Response to CRT: The interrelation of electrical activation patterns with mechanical contraction patterns and dyssynchrony measures.

 

Dr. Tom Jackson BSc(hons) MBBS MRCP

Contents
Abstract 2
Introduction 4
Materials and Methods 17
Preliminary Results 24
Discussion 33
References 38
Appendix 44
Timetable for completion of thesis 47
Abstract
Background. Current guidelines recommend cardiac resynchronisation therapy for patients with an ejection fraction less than 35% and a QRS duration over 120ms, with a positive response seen in 70% of patients. Imaging and electrophysiological studies have shown heterogeneous contraction and activation patterns amongst these patients, which may explain why some patients do not respond. Further studies have also shown a reasonable response rate amongst patients with a narrow QRS raising the question whether QRS duration accurately predicts who stands to benefit from this therapy.
Aims. To provide mechanistic insight into response to cardiac resynchronisation therapy using high fidelity imaging and electrical activation visualisation techniques. In particular to investigate potential mechanisms for response in patients not currently considered for CRT due to a normal QRS duration on the surface ECG. In addition to consider how the results of pre-implant investigations can best be combined using computer modelling protocols, in order that the clinician can best predict the chance of response in any individual patient.
Methods. This work is comprised of three prospective studies. The first uses cardiac MRI dyssynchrony and contraction pattern measures along with body surface ECG mapping to predict acute haemodynamic response to temporary biventricular and multisite pacing in patients with narrow QRS. In the second study electrical activation patterns are investigated using body surface mapping both before and during pacing in patients who are guideline indicated for resynchronisation. The final study tests computer modelling predictors of response to therapy within the clinical workflow.
Retrospective data has been used to investigate mechanical contraction patterns and response rates amongst patients with strict LBBB on the surface ECG (QRS = 140ms for men, =130ms for women and mid QRS notching/slurring in =2 contiguous leads), and MRI dyssynchrony measures amongst patients with a normal QRS duration and systolic heart failure.
Results. Thirty-seven patients with strict LBBB were investigated, 19 (51%) had a type II contraction pattern. A total of 25 (68%) of the cohort reverse remodeled. In the type II contraction group all 19 (100%) of patients reverse remodeled, compared to 6 (33%) in the type I group (p<0.01). Super-response was achieved in 21 (57%) of the total cohort, 5 (28%) of type I patients and 16 (84%) of type II patients (p<0.01).
Forty-five patients with a narrow QRS were investigated with a cardiac MRI. The systolic dyssynchrony index was 10.2 ± 7.9% for the whole group (p<0.01). Mechanical dyssynchrony was defined as >10% and was present in 14 patients (31%).
Conclusion. Varying left ventricular electrical and mechanical properties in systolic heart failure impact the ability for a patient to respond to cardiac resynchronisation. Greater personalisation using combinations of new electrical and mechanical imaging techniques represents a significant development in selection of candidates for this therapy.

Introduction including survey of the existing literature
Background
Heart failure is the most common debilitating chronic cardiac disease, affecting approximately 1 million people in the UK and still has a 1-year mortality of over 30% after a hospital admission.1 CRT is an established therapy for patients with systolic heart failure and a prolonged QRS duration (QRSd>120ms) usually manifested as left bundle branch block (LBBB) on the surface ECG. Early CRT trials, MUSTIC2 and MIRACLE3, showed improvements in symptoms, exercise tolerance, quality of life and adverse ventricular remodeling in the broad QRS heart failure population. Subsequent larger randomised clinical trials, COMPANION4 and CARE-HF5, showed reduction in mortality and hospitalisation with heart failure in this group. More recently benefits have been demonstrated in patients with milder degrees of heart failure, MADIT-CRT6 and RAFT7. CRT is therefore beneficial for the majority of patients with systolic heart failure, however despite 20 years experience 30% of patients implanted with CRT are still non-responders.8 Because of this the ongoing question remains as to whether current selection criteria are robust enough, or whether newer imaging technology, electrophysiology tools and data processing techniques should be used to select candidates for CRT.

Defining Electrical Dyssynchrony
CRT is established as a technique to counteract delayed interventricular electrical activation, however interest in response to CRT has added complexity to how delayed activation is best defined. Current European Society of Cardiology guidelines suggest that CRT is indicated in patients with an ejection fraction (EF) less than 35%, and New York Heart Association (NYHA) class II-IV symptoms with LBBB and a QRSd>150ms (class I level A evidence), LBBB and QRSd 120-150ms (class I level B), non LBBB QRSd>150ms (class IIa level B) and non LBBB QRS 120-150ms (class IIb level B).9 These guidelines also define LBBB as a QRSd=120ms, QS or rS in V1, broad frequently notched or slurred R wave is I, aVL, V5 or V6 and absent Q waves in V5 or V6. Recently, however, this definition has been challenged, because of an increased understanding of left ventricular (LV) activation and conduction velocities; therefore new ‘strict’ LBBB criteria have been proposed by Strauss et al.10 These are a QRS duration =140 ms in men and =130 ms in women with mid-QRS notching or slurring in at least 2 contiguous leads of I, aVL, V1, V2, V5 or V6. The first notch is due to delayed endocardial breakout on the left side of the interventricular septum and the second notch is the arrival of activation at the epicardium of the posterolateral wall. These criteria have shown concordance with the mechanical contraction pattern on 2-Dimensional echocardiographic strain imaging11 and mechanical dyssynchrony on tagged cardiac MRI.12 Patients with strict LBBB have demonstrated increased reverse remodeling (RR), symptomatic response and event free survival rate with CRT in comparison with non strict LBBB and intraventricular conduction delay.13-15 Another consideration is that traditional LBBB does not represent a single activation pattern, but rather a heterogeneous phenomenon associated with two electrical activation patterns: a type I pattern characterized by homogenous spread of electrical activation from the septum to lateral wall, and a type II pattern characterized by a ‘U-shaped’ spread of activation mandated by a line of functional block between the septum and lateral wall and delayed trans-septal conduction times.16-20 These activation patterns were defined using contact mapping, non-contact mapping and body surface mapping. A type II pattern of electrical activation has been shown to predict increased response to CRT.21 As yet it is unknown whether Strauss’ criteria represent a single activation pattern, possibly a type II pattern, or rather a heterogeneous group as seen in traditional LBBB.

Defining Mechanical Dyssynchrony
The presence of mechanical dyssynchrony has been linked with improved response rates over QRSd in multiple single centre studies using echocardiography.22-24 However, outside single centre studies these measures have been shown to be both poorly predictive and poorly reproducible.25,26 In addition the poor acoustic windows seen in heart failure patients often limit echo measures, and in particular 3D acquisitions, strain or SPEQLE analysis.27 Cardiac MRI is not limited by acoustic windows and has excellent spatial and temporal resolution.28 It is considerably more reproducible than echocardiography particularly for more complex measures.29
Novel cardiac MR acquisition (including 3D cardiac tagging) and image processing algorithms (to calculate a wide range of mechanical parameters including regional strain) have been developed over the last 5 years in our department.30,31 These show that measuring a volume derived systolic dyssynchrony index (SDI) strongly predicts ventricular remodeling post CRT in the left bundle branch block population when compared to standard echo measures of dyssynchrony, and is a highly reproducible measurement. In addition this index can be calculated from standard MRI cine imaging sequences, which are routine during a clinical MRI scan.
A further consideration is whether mechanical dyssynchrony is measured using deformation imaging (for example segmental volume change), strain imaging in any particular direction (radial, longitudinal or circumferential) or strain rate imaging. New processing tools make acquisition of each of these parameters possible both for 3D echocardiography and MRI.32-34

