Aims: To explore the clinical effect on exercise tolerance and quality of life, safety and tolerability of pacing at higher outputs in patients with chronic heart failure and a pacemaker.

Background: Heart failure (HF) is a very common condition of breathlessness or fatigue associated with heart muscle weakness. In around 30% of people with HF, a pacemaker-based treatment known as cardiac resynchronization therapy (CRT) can improve symptoms and prognosis by retuning the timing of the contraction of the heart. However, the effect of CRT is variable and unpredictable, with around 1 in 3 of people obtaining no obvious symptomatic benefit. One of the reasons for this might be that the pacemaker pulse does not activate all of the heart muscle cells at the same time or at all. In order to provide the longest possible battery life span, the default programming for all pacemakers is to provide a stimulus at an arbitrary level above the capture threshold (at which the spike leads to contraction). Whilst this is reasonable in a normal heart where the aim is to treat a slow heart rate, in heart failure, where the aim is to retune all parts of the heart, it is possible that this is not enough to provide consistent contraction of all heart muscle cells. It is possible that providing a higher output electrical signal from the pacemaker will activate more of the heart muscle cells immediately and thereby improve the contraction of the heart. The investigators think that this might be important at rest, but even more important during activity. This concept has never been tested before in a systematic manner but could have large implications for people with heart failure and existing CRT devices which could simply be reprogrammed to derive greater benefits for patients during everyday activities.

Design: The proposed project has two parts:

Study 1 - 105 patients with a CRT pacemaker for heart failure but ongoing symptoms will be invited to attend the National Institute of Health Research Clinical Research Facility. Symptoms, medication, hospitalisation information will be collected and a heart ultrasound scan using the pacemaker to increase the heart rate will be done to describe the force frequency relationship. Patients will perform a cardiopulmonary exercise test.

Of these patients, 40 will be invited to return for two further visits, to perform an exercise test each time with the pacemaker programmed to its usual output or high output pacing. At each visit, including the heart scans, the order of the programming will be random, and neither the observer nor the patient will know how the device has been programmed.

Study 2 - 70 patients will be invited to participate in a longer term study of whether high output pacing is safe, well tolerated and has effects on walk time (on a treadmill) and heart pumping function. Participants will be randomly allocated to one of two groups: high output or standard pacemaker settings. In the high output group, the pacemaker will be programmed to deliver the highest output possible or tolerated. In the standard care group patients will have standard output settings.

ORIGINAL HYPOTHESIS Increasing pacemaker left ventricular stimulation output is safe and well-tolerated, and increases acute and longer term exercise capacity and quality of life through improved left ventricular function over a range of heart rates.

The project consists of two closely related work packages.

AIMS -WORK PACKAGE 1

1. To determine whether high output pacing from the left ventricular (LV) lead in CRT recipients acutely increases left ventricular contractility over a range of heart rates and the baseline clinical features that predict this;
2. To establish the proportion of patients in whom high output pacing is not possible due to phrenic nerve stimulation;
3. To determine whether high output pacing acutely improves exercise time on a treadmill.

AIMS -WORK PACKAGE 2

1. To determine whether longer term (6 months) high output left ventricular pacing is associated with patient orientated benefits on treadmill exercise time and quality of life
2. To determine whether longer-term (6m) high output left ventricular pacing is safe and tolerated and what effects this approach has on battery longevity

BACKGROUND Chronic heart failure and cardiac resynchronisation therapy Even when prescribed optimal medical and device therapy, patients with chronic heart failure suffer a persistent reduction in quality of life mainly due to breathlessness and fatigue on exercise. The origin of these symptoms is multifactorial, but despite increasing appreciation of the influence of peripheral adaptations on exercise tolerance the key factor in initiation and persistence of the symptoms is impaired cardiac contractility both at rest and over the relevant heart rate range. Whilst cardiac resynchronization therapy (CRT) provides a powerful adjunct to medical therapy in the third of patients with dyssynchronous contraction, many patients remain symptomatic despite CRT, prompting a series of pacing adaptations to try and improve the technology further. Standard left ventricular leads now include four poles from which a pacing stimulus can be delivered selectively or in parallel (multi-point pacing), and each of the manufacturers includes automated software to adjust the intraventricular and interventricular timing. Despite this, the 'response rate' in clinical practice, however measured, remains solidly at around 30%. Although overall, these advances might be neutral, more pacing electrodes, and automatic programs to measure and adjust timing come at a cost of battery longevity. To counter this many systems now include automatic capture assessment algorithyms that adjust the pacing stimulus downwards in response to successful LV capture and upwards in its absence. These have demonstrated reduced average pacing stimulus amplitude.

Although device longevity is a key contributor to cost-effectiveness, battery longevity may not be an important variable for people with heart failure, for whom the primary aim is to maintain or improve their quality of life. Data from 76 consecutive patients implanted with a CRT device between 2008-2010 demonstrate that over 60% did not survive long enough to require a second device. Failure to survive to generator replacement was higher in people with worse LV function, worse symptoms, and those in whom there was no symptomatic improvement. Our data suggest therefore that in contrast to the situation for pacemakers implanted to treat bradycardia, battery longevity might be less important in people with heart failure. Our Patient and Public Involvement and Engagement group was indeed split along these lines. People with a good quality of life want a long battery life, whereas those with persistent symptoms were content to accept a reduced battery longevity if this would benefit their symptoms.

How to determine force-frequency relationship in humans Most work on the Force Frequency Relationship (FFR) including the expression and function of cellular mechanisms and contractile responses has been done ex-vivo in myocardial strips either from explanted hearts during transplantation, or from biopsy material. Assessment of contractility, and the FFR specifically, in vivo requires a controlled increase in heart rate, and a reliable measure of contractility which can be achieved non-invasively using cardiac ultrasound to measure LV end-systolic pressure (LVESP) and end systolic LV volume (LVESV). These are divided by each other (LVESP/LVESV) to give an end-systolic pressure volume ratio (LVESPVR). LVESP and LVESV can be measured by invasive and estimated by non-invasive techniques. Following discussion with our PPI forum to check acceptability, I selected an echocardiographic (cardiac ultrasound) approach. Using cardiac ultrasound, contractility can be estimated either from 2-dimensional images or by tissue Doppler imaging. Using 2D images to measure the end-systolic volume index (LVESVi =LVESV/Body surface area (BSA)), and using SBP as a surrogate of the LV end-systolic pressure one can then estimate contractility as SBP/LVESVi. This surrogate of contractility has been validated against invasive methods. , Repeated measurements can be taken and the slope of the FFR can then be calculated as the ratio between SBP/LVESVi change from baseline / HR increase from baseline. Critical heart rate (optimal heart rate) is the heart rate at which the SBP/LVESVi reaches the maximum value or that at which beyond the SBP/LVESVi has declined by 5%. In a negative test, (one where there is no increase in contractility with HR) the critical heart rate is the baseline heart rate. Non-invasive methods require an assumption that LVESP is closely related to systolic blood pressure (SBP). This introduces an approximation, especially in younger subjects, but mostly there is a tight relationship between peripheral systolic and end-systolic LV pressure and it will be assumed that any error is systematically distributed along the whole FFR within individuals.

Rationale for increasing pacemaker output in people with heart failure Limited published data suggest that higher output pacing can lead to improved timing of depolarization and better LV hemodynamics. Whether this will translate into improved exercise tolerance or quality of life with acceptable loss of battery longevity is unknown. This balance is likely to depend upon