) Baseline LA-RA gradients in unsuccessful cases; (D) LA vs RA mean DF response to HDF ablation.

Europace. 2016 Feb 22;17(Suppl 5):v1–v2.

6: Feasibility of fully mr-guided ablation with active tracking: from pre-clinical to clinical application

H Chubb 1, J Harrison 1, SE Williams 1, J Whitaker 1, S Weiss 2, S Krueger 2, G Stenzel 3, T Schaeffter 1, R Razavi 1, M O'Neill 1

Abstract

Introduction: MR-guided electrophysiology (MR-EP) has the potential to improve catheter navigation, to visualize arrhythmia substrate and real-time ablation injury and to avoid ionizing radiation. This study investigated the feasibility of an actively-tracked, fully MR-guided, electroanatomical mapping and ablation system in the ablation of atrial flutter in humans.

Methods: The MR-EP system integrated a Philips 1.5T Achieva scanner, an EP recording system (Horizon System, Imricor) and a real-time image guidance platform (iSuite, Philips). Active tracking of the ablation catheter (Vision Catheter, Imricor) was achieved through the integration of two receiver coils at the tip of the catheter, receiving a 10Hz tracking signal with theoretical tracking accuracy <0.8mm. The catheter position was projected in real-time on a multiplanar reconstruction or anatomical shell (Figure 1).

graphic file with name euv32406.jpg

Figure 1 Screenshot of iSuite system during human CTI ablation. Multiplanar reconstruction (three black and white panels) with isochrones projected onto RA shell (bottom right).

A pre-clinical study was performed in 5 pigs, with 3D RA shells segmented from a whole-heart MR sequence, enabling an electroanatomical mapping system and subsequent linear intercaval ablation. Electrical, MR imaging and histological validation of ablation lesions was performed.

Ten patients with typical right atrial flutter were put forward for fully MR-guided ablation. Pre-ablation isochronal activation maps were created and RF ablation (35-45W, 40-60s) of the cavotricuspid isthmus (CTI) performed under active MR-guidance, with brief cine MR sequences to confirm catheter position. Post-ablation, activation maps were repeated and immediate T2 weighted imaging of the lesions performed. Repeat MRI was performed at 3 months.

Results: In pre-clinical studies, irrigated RF ablation was successfully performed in all animals and activation maps were created before and after ablation (34 ± 5 points; 20 ± 8 minutes). Correct lesion localization was confirmed on immediate MR imaging and histologically validated.

In the clinical phase, there were no intraprocedural complications. Nine patients underwent MR guided ablation, with non-investigational equipment failure necessitating conventional ablation in the final patient. Two patient required completion of CTI conduction block, using conventional 8mm Blazer catheter under fluoroscopy, due to CTI morphology. At three month follow-up there have been two recurrences. Mean total procedure time (including pre-and post-ablation imaging) was 314 ± 54min. Time for the MR-guided ablation phase of the procedure (first to final RF energy delivery) was 77 ± 40min, and total ablation time 19.4 ± 7.1min.

Conclusion: This study confirms safety and feasibility of actively-tracked MR-guided ablation, and its successful application in the ablation of typical atrial flutter in humans.

Articles from Europace are provided here courtesy of Oxford University Press

RESOURCES