The inter-relation of electrical and mechanical dyssynchrony
There is yet to be any consistent finding of a strict association between presence of mechanical dyssynchrony and electrical dyssynchrony as defined by QRSd.35 In particular approximately 30-50% of heart failure patients with a narrow QRS have been shown to have mechanical dyssynchrony using echocardiography or cardiac MR tissue synchronization imaging.36,37 Newer technologies such as body surface mapping provide further insight into electrical dyssynchrony. These have suggested little association between QRSd and the presence of significant electrical dyssynchrony defined as the standard deviation of activation times at 500 sites on the LV epicardium, including the epicardial aspect of the septum, and are yet to be tested against the presence or absence of mechanical dyssynchrony.38 We therefore have a dilemma, is the QRS complex from a surface 12 lead ECG a suitable predictor for electrical dyssynchrony amenable to resynchronization? Secondly, if it is not, is there a marker of mechanical dyssynchrony which may better tally with this treatable electrical dyssynchrony, and thereby be a better predictor of response?
The explanation for the presence of mechanical dyssynchrony despite the presence of electrical synchrony demonstrated on the surface ECG (a narrow QRS) may not be because of misinterpretation of the investigations, but rather a true representation of the pathological state. The potential for pathophysiological processes leading to mechanical dyssynchrony despite maintained electrical synchrony has not been fully investigated, but potential mechanisms do exists in the literature. In a dog model, Yano et al showed that increasing afterload with an aortic band led to relaxation delay and dyssynchrony despite normal temporal electrical activation.39 However this seems an unlikely mechanism in patients. Tomaselli and Kass have done extensive work in dog models of dilated heart failure with left bundle branch block.40-44 They have shown heterogeneous changes in basic cellular and molecular pathways. Systems affected include calcium handling, ? adrenergic stimulated myocyte contractility, cell survival signaling and mitochondrial activity. Regional specificity is seen in multiple pathways, including activation of stress kinases in the lateral wall and transcriptional changes in the anterior wall; CRT can reverse some of these changes.45 It is possible to consider that this cellular and molecular variation may also occur in heart failure with normal QRS duration and could well lead to disparate excitation-contraction coupling and contractility. Contractility variability may also be secondary to fibrosis and scar. One hypothesis to explain the discoordinate overall motion in patients is this regional variability in contractility or excitation-contraction coupling leading to a strain mismatch, with the weaker wall being pushed out by its stronger partner despite synchronous temporal activation.35 Another theory is that mechanical dyssynchrony occurs in the setting of small heterogeneous areas of myocardial scar and fibrosis. Therefore although overall contraction is discoordinate, only small areas are affected and their electrical impact on the QRS morphology is negligible.46

Studies of CRT in narrow QRS patients
Positive studies
Achilli et al published the first study in this population in 2003.47 52 patients were implanted with CRT regardless of the QRS duration. 14 had a QRS duration of less than 120ms. All patients had echo evidence of inter and intraventricular dyssynchrony (using m-mode and pulsed Doppler). All patients including the group with a QRS<120ms had a significant improvement in NYHA class, LV ejection fraction (EF), degree of mitral regurgitation (MR), LV end systolic volume (LVESV), LV end diastolic volume (LVEDV), interventricular delay and 6 minute walk test (6MWT) at follow up. Mean follow up was 18 months.
Two single centre prospective studies were published in 2006. Yu et al enrolled 102 patients with NYHA III-IV heart failure.48 51 had a QRS duration over 120ms, and for those who had a narrow QRS, intraventricular dyssynchrony was demonstrated with echo tissue Doppler (TDI dyssynchrony index – SD 12 basal segments>32.6ms). Clinical and echo assessments were performed at baseline and at 3 months follow up. There was evidence of positive remodeling in both groups; in the narrow QRS group LVESV reduced by 19mls (p<0.01). Improvements were also seen in both groups with regard to NYHA class, degree of MR, 6MWT and EF. Furthermore the degree of baseline mechanical dyssynchrony determined LV reverse remodeling to a similar extent for both groups. Bleeker et al defined dyssynchrony as a peak systolic velocity difference between 2 LV walls >65ms assessed also using TDI.49 Other inclusion criteria were an EF<35% and an NYHA class of III-IV. They studied 33 patients with a QRS duration <120ms and 33 controls (QRS>120ms). There was no significant difference between the groups in reverse remodeling (LVESV reduction of 39?34 ml vs. 44?46 ml) or degree of improvement in the Minnesota living with heart failure questionnaire (MLHFQ), NYHA class, 6MWT and EF between the two groups at 6 months.
In the DESIRE study Cazeau et al assessed 60 patients with a QRS duration <150ms (mean 121ms?19ms) and NYHA III-IV heart failure from multiple centres.50 All patients were implanted with CRT and followed up for 1 year. 27 had echo evidence of mechanical dyssynchrony be it either atrioventricular, interventricular or intraventricular. Methods used to assess mechanical dyssynchrony included Doppler and m-mode imaging but not tissue Doppler. They used a composite of clinical outcomes as their primary endpoint including death from any cause, heart failure related hospitalisations and NYHA class. Mechanical dyssynchrony measures were superior to QRS duration in predicting response to CRT. At 6 months 70% of those with dyssynchrony improved in comparison to 42% without dyssynchrony (P<0.04). Among patients with a QRS<120ms 42% improved vs. 57% with QRS>120ms (P=ns).
Van Bommel et al evaluated different echo indices of mechanical dyssynchrony and their impact on selection for CRT in the narrow QRS group.51 In this 2 centre prospective study 123 patients with a QRS duration <120ms on maximal therapy for NYHA III systolic heart failure (LVEF<35%) underwent CRT. At 6 months follow up 48% fulfilled a predefined response to CRT of a ?15% reduction in LVESV. A reduction in NYHA functional class was seen in 72%. Intraventricular dyssynchrony assessed with both TDI and speckle tracking was substantially superior to interventricular dyssynchrony measures in predicting reverse remodeling. Receiver operator characteristics (ROC) analysis of opposite wall delay >75ms and anteroseptal to posterior wall delay >107ms had an area under the curve (AUC) of 0.661 and 0.722 respectively. Oyenuga et al also investigated the predictive power of echo measures in a similar group of patients.52 They assessed 221 patients prior to CRT implant; 86 had a QRS duration of 100-130ms and 135 had a QRS >130ms. Response to CRT was defined as a >15% improvement in EF and reverse remodeling as a >10% decrease in LVESV. In total 53% of the shorter QRS patients responded. Importantly interventricular mechanical delay >40ms and opposing wall delay >65ms were not predictors of response in the narrow QRS group (see RethinQ later). Speckle tracking radial dyssynchrony >130ms did predict response (ROC AUC 0.83).
PROSPECT was a large multicentre randomised controlled trial (RCT) appraising echo markers for mechanical dyssynchrony predicting response to CRT.25 41 patients were enrolled into a substudy for narrow QRS patients (<130ms).53 All had echo evidence of mechanical dyssynchrony (by any of seven measures including TDI). Using a clinical composite score after 6 months 63.4% had improved, 9.8% were unchanged and 26.8% had worsened. NYHA class, 6MWT and MLHFQ all showed significant improvement. Reverse remodeling was demonstrated with a decrease in LVESV from 59?9 to 55-?12mls (p=0.002).
RESPOND was a single centre RCT studying 60 patients with NYHA III-IV systolic heart failure and a QRS duration <120ms.54 No evidence of mechanical dyssynchrony was sought. 29 were randomised to CRT-P and 31 were randomised to optimum pharmacological therapy (OPT). The primary end point was determined as a >25% increase at 6 months in 6MWT. 51.7% of the CRT-P patients achieved this, vs. 12.9% of the OPT patients (p=0.0019). Other measures of response including improvement in MLHFQ, NYHA class and clinical composite score all showed a significant benefit to CRT-P.
NARROW-CRT was a multicentre prospective study implanting either CRT-D or and ICD into 111 patients with ischaemic cardiomyopathy, an EF<35% and mechanical dyssynchrony demonstrated on tissue Doppler.55 The primary endpoint was a clinical composite response and was met in 43% of the CRT group vs. 16% of the ICD group (p=0.004); there was also a significantly higher survival from the combined end point of hospitalization with heart failure, heart failure death and spontaneous ventricular fibrillation (p=0.028).

Negative studies
RethinQ is the largest double blind multicentre RCT examining CRT in patients with narrow QRS to date.56 172 patients with a standard indication for an ICD, NYHA III heart failure, a QRS duration <130ms and echo evidence of dyssynchrony using TDI opposite wall delay of >65ms or M-mode septal/posterior difference >130ms were implanted with CRT-D. For the majority of patients intraventricular dyssynchrony was demonstrated by opposing wall delay (see Oyenuga et al above). These patients were followed up at 6 months and after attrition 76 were analysed in the CRT-on group and 80 in the CRT-off group. The study did not meet its primary end point of an increase in VO2max of 1 ml/kg/min in the CRT-on group. A prespecified group of QRS>120ms did meet this end point. There was a significant improvement in at least one NYHA class in the CRT group (54 vs. 29% p=0.006). Other endpoints of 6MWT, MLHFQ and remodeling improvements did not differ between the groups.
In ESTEEM-CRT Donahue et al implanted 68 patients with a narrow QRS, EF of <35%, TDI derived mechanical dyssynchrony (Ts-SD-12>28.7ms) and an indication for an ICD with a CRT-D.26 They performed an acute haemodynamic study in 47 of these at the time of implant. Follow up was at 6 and 12 months. There was no significant improvement in the acute haemodynamic study (average increase in LV dP/dtmax of 2?2%). NYHA and quality of life scores were substantially improved at 6 and 12 months (P<0.001), however exercise capacity and LV volumes were unchanged. Interestingly, when reevaluating the echos at the core lab, 48% of patients appeared to have an EF>35% upon inclusion.
Ploux et al performed acute haemodynamic studies in 82 consecutive heart failure patients who underwent CRT implantation irrespective of their QRS duration.57 These patients were not assessed for mechanical dyssynchrony. 34 had a QRS <120ms, 11 had a QRS ?120 to <150ms, and 37 had a QRS?150ms. They found that there was a high correlation between changes in LV dP/dtmax and baseline QRS duration. For the narrow QRS patients no change was seen (+0.4% ? 6.1%; P=ns). In the moderately prolonged QRS group no significant increase was seen (+4.4% ? 6.9%; P=.06). There was a significant increase seen in the patients with the broadest QRS durations (+17.1% ? 13.4%; P<0.001).
LESSER-EARTH is the latest trial to be published in which patients with an EF<35%, heart failure symptoms limiting their 6MWT to less than 400m, a QRS<120ms and no assessment of mechanical dyssynchrony were implanted with CRT.58 It was a multi-centre double-blind study in which patients were implanted with CRT, and either randomised to CRT-on or CRT-off after a run-in period. Following randomisation of 85 patients it was stopped early as there was no difference in the primary end-point of maximal exercise duration, and safety concerns due to a trend towards increased heart failure hospitalisations in the CRT-on group. Of note only 53% of patients had NYHA III-IV symptoms initially.
EchoCRT is a multicentre prospective RCT in which 809 patients with an EF=35%, a QRS duration <130ms and echo evidence of dyssynchrony (demonstrated with tissue Doppler opposing wall delay >80ms or an anterior-posterior speckle tracking radial strain delay >130ms) were all implanted with a CRT-D and followed up for a mean of 19.4 months.59 404 patients had the left ventricular lead programmed on, with 405 randomised to ICD only. There was no significant difference in the primary outcome of death from any cause or hospitalisation with heart failure, and the conclusions were that CRT in this group may increase risk of death.
It therefore remains unclear whether CRT has a role to play in narrow QRS patients and if it does is this only in patients with mechanical dyssynchrony? There is specifically very little information as to the potential mechanisms of mechanical dyssynchrony leading to a possible therapeutic target for CRT. The RCTs that have been published (RethinQ, LESSER-EARTH and EchoCRT) either used dyssynchrony measures which have since been shown to be poorly predictive of echocardiographic response, or did not assess for dyssynchrony at all.
Advanced pacing techniques to maximize response
Traditional transvenous LV lead placement utilizes the coronary sinus to place a single lead on the epicardial surface of the LV. Endocardial pacing and multisite pacing may offer supplementary techniques to help improve response to CRT in narrow QRS. Endocardial LV pacing offers a mechanism by which the constraints of coronary sinus anatomy are overcome. This can be useful in the avoidance of postero-lateral scar19,60-62 and to reproduce the physiological gradient of LV contraction in an endocardial to epicardial direction.63-65 Studies at our centre have shown significant improvement in acute haemodynamic response (AHR) with endocardial biventricular pacing in comparison with traditional CRT in the left bundle branch block population (increase in LV dP/dtmax of 79.8±49% vs. 59.6±49.5% (p<0.05)).62 Multisite pacing employs multiple coronary sinus leads or a combination of coronary sinus and endocardial leads. Small studies at centres including our own have seen an improved AHR over that of traditional biventricular pacing.19,66,67 Two further studies have shown improved reverse LV remodeling with triventricular pacing when compared with biventricular pacing at 9 and 12 months.68,69

Computer modeling coordination of investigation results
Through the application of multi-scale and multi-physics approaches, these multi-level (from cell to organ) models have been coupled together to produce increasingly sophisticated simulations of cardiac physiology incorporating excitation, contraction, coronary perfusion and ventricular flows.70 A good example has been predicting the effect of different pacing locations on the heart function following CRT.71 However, simulating the heart of patients with heart failure comprising multiple aetiologies requires further integration of blood flow, electrophysiology and wall mechanics simulations. Combining these inputs into a usable clinical workflow remains a challenge due to the timescale for segmentation, parameterization and running of the model. However the efficiency of each process continues to improve and the workflow should now be applied within a clinical system.
Materials and methods
Overview of methods
For this investigation there are three separate study protocols each designed with the questions laid out in the introduction in mind.

Study 1: Cardiac Resynchronisation Therapy In Patients With Narrow QRS Morphology And Heart Failure: Mechanistic Insights From Cardiac MRI And Electroanatomical Mapping
This is an observational cohort study mechanistic study designed to recruit 30 patients from the Guy’s and St. Thomas’ heart failure clinic with narrow QRS systolic heart failure (EF<35%).
Primary Endpoint:
• Improvement in LV dP/dtmax >10% during temporary multisite pacing
Secondary Endpoint:
• Electrical dyssynchrony measures from Body Surface Mapping.
The participants are expected to need to attend a maximum of 3 visits (excluding unscheduled visits). These visits are listed below; further details of each visit are summarised in the flow chart in the appendix.
Visit 1 – Screening and cMR
Visit 2 – Body surface mapping
Visit 3 – Temporary pacing study

Study 2: Evaluation of a novel method of non-invasive surface electrocardiographic mapping in predicting clinical, structural and neurohormonal responses in patients undergoing cardiac resynchronization therapy
This is a prospective, single-arm study which will recruit thirty consecutive HF patients with a conventional indication for CRT-P/CRT-D implantation from St Thomas’ Hospital and Royal Brompton Hospital Heart Failure/Pacing clinics.
Combined Primary endpoint:
• Change in distance travelled during six-minute walk test (6MWT)
• Echocardiographic: signs of LV reverse remodeling defined as an increase by >5% in left ventricular ejection fraction with an associated decrease in LV end-diastolic (LVEDV) and end-systolic (LVESV) volumes
Secondary endpoints:
• Symptoms: change in symptoms severity assessed by Minnesota Living With Heart Failure Questionnaire (MLHFQ)
• Neurohormonal status: change in neurohormonal activation assessed by brain-natriuretic peptide (BNP)
• Pacing: atrial and ventricular arrhythmic burden, percentage of bi-ventricular pacing
The participants are expected to need to attend a maximum of 6 visits (excluding unscheduled visits). These visits are listed below; further details of each visit are summarised in the flow chart in the appendix.
Visit 1 – Screening visit and baseline assessment
Visit 2 – CRT Implant
Visit 3 – Day 1 post implant – Echo optimisation and body surface mapping in each CRT setting.
Visit 4 – 1 month follow up visit
Visit 5 – 3 month interim visit
Visit 6 – 6 month final visit and assessment with repeat body surface mapping study.

Study 3: Computer model derived indices for optimal patient-specific treatment selection and planning in heart failure.
This is a pilot study designed to generate complete datasets that can be used to parameterise computational models. The first stage of this study is to database historical datasets in order to run these through the planned modeling pipeline. Following this prospective patients will undergo a detailed imaging and bio-signal (including electrical and haemodynamic) assessment at a number of time points during their CRT implant timeline.
Primary Endpoint:
• Computational model prediction of reverse remodelling as defined by a reduction in LVESV (on echo or MRI) 6 months post intervention (CRT or MVR/r).
Secondary Endpoints:
• Computational model prediction of response to intervention as defined by an improvement 6 months following intervention in:
o MLHFQ
o 6MWT
o VO2 max
o BNP
o Non-invasive central blood pressure measurement
The prospective patients are expected to need to attend a maximum of 4 visits (excluding unscheduled visits). These visits are listed below; further details of each visit are summarised in the flow chart in the appendix.
Visit 1 – Screening and baseline assessment
Visit 2 – CRT implant
Visit 3 – Immediate post procedure assessment
Visit 4 – 6 month follow up visit

Each study protocol contains a combination of the following investigation methods:
• Twelve-lead ECGs acquired with a GE Mac 5000 ECG system (General Electric-Vingmed, Milwaukee, Wisconsin) using standard AHA recommended filter settings72 at a sweep speed of 25mm/s and a gain of 10mm/mV.
• A CMR performed with a 1.5T scanner (Achieva or Ingenia, Philips Healthcare, Best, The Netherlands) with a 32-element cardiac coil prior to CRT. Cardiac synchronization will be performed with vector electrocardiography. After localization and a coil sensitivity reference scan, an interactive real-time scan will be performed to determine the geometry of the short axis, 4 chamber (4CH), 3 chamber (3CH) and 2 chamber (2CH) orientations. A multiple slice cine steady state free precession (SSFP) scan will be performed in a stack of short axis slices covering the left ventricle (LV) in the 4CH, 3CH and 2CH orientations (FA=60°, TR/TE=2.9/1.5ms, resolution 2.2×2.2x10mm, 35 heart phases – temporal resolution approximately 30ms).
• Standard and 3D transthoracic echocardiographic assessment using an IE33 or an EPIC model scanner (Philips Healthcare, Best, The Netherlands). Analysis will be performed in Q-Lab providing data for LV function and volumes pre-implantation and at 6 months follow-up. EF and LV dimensions are measured using the 2-dimensional (2D) modified biplane Simpson’s method. End-systolic volumes will be recorded pre-implantation and at 6 months. Reverse remodeling (RR) will be defined as a reduction in an end systolic volume (ESV) of =15%.73
• Body surface ECG mapping using the CardioInsight ecVUE system (CardioInsight Technologies Inc., Cleveland, Ohio) to generate epicardial activation maps. Body surface potentials are recorded from 252 electrodes around the surface of the torso, a thoracic computed tomography (CT) scan is performed with the electrodes attached. The CT images are combined with the body surface potentials to reconstruct 1500 unipolar electrograms. Ventricular activation times are calculated from the onset of the QRS to the maximal negative slope of each electrogram and combined for an isochrone map. LV total activation time (LVTAT), RV total activation times (RVTAT) and a value for ventricular electrical uncoupling (VEU – difference between LV and RV mean activation times) are calculated by ecSYNC software.
• Assessment of mechanical contraction patterns and dyssynchrony measures from MRI will be done using a prototype software platform (TomTec, Unterschleissheim, Germany). This software uses long and short axis SSFP images to track endocardial and epicardial contours in order to generate left ventricular contraction maps and segmental volume change curves (figure 1).

Figure 1. Schematic representation of a Type II contraction pattern in a strict LBBB patient superimposed on a still image from the TomTec output for the contraction propagation. Bold line represents an anteroseptal line of contraction block, arrow shows contraction front direction of travel.

•
• Temporary pacing electrophysiological study (narrow QRS study) in which catheters will be placed within the RA, RV, coronary sinus (LV pacing lead) and LV via a retrograde aortic approach. Further arterial access is needed in order to place a pressure wire (RADI wire – St Jude Medical, St. Paul, Minnesota) within the LV in order to measure the acute haemodynamic change (LV dP/dt). A pacing protocol is followed where the effects of biventricular, triventricular and endocardial pacing on dP/dt is tested. Non-contact mapping will also be undertaken in 10 patients (EnSite balloon – St Jude Medical) as a comparator with body surface mapping and in order to acquire endocardial electrical activation data.
• CRT-D/P insertion as per routine implant techniques.
• Clinical assessment to include grading of NYHA class, Minnesota living with heart failure questionnaire score, 6 minute walk test distance, and cardio-pulmonary exercise testing using a cycle ergometer for VO2max.
• Serum analysis for renal profile, full blood count and BNP levels.
Preliminary results
Electrical and mechanical dyssynchrony assessment of historical data
Strict LBBB and contraction patterns
Thirty-seven patients implanted with CRT who had strict LBBB as defined by Strauss et al10 prior to implant, had analysis of their contraction pattern. They were aged 66.0 ± 11.7 years; 31 (84%) were male. The aetiology was ischemic in 14 (38%) and the majority of patients were NYHA class III. The echocardiographic mean LV ejection fraction pre-implant was 22.0% ± 8.2%. The mean QRS duration was 168±18ms. On the basis of previously described criteria, a type II contraction pattern was present in 19 patients (51%). There were no significant differences in baseline characteristics between patients with a type I or a type II contraction pattern (table 1). In total 25 patients (68%) were deemed echocardiographic responders with a >15% reduction in LVESV (table 2). Responders had a non-significant trend towards a broader QRS (173.1 ± 15.0 vs162.9 ± 19.3ms, p=0.08). Nine (64%) ischemic patients were responders compared to 16 (70%) non ischemic (p=1.0).
All (n=37) Type I (n=18) Type II (n=19) p value
Age 66.0 ± 11.7 65.5 ± 14.7 66.4 ± 8.1 0.82
Sex M 31 (84) 16 (89) 15 (79) 0.66
F 6 (16) 2 (11) 4 (21)
Etiology ICM 14 (38) 9 (50) 5 (26) 0.18
NICM 23 (62) 9 (50) 14 (74)
NYHA Class II 7 (19) 3 (17) 4 (21) N/A
III 27 (73) 12 (66) 15 (79)
IV 3 (8) 3 (17) 0 (0)
HF score 51.0 ± 22.2 51.8 ± 20.4 50.1 ± 24.4 0.82
QRS duration (ms) 168 ± 18 162.9 ± 19.3 173.1 ± 15.0 0.08
LVESV (ml) 186.5 ± 61 190 ± 63.4 183.1 ± 61.0 0.45
LVEDV (ml) 236.3 ± 68 245.1 ± 68.7 227.6 ± 67.5 0.74
Ejection Fraction 22.0 ± 8.2 24.2 ± 9.5 19.7± 6.1 0.10
Table 1: Baseline characteristics and characteristics by contraction pattern. Values are mean ± SD or n (%).
(HF – Heart failure, ICM – Ischemic cardiomyopathy, LVEDV – Left ventricular end diastolic volume, LVESV – Left ventricular end systolic volume, NICM – Non-ischemic Cardiomyopathy)
Responder (n=25) Non-Responder (n=12) p value
Age 65.7±8.3 66.5±17.1 0.84
Sex M 21 (68) 10 (32) 1.0
F 4 (67) 2 (33)
Etiology ICM 9 (64) 5 (36) 1.0
NICM 16 (70) 7 (30)
HF Score 52.6±22.9 47.4±21.3 0.51
QRS duration (ms) 171.5 ± 17.7 161.0 ± 16.3 0.09
LVESV (ml) 185.1 ± 61.1 189.2 ± 64.4 0.86
LVEDV (ml) 235.0 ± 69.2 239.0 ± 67.6 0.87
Ejection Fraction 21.2 ± 7.4 23.6 ± 9.7 0.41

Table 2. Baseline characteristics divided by responders and non-responders on basis of a decrease in LVESV >15%. Values are mean ± SD or n (%).
(HF – Heart failure, ICM – Ischemic cardiomyopathy, LVEDV – Left ventricular end diastolic volume, LVESV – Left ventricular end systolic volume, NICM – Non-ischemic Cardiomyopathy)

Contraction pattern
Notably all 19 patients with strict LBBB and a type II contraction fulfilled the criteria for RR compared to only 6/18 patients (33%) with a type I contraction pattern (p<0.01). A type II contraction pattern also predicted a significantly greater reduction in LVESV with CRT; 32.8 ± 14.1% vs. 10.8 ± 20.1% (p<0.01, figure 3). Patients with type II contraction had broader QRS but this did not reach statistical significance (162.9 ± 19.3ms vs. 173.1 ± 15.0ms, p=0.08 – see table 1). Figure 2 shows 12 lead ECGs in patients with a type I and II contraction pattern. There was no significant difference in clinical response between patients with a type I or type II contraction pattern (table 3).

Figure 2. 12 lead ECGs of A) Patient with a type I contraction pattern. B) Patient with a type II contraction pattern. Note: absence of definite notching in any leads in A, whereas there is definite notching in I, V5 and V6 in B. Examples of QRS complexes in aVL from patients who fulfill criteria for strict LBBB with prominent mid QRS notching in C and mid QRS slurring in D.

Figure 3. Boxplots of % end systolic volume (ESV) and absolute Ejection Fraction (EF) change following 6 months CRT pacing grouped by contraction pattern

Combined predictors of RR and super-response
In total 21patients (57%) fulfilled any criteria for super-response whereas only 13 (35%) were super-responders on basis of an ESV reduction =30%. Sixteen (84%) patients with type II contraction were super-responders on the basis of any criteria (reduction in LVESV=30%, post CRT EF=50%, EF increase by =10% and reduction in LVEDV =20% or NYHA I/II post CRT and =20% increase in EF)13,73,74 compared to only 5 (27%) with a type 1 contraction (p<0.01). On the basis of >30% LVESV reduction 9 (47%) patients with type II contraction were super responders compared to only 4 (22%) with type I contraction (p=0.11) (see table 4). Amongst ischemic patients a type II pattern predicted RR, but was not statistically significant (p=0.09). Amongst non-ischemic patients a type II pattern did predict RR (p<0.01). On multivariate analysis of contraction pattern, etiology and QRSd > 150ms a type II contraction pattern was the only independent predictor of super-response by any definition (see figure 4).

Figure 4: Predictors of super-response. Panel A: Response and different type of super-response rate separated by contraction pattern. Blue bars all patients, red type I contraction and green bars type II contraction. Panel B: Forest plot of the odds ratios for CRT super-response in a multivariate regression model including 3 variables. OR – Odds Ratio, CI – Confidence Interval.
All (n=37) Type I (n=18) Type II (n=19) P value
Average change NYHA Class 1.03 ± 0.55 0.9 ± 0.6 1.2 ± 0.5 0.14
Change NYHA class >1 Yes 32 (86) 14 (78) 18 (95) 0.18
No 5 (14) 4 (22) 1 (5)
Change in HF questionnaire score 23.3 ± 22.1 22.5 ± 25.4 23.9 ± 20.0 0.86
Absolute change EF (%) 10.9 ± 2.0 3.7 ± 8.7 17.2 ± 10.0 <0.01
Change in LVESV (%) 22.1 ± 20.4 10.8 ± 20.1 32.8 ± 14.1 <0.01
Responder Yes 25 (68) 6 (33) 19 (100) <0.01
No 12 (32) 12 (67) 0 (0)

Table 3. Outcome measures separated on the basis of contraction pattern. Values are mean ± SD or n (%)
(EF – Ejection Fraction, HF – Heart failure, LVESV – Left ventricular end systolic volume, NYHA – New York Heart Association)

Type of Super-Response Contraction Pattern Type I Type II ?2 OR (95% CI) p value
Reduction LVESV = 30% Yes 4 (22) 9 (47) 2.6 3.15 (0.75-13.17) 0.11
No 14 (78) 10 (53)
Post CRT EF =50% Yes 0 (0) 2 (11) N/A N/A 0.49
No 18 (100) 17 (89)
EF Increase =10% Yes 2 (11) 14 (74) 14.7 22.40 (3.74-134.14) <0.01
No 16 (89) 5 (26)
Reduction LVEDV =20% Yes 4 (22) 5 (26) 0.1 1.25 (0.28-5.65) 0.77
No 14 (78) 14 (74)
NYHA I/II & =20% Increase in EF Yes 2 (11) 6 (32) 2.3 3.69 (0.64-21.46) 0.13
No 16 (89) 13 (68)
Any type super-response Yes 5 (28) 16 (84) 12.0 13.87 (2.78-69.21) <0.01
No 13 (72) 3 (16)

Table 4. Number of super-responders depending on classification of super-response amongst strict LBBB patients categorized by contraction pattern.
(CRT – Cardiac Resynchronization Therapy, EF – Ejection Fraction, LVEDV – Left Ventricular End Diastolic Volume, LVESV – Left Ventricular End Systolic Volume, NYHA – New York Heart Association)
Narrow QRS and mechanical dyssynchrony
Forty-five patients with NYHA III-IV symptomatic heart failure and an ejection fraction (EF)<35% had a clinical CMR scan at our institution, 18 of these patients had the decision to have a CRT-D implanted on clinical grounds by the heart failure multi-disciplinary team.

Presence of Mechanical Dyssynchrony
The SDI for the whole cohort was 10.2 ± 7.9% (p<0.01), with no significant difference between groups when separated for gender, aetiology or presence of sinus rhythm (tables 5&6). Mechanical dyssynchrony was present in 31% of patients. When dichotomizing for presence of mechanical dyssynchrony there was no statistical difference between age, QRS duration, left ventricular volumes and EF; however there was an almost significant difference in ejection fraction between groups with those with mechanical dyssynchrony tending to have a lower EF when compared to those without dyssynchrony (24.7 ± 11.8 vs. 30.2 ± 7.9%, p=0.08, table 7).
Baseline Characteristics (n=45)
Age (yrs.) 58.5 ± 15.9
Gender 37(82) M, 8(18) F
Aetiology 27(60) ICM; 18(40) NICM
LVEDV (mls.) 241.8 ± 112.7
LVESV (mls.) 180.4 ± 106.6
EF (%) 28.5 ± 9.5
QRS duration (msec.) 106.6 ± 10.6
ECG Rhythm 39(87) SR; 4(9) AF; 2(4) AFl
Table 5. Baseline characteristics. Values are mean ± SD or n (%). ICM – Ischaemic Cardiomyopathy, NICM – Non-Ischaemic Cardiomyopathy, SR – Sinus rhythm, AF – Atrial Fibrillation, AFl – Atrial Flutter.

SDI (%) p value
All (45) 10.2 ± 7.9
Gender M (37) 10.8 ± 2.7 0.26
F (8) 7.3 ± 2.7
Aetiology ICM (27) 10.5 ± 7.3 0.76
NICM (18) 9.7 ± 8.9
Rhythm SR (39) 10.5 ± 8.3 0.53
AF/AFl (6) 8.3 ± 3.7
Table 6. Systolic dyssynchrony index amongst all patients with narrow QRS and separated by gender, aetiology and rhythm. Values are mean ± SD.

Implanted Patients
Eighteen patients were implanted with CRT-D within this cohort, table 8 shows their baseline characteristics. There was follow up 2D echocardiographic data in 15 of the 18 patients, and time from implant to follow up echo was 6.3 ± 4.0 months, with a range of 2 to 17 months. One patient died during 4 months following implant form a heart failure related cause.
A positive clinical response was seen with an improvement of 0.7±0.8 NYHA classes (p<0.01) and 67% of patients who improving by >1 NYHA class. Mean reduction in ESV was 3.2±22.8% (p=0.6), and improvement in absolute EF was 1.5±6.6% (p=0.4). RR was seen in 40% of patients (table 9). When dichotomized for presence of mechanical dyssynchrony and response there were no significant differences between groups regarding age, gender, aetiology, LV volumes, EF and QRS duration. Responders tended to have a lower EF and a shorter QRS duration.
Comparing response and RR between groups with and without mechanical dyssynchrony there was no difference between groups; for those with dyssynchrony there was a non significant reduction in LVESV (15.6±18.5% vs. -3.0±22.9% (p=0.14)), and improvement in EF (5.4±5.2 vs. -0.5±6.5% (p=0.1)). There was a significant difference between groups with regards to duration until follow up echo, those with mechanical dyssynchrony had their echo 3.2±1.1 months vs. 7.9±4.0 months for those without (p=0.02, see table 11).
Figures 5 and 6 show scatter plots and regression lines for SDI plotted against change in LVESV and change in EF following pacing. These both show a loose non-significant association between increasing SDI and improvement; for LVESV R=0.42,p=0.12, for EF R=0.41,p=0.13.

Mechanical Dyssynchrony (n=14) No Mechanical Dyssynchrony (n=31) p value
Age 58.2 ± 16.8 58.7 ± 15.7 0.93
QRS duration (msec.) 108.3 ± 11.7 105.9 ± 10.2 0.51
LVEDV (mls.) 237.2 ± 98.6 243.9 ± 120.0 0.86
LVESV (mls.) 191.4 ± 105.1 175.5 ± 108.7 0.65
EF (%) 24.7 ± 11.8 30.2 ± 7.9 0.08
Table 7. Age, QRS duration, LV volumes and ejection fraction dichotomized by presence or absence of mechanical dyssynchrony. Values are mean ± SD.

Baseline Characteristics (n=18)
Age (yrs.) 63.8 ± 14.7
Gender 14(78) M, 4(22) F
Aetiology 15(83) ICM; 3(17) NICM
LVEDV (mls.) 260.9 ± 140.9
LVESV (mls.) 194.6 ± 128.5
EF (%) 27.9 ± 9.5
QRS duration (msec.) 112.0 ± 8.6
SDI (%) 10.0 ± 6.0
Table 8. Baseline characteristics for those implanted with CRT. Values are mean ± SD or n (%). ICM – Ischaemic Cardiomyopathy, NICM – Non-Ischaemic Cardiomyopathy, SR – Sinus rhythm, AF – Atrial Fibrillation, AFl – Atrial Flutter.

Measures of Response
Improvement in NYHA Class 0.7 ± 0.8
NYHA class improvement>1 12 (67)
Time to Follow-Up Echocardiogram* (months) 6.3 ± 4.0
Change in ESV (Pre-Post)(%)* 3.2 ± 22.8
Change in EF (Post-Pre)(%)* 1.5 ± 6.6
Responders (?ESV>15%)* 6 (40)
Table 9. Measures of response for whole cohort. *Follow-up echo data only available in 15 patients. Values are mean ± SD, or n (%).

MD No MD p value Response No Response p value
Age (yrs.) 62.5 ± 13.4 64.4 ± 15.9 0.80 62.7 ± 10.9 63.3 ± 19.5 0.94
Gender M 6 (33) 8 (45) 0.25 5 (33) 6 (40) 0.60
F 0 (0) 4 (22) 1 (7) 3 (20)
Aetiology ICM 5 (28) 10 (55) 1.0 8 (56) 5 (33) 1.0
NICM 1 (6) 2 (11) 1 (7) 1 (7)
LVEDV (mls.) 234.0 ± 68.7 274.3 ± 167.1 0.75 230.4 ± 55.9 293.7 ± 192.3 0.45
LVESV (mls.) 183.5 ± 84.8 200.2 ± 148.8 0.82 172.0 ± 67.3 225.8 ± 171.8 0.48
EF (%) 25.0 ± 14.2 29.3 ± 6.5 0.10 27.4 ±11.7 25.9 ± 9.3 0.79
QRS duration (msec.) 108.8 ± 12.5 113.6 ± 6.0 0.28 106.8 ± 11.3 115.2 ± 4.1 0.06
SDI (%) 17.2 ± 5.1 6.5 ± 1.5 <0.01 12.4 ± 7.4 9.1 ± 5.3 0.33
Table 10. Baseline characteristics for those implanted with CRT separated by those with and without mechanical dyssynchrony (MD), and response. Values are mean ± SD or n(%).

Measures of Response Mechanical Dyssynchrony No Mechanical Dyssynchrony P value
Improvement in NYHA Class 0.8 ± 0.8 0.6 ± 0.9 0.57
NYHA class improvement>1 Yes 4 (22) 8 (45) 1.0
No 2 (11) 4 (22)
Time to Follow-Up Echocardiogram* (months) 3.2 ± 1.1 7.9 ± 4.0 0.02
Change in ESV (Pre-Post)(%)* 15.6 ± 18.5 -3.0 ± 22.9 0.14
Change in EF (Post-Pre)(%)* 5.4 ± 5.2 -0.5 ± 6.5 0.10
Responders (?ESV>15%)* Yes 3 (20) 3 (20) 0.32
No 2 (13) 7 (47)
Table 11. Measures of response dichotomized by presence or absence of mechanical dyssynchrony. *Follow-up echo data only available in 15 patients. Values are mean ± SD, or n (%).
Figure 5. Linear regression plot for SDI against change in ESV after CRT; lines are regression line and 95% confidence intervals. R=0.42, p=0.12.
Figure 6. Linear regression plot for SDI against change in EF after CRT; lines are regression line and 95% confidence intervals. R=0.41, p=0.13.

Discussion
Strict LBBB and mechanical contraction patterns
The main findings of this study are that:
1) Amongst patients with strict LBBB planned for CRT both type I and type II contraction patterns are demonstrated on CMR.
2) Not all patients with strict criteria for LBBB significantly remodeled with CRT, however an overall 68% RR in our study is high compared to series with guideline CRT indications.25
3) Knowledge of contraction pattern in addition to strict LBBB gives further predictive power with 100% of patients with strict LBBB and type II contraction being volumetric responders.
4) 84% of patients with strict LBBB and type II contraction patterns were classified as super-responders.
Our results would suggest that strict LBBB criteria alone are insufficient to predict volumetric CRT response. This may be explained by the fact that either these patients still do not have true complete LBBB, or true LBBB has many electrical manifestations not all of which are amenable to CRT, or that they may have true LBBB but are unable to respond due to ischemic scar. The failure to see a significant difference between ischemic vs. non-ischemic patients makes the latter less likely. These results suggest that a type II contraction pattern in conjunction with strict ECG criteria represents a true LBBB amenable to electrical correction with CRT.

Narrow QRS and MRI derived SDI
This is the first in man study demonstrating a significant amount of MD amongst narrow QRS heart failure patients using CMR. The proportion of patients with MD is comparable to previous echo studies. Yu et al36 used tissue Doppler to demonstrate MD in 51% of narrow QRS patients. Matsumoto et al75 and Tatsumi et al76 both used 3D speckle tracking and identified MD in 39% and 22% of narrow QRS patients respectively. There is a very small CMR study in a canine model using myocardial tissue tagging and cine displacement encoding which demonstrated some increased radial dyssynchrony in narrow QRS heart failure when compared with no heart failure, but far less than in the LBBB model.77
Of our 18 implanted patients we have follow up data on 15 patients. This data is variable as the duration to follow up echo ranged from 2 to 17 months. In total 40% of these patients reverse remodeled. This is a similar finding to previous studies, Cazeau et al50 found 42% RR, Van Bommel et al51 found 48% RR and Oyenuga et al52 found 52% RR. We did not find that presence of mechanical dyssynchrony from CMR SDI predicted RR, however this was retrospective and heterogenous data and therefore is not powered for this. We did see a signal that increasing SDI may predict degree of RR in this group of patients, this finding is in agreement with Yu et al48 who used echo tissue Doppler to assess mechanical dyssynchrony. Interestingly there was a non-significant trend for those who responded to have a shorter QRS duration; this is in contradiction to much larger published data78 and is likely only because of our small sample.
Why should we investigate for mechanical dyssynchrony and CRT response in narrow QRS patients?
There have been 3 multicentre randomized control studies showing no improvement and a possible signal for harm for CRT in this population (RethinQ, Lesser-Earth, Echo-CRT) and one multicenter randomized control study showing an improvement (NARROW-CRT). RethinQ, Echo-CRT and NARROW-CRT all used echo indices of MD. These indices included tissue Doppler and speckle tracking strain imaging. Tissue Doppler dyssynchrony measures have been since shown to be poorly predictive and poorly reproducible.25,79 3D speckle tracking is more robust, however continues to be hampered by echo windows and intra-operator variability.33 In the two negative studies using echo markers of dyssynchrony all of the patients in RethinQ used TDI, and 25% of the Echo-CRT patients demonstrated MD with TDI alone. If we are going to assess for MD in narrow QRS patients, why would we use biomarkers found not to be useful in patients with broader QRS complexes?
Furthermore newer techniques such as multi-spot80, multi-vein68 and endocardial pacing81 are evolving providing increasing options for implantation that show promise in broad QRS non-responders and may assist in narrow QRS patients with mechanical dyssynchrony. Therefore although not all narrow QRS heart failure patients will benefit from CRT, there may well be a group who stand to improve with CRT and as yet we do not know how to identify them.

What role will computer modeling have in selection for CRT?
As demonstrated in this historical data there are multiple factors to consider in order to predict response to CRT. It is also important to consider how to represent this prediction of ‘response’. For the historical data presented here a definition of ESV improvement >15% was chosen as this has been shown to be associated with a favourable long-term outcome73; there is however an ongoing discussion as to other definitions such as a failure to deteriorate.82 A tool where using baseline patient parameters, such as imaging and electrical data, to present to the clinician and patient information on degree of response expected (by any measure) would be clinically very useful. A simplified version of this predictive tool is currently available in the form of current guidelines9, however computer modeling and statistical machine learning tools provide mechanisms by which the output can be more specific, providing information on degree of response expected or chance of a predefined response such as RR. These models also can provide information on which pacing combination and lead placements are most likely to illicit the best response.83

Limitations
The historical data presented is in a relatively small population from a single center and the findings will need to be validated in a larger population. The technique used to predict mechanical dyssynchrony is not yet validated in the heart failure population, however it is currently under consideration for publication in patients with guideline indications for CRT. The narrow QRS response data is retrospective and heterogeneous in nature with variation in whom was implanted and duration to follow up echo; our results from this data are not statistically significant, setting up a prospective study will help discover if the positive signal seen here translates into a significant result. Finally although MRI acquisition is increasingly ubiquitous for heart failure patients, assessment of contraction patterns is not yet routine practice.

Conclusions
Systolic heart failure represents a group of heterogeneous left ventricular electrical activation and mechanical contraction patterns. Even amongst patients with traditional and strict LBBB these patterns differ. The relationship between electrical activation and mechanical contraction in these patients is yet to be fully explained, especially when looking beyond the QRS duration and morphology with more precise electrical investigations. This understanding is important as the strict LBBB data suggests that knowledge of the contraction pattern has an impact on the response rate to CRT in this group.
CRT is not currently recommended for any patients with narrow QRS; there have however been some studies showing an improvement in this group. This improvement has not been translated into larger multicentre studies, but the methodologies are open to critisism primarily due to the method used to demonstrate mechanical dyssynchrony. Focusing on improved characterization of the underlying pathophysiological state may expose a group of patients with narrow QRS who stand to benefit from CRT.
Combining all of the imaging and electrical data generated in the work up of any individual for CRT into a computer model with predictive outputs for response stands to revolutionise assessment of these patients; providing the clinician and patient with the ability to make an informed decision as to whether the implant is appropriate.

References
1. Cleland, J., Dargie, H., Hardman, S., McDonagh, T. & Mitchell, P. National Heart Failure Audit. April 2011 – March 2012. NICOR 1–64 (2012).
2. CAZEAU, S. et al. Effects of multisite biventricular pacing in patients with heart failure and intraventricular conduction delay. N Engl J Med 344, 873–880 (2001).
3. Abraham, W. T. et al. Cardiac resynchronization in chronic heart failure. N Engl J Med 346, 1845–1853 (2002).
4. Bristow, M. R. et al. Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med 350, 2140–2150 (2004).
5. Cleland, J. G. F. et al. The effect of cardiac resynchronization on morbidity and mortality in heart failure. N Engl J Med 352, 1539–1549 (2005).
6. Moss, A. J. et al. Cardiac-Resynchronization Therapy for the Prevention of Heart-Failure Events. N Engl J Med 361, 1329–1338 (2009).
7. Tang, A. S. L. et al. Cardiac-Resynchronization Therapy for Mild-to-Moderate Heart Failure. N Engl J Med 363, 2385–2395 (2010).
8. Auricchio, A. & Prinzen, F. W. Non-Responders to Cardiac Resynchronization Therapy. Circ J 75, 521–527 (2011).
9. Authors/Task Force Members et al. 2013 ESC Guidelines on cardiac pacing and cardiac resynchronization therapy: The Task Force on cardiac pacing and resynchronization therapy of the European Society of Cardiology (ESC). Developed in collaboration with the European Heart Rhythm Association (EHRA). European Heart Journal 34, 2281–2329 (2013).
10. Strauss, D. G., Selvester, R. H. & Wagner, G. S. Defining Left Bundle Branch Block in the Era of Cardiac Resynchronization Therapy. The American Journal of Cardiology 107, 927–934 (2011).
11. Risum, N. et al. Left bundle-branch block: The relationship between electrocardiogram electrical activation and echocardiography mechanical contraction. American Heart Journal 166, 340–348 (2013).
12. Andersson, L. G. et al. Left ventricular mechanical dyssynchrony by cardiac magnetic resonance is greater in patients with strict vs nonstrict electrocardiogram criteria for left bundle-branch block. American Heart Journal 165, 956–963 (2013).
13. Tian, Y. et al. True complete left bundle branch block morphology strongly predicts good response to cardiac resynchronization therapy. Europace 15, 1499–1506 (2013).
14. van Deursen MD, C. J. M. et al. The value of the 12-lead ECG for evaluation and optimization of cardiac resynchronization therapy in daily clinical practice. Journal of Electrocardiology 1–10 (2014). doi:10.1016/j.jelectrocard.2014.01.007
15. Mascioli, G. et al. Electrocardiographic criteria of true left bundle branch block: a simple sign to predict a better clinical and instrumental response to CRT. Pacing Clin Electrophysiol 35, 927–934 (2012).
16. Rodriguez, L.-M., Timmermans, C., Nabar, A., Beatty, G. & Wellens, H. J. J. Variable patterns of septal activation in patients with left bundle branch block and heart failure. Journal of Cardiovascular Electrophysiology 14, 135–141 (2003).
17. Auricchio, A. et al. Characterization of left ventricular activation in patients with heart failure and left bundle-branch block. Circulation 109, 1133–1139 (2004).
18. Fung, J. W. H. et al. Variable left ventricular activation pattern in patients with heart failure and left bundle branch block. Heart 90, 17–19 (2004).
19. Ginks, M. R. et al. Benefits of Endocardial and Multisite Pacing Are Dependent on the Type of Left Ventricular Electric Activation Pattern and Presence of Ischemic Heart Disease: Insights from Electroanatomic Mapping. Circulation: Arrhythmia and Electrophysiology 5, 889–897 (2012).
20. Jia, P. et al. Electrocardiographic imaging of cardiac resynchronization therapy in heart failure: Observation of variable electrophysiologic responses. Heart Rhythm 3, 296–310 (2006).
21. Duckett, S. G. et al. Relationship between endocardial activation sequences defined by high-density mapping to early septal contraction (septal flash) in patients with left bundle branch block undergoing cardiac resynchronization therapy. Europace 14, 99–106 (2011).
22. Bax, J. J. et al. Left ventricular dyssynchrony predicts response and prognosis after cardiac resynchronization therapy. JAC 44, 1834–1840 (2004).
23. Tanaka, H. et al. Dyssynchrony by speckle-tracking echocardiography and response to cardiac resynchronization therapy: results of the Speckle Tracking and Resynchronization (STAR) study. European Heart Journal 31, 1690–1700 (2010).
24. Lim, P. et al. Multicentre study using strain delay index for predicting response to cardiac resynchronization therapy (MUSIC study). European Journal of Heart Failure 13, 984–991 (2011).
25. Chung, E. S. et al. Results of the Predictors of Response to CRT (PROSPECT) Trial. Circulation 117, 2608–2616 (2008).
26. on behalf of the ESTEEM-CRT Investigators et al. Acute and Chronic Response to CRT in Narrow QRS Patients. J. of Cardiovasc. Trans. Res. 5, 232–241 (2011).
27. Bilchick, K. C. et al. Cardiac Magnetic Resonance Assessment of Dyssynchrony and Myocardial Scar Predicts Function Class Improvement Following Cardiac Resynchronization Therapy. JACC: Cardiovascular Imaging 1, 561–568 (2008).
28. Ypenburg, C. et al. Noninvasive imaging in cardiac resynchronization therapy–part 1: selection of patients. Pacing Clin Electrophysiol 31, 1475–1499 (2008).
29. Keenan, N. G. & Pennell, D. J. CMR of Ventricular Function. Echocardiography 24, 185–193 (2007).
30. Duckett, S. G. et al. Systolic Dyssynchrony Index derived from cardiac magnetic resonance imaging predicts left ventricular remodeling in heart failure patients undergoing CRT. J Cardiovasc Magn Reson 13, P175 (2011).
31. Shetty, A. et al. 152 Real-time cardiac MR anatomy and dyssynchrony overlay to guide left ventricular lead placement in CRT. Heart 97, A84–A85 (2011).
32. Hui Wang & Amini, A. A. Cardiac Motion and Deformation Recovery From MRI: A Review. IEEE Trans. Med. Imaging 31, 487–503 (2012).
33. Kalogeropoulos, A. P., Georgiopoulou, V. V., Gheorghiade, M. & Butler, J. Echocardiographic evaluation of left ventricular structure and function: new modalities and potential applications in clinical trials. J. Card. Fail. 18, 159–172 (2012).
34. Bilchick, K. C. et al. Impact of Mechanical Activation, Scar, and Electrical Timing on Cardiac Resynchronization Therapy Response and Clinical Outcomes. Journal of the American College of Cardiology (2014). doi:10.1016/j.jacc.2014.02.533
35. Kass, D. A. An Epidemic of Dyssynchrony. Journal of the American College of Cardiology 51, 12–17 (2008).
36. Yu, C. M., Lin, H., Zhang, Q. & Sanderson, J. E. High prevalence of left ventricular systolic and diastolic asynchrony in patients with congestive heart failure and normal QRS duration. Heart 89, 54–60 (2003).
37. Chalil, S. et al. Intraventricular Dyssynchrony Predicts Mortality and Morbidity After Cardiac Resynchronization Therapy. Journal of the American College of Cardiology 50, 243–252 (2007).
38. Ghosh, S. et al. Electrophysiologic substrate and intraventricular left ventricular dyssynchrony in nonischemic heart failure patients undergoing cardiac resynchronization therapy. Heart Rhythm 8, 692–699 (2011).
39. Yano, M. et al. Influence of left ventricular regional nonuniformity on afterload-dependent relaxation in intact dogs. Am. J. Physiol. 267, H148–54 (1994).
40. Spragg, D. D. Regional Alterations in Protein Expression in the Dyssynchronous Failing Heart. Circulation 108, 929–932 (2003).
41. Chakir, K. et al. Reversal of Global Apoptosis and Regional Stress Kinase Activation by Cardiac Resynchronization. Circulation 117, 1369–1377 (2008).
42. Chakir, K. et al. Mechanisms of Enhanced -Adrenergic Reserve From Cardiac Resynchronization Therapy. Circulation 119, 1231–1240 (2009).
43. Aiba, T. et al. Electrophysiological Consequences of Dyssynchronous Heart Failure and Its Restoration by Resynchronization Therapy. Circulation 119, 1220–1230 (2009).
44. Barth, A. S. et al. Cardiac Resynchronization Therapy Corrects Dyssynchrony-Induced Regional Gene Expression Changes on a Genomic Level. Circulation: Cardiovascular Genetics 2, 371–378 (2009).
45. Cho, H., Barth, A. S. & Tomaselli, G. F. Basic Science of Cardiac Resynchronization Therapy: Molecular and Electrophysiological Mechanisms. Circulation: Arrhythmia and Electrophysiology 5, 594–603 (2012).
46. Auricchio, A. Beyond the measurement of QRS complex toward mechanical dyssynchrony: cardiac resynchronisation therapy in heart failure patients with a normal QRS duration. Heart 90, 479–481 (2004).
47. Achilli, A. et al. Long-term effectiveness of cardiac resynchronization therapy in patients with refractory heart failure and ‘narrow’ QRS. Journal of the American College of Cardiology 42, 2117–2124 (2003).
48. Yu, C.-M. et al. Benefits of Cardiac Resynchronization Therapy for Heart Failure Patients With Narrow QRS Complexes and Coexisting Systolic Asynchrony by Echocardiography. Journal of the American College of Cardiology 48, 2251–2257 (2006).
49. Bleeker, G. B. et al. Cardiac Resynchronization Therapy in Patients With a Narrow QRS Complex. Journal of the American College of Cardiology 48, 2243–2250 (2006).
50. CAZEAU, S. Responders to cardiac resynchronization therapy with narrow or intermediate QRS complexes identified by simple echocardiographic indices of dyssynchrony: The DESIRE study. European Journal of Heart Failure 10, 273–280 (2008).
51. van Bommel, R. J. et al. Association of intraventricular mechanical dyssynchrony with response to cardiac resynchronization therapy in heart failure patients with a narrow QRS complex. European Heart Journal 31, 3054–3062 (2010).
52. Oyenuga, O. et al. Usefulness of Echocardiographic Dyssynchrony in Patients With Borderline QRS Duration to Assist With Selection for Cardiac Resynchronization Therapy. JACC: Cardiovascular Imaging 3, 132–140 (2010).
53. van Bommel, R. J. et al. Effects of cardiac resynchronisation therapy in patients with heart failure having a narrow QRS Complex enrolled in PROSPECT. Heart 96, 1107–1113 (2010).
54. Foley, P. W. X. et al. Cardiac resynchronisation therapy in patients with heart failure and a normal QRS duration: the RESPOND study. Heart 97, 1041–1047 (2011).
55. Muto, C. et al. A Randomized Study of Cardiac Resynchronization Therapy Defibrillator Versus Dual-Chamber Implantable Cardioverter-Defibrillator in Ischemic Cardiomyopathy With Narrow QRS: The NARROW-CRT Study. Circulation: Arrhythmia and Electrophysiology 6, 538–545 (2013).
56. Beshai, J. F. et al. Cardiac-Resynchronization Therapy in Heart Failure with Narrow QRS Complexes. N Engl J Med 357, 2461–2471 (2007).
57. Ploux, S. et al. Acute hemodynamic response to biventricular pacing in heart failure patients with narrow, moderately, and severely prolonged QRS duration. Heart Rhythm 9, 1247–1250 (2012).
58. Thibault, B. et al. Cardiac Resynchronization Therapy in Patients With Heart Failure and a QRS Complex. Circulation 127, 873–881 (2013).
59. Ruschitzka, F. et al. Cardiac-Resynchronization Therapy in Heart Failure with a Narrow QRS Complex. N Engl J Med 130902230038007 (2013). doi:10.1056/NEJMoa1306687
60. Derval, N. et al. Optimizing Hemodynamics in Heart Failure Patients by Systematic Screening of Left Ventricular Pacing Sites. Journal of the American College of Cardiology 55, 566–575 (2010).
61. Spragg, D. D. et al. Optimal Left Ventricular Endocardial Pacing Sites for Cardiac Resynchronization Therapy in Patients With Ischemic Cardiomyopathy. Journal of the American College of Cardiology 56, 774–781 (2010).
62. Ginks, M. R. et al. A Simultaneous X-Ray/MRI and Noncontact Mapping Study of the Acute Hemodynamic Effect of Left Ventricular Endocardial and Epicardial Cardiac Resynchronization Therapy in Humans. Circulation: Heart Failure 4, 170–179 (2011).
63. van Deursen, C. et al. Left Ventricular Endocardial Pacing Improves Resynchronization Therapy in Canine Left Bundle-Branch Hearts. Circulation: Arrhythmia and Electrophysiology 2, 580–587 (2009).
64. Derumeaux, G. et al. Assessment of Nonuniformity of Transmural Myocardial Velocities by Color-Coded Tissue Doppler Imaging : Characterization of Normal, Ischemic, and Stunned Myocardium. Circulation 101, 1390–1395 (2000).
65. Strik, M. et al. Endocardial Left Ventricular Pacing Improves Cardiac Resynchronization Therapy in Chronic Asynchronous Infarction and Heart Failure Models. Circulation: Arrhythmia and Electrophysiology 5, 191–200 (2012).
66. Ginks, M. R. et al. Multi-site left ventricular pacing as a potential treatment for patients with postero-lateral scar: insights from cardiac magnetic resonance imaging and invasive haemodynamic assessment. Europace 14, 373–379 (2012).
67. Pappone, C. et al. Cardiac pacing in heart failure patients with left bundle branch block: impact of pacing site for optimizing left ventricular resynchronization. Ital Heart J 1, 464–469 (2000).
68. Leclercq, C. et al. A Randomized Comparison of Triple-Site Versus Dual-Site Ventricular Stimulation in Patients With Congestive Heart Failure. Journal of the American College of Cardiology 51, 1455–1462 (2008).
69. Rogers, D. P. S., Lambiase, P. D., Lowe, M. D. & Chow, A. W. C. A randomized double-blind crossover trial of triventricular versus biventricular pacing in heart failure. European Journal of Heart Failure 14, 495–505 (2012).
70. Smith, N. et al. euHeart: personalized and integrated cardiac care using patient-specific cardiovascular modelling. Interface Focus 1, 349–364 (2011).
71. Sermesant, M. et al. Patient-specific electromechanical models of the heart for the prediction of pacing acute effects in CRT: A preliminary clinical validation. Medical Image Analysis 16, 201–215 (2012).
72. Kligfield, P. et al. Recommendations for the Standardization and Interpretation of the Electrocardiogram. Journal of the American College of Cardiology 49, 1109–1127 (2007).
73. Ypenburg, C. et al. Long-term prognosis after cardiac resynchronization therapy is related to the extent of left ventricular reverse remodeling at midterm follow-up. Journal of the American College of Cardiology 53, 483–490 (2009).
74. Steffel, J. et al. Characteristics and long-term outcome of echocardiographic super-responders to cardiac resynchronisation therapy: ‘real world’ experience from a single tertiary care centre. Heart 97, 1668–1674 (2011).
75. Matsumoto, K. et al. Regional Heterogeneity of Systolic Dysfunction Is Associated with Ventricular Dyssynchrony in Patients with Idiopathic Dilated Cardiomyopathy and Narrow QRS Complex. Echocardiography 29, 1201–1210 (2012).
76. Tatsumi, K. et al. Mechanical Left Ventricular Dyssynchrony in Heart Failure Patients With Narrow QRS Duration as Assessed by Three-Dimensional Speckle Area Tracking Strain. The American Journal of Cardiology 108, 867–872 (2011).
77. Budge, L. P. et al. MR Cine DENSE Dyssynchrony Parameters for the Evaluation of Heart Failure. JACC: Cardiovascular Imaging 5, 789–797 (2012).
78. Gold, M. R. et al. Effect of QRS duration and morphology on cardiac resynchronization therapy outcomes in mild heart failure: results from the Resynchronization Reverses Remodeling in Systolic Left Ventricular Dysfunction (REVERSE) study. Circulation 126, 822–829 (2012).
79. Ellims, A. H. et al. Utility of cardiac magnetic resonance imaging, echocardiography and electrocardiography for the prediction of clinical response and long-term survival following cardiac resynchronisation therapy. Int J Cardiovasc Imaging 29, 1303–1311 (2013).
80. Shetty, A. K. et al. Initial Single-Center Experience of a Quadripolar Pacing Lead for Cardiac Resynchronization Therapy. Pacing and Clinical Electrophysiology 34, 484–489 (2011).
81. Auricchio, A. et al. First-in-man implantation of leadless ultrasound-based cardiac stimulation pacing system: novel endocardial left ventricular resynchronization therapy in heart failure patients. Europace 15, 1191–1197 (2013).
82. Cleland, J. G. F., Tageldien, A., Buga, L., Wong, K. & Gorcsan, J. Should We Be Trying to Define Responders to Cardiac Resynchronization Therapy? JCMG 3, 541–549 (2010).
83. NIEDERER, S. A. et al. Biophysical Modeling to Simulate the Response to Multisite Left Ventricular Stimulation Using a Quadripolar Pacing Lead. Pacing and Clinical Electrophysiology 35, 204–214 (2011).

Appendix
Study 1 Flowchart – Narrow QRS Study
Study 2 Flowchart – Body Surface Mapping Study

?Or earliest available opportunity

Study 3 Flowchart – Computer Modelling Study
Timetable for completion of the thesis
Current Position
Narrow QRS study started recruiting April 2014
Body surface mapping study to start recruiting in May 2014
Computer Modelling Study awaiting final ethical and R&D approval

Narrow QRS Study timeline
Finish recruitment of 30 patients by end of September 2015
(recruitment rate of 2 patients/month)
Interim analysis December 2014
Full analysis September 2015-November 2015

Body surface mapping study timeline
Complete recruitment of 15 patients at GSTT by January 2015
Full analysis February 2015-May 2015
(recruitment rate of 2 patients/month)

Computer modelling study timeline
Retrospective database to be complete by May 2014
Prospective recruitment to commence in June 2014
Definition of computer modelling pipeline by January 2015, with analysis of predictive power on interim analysis in September 2015

Completion of 1st Draft of Thesis – January 2016
Completion of Thesis – April 2016
TO HAVE YOUR ASSIGNMENTS DONE AT CHEAPER PRICE, PLACE THIS ORDER OR A SIMILAR ORDER WITH US NOW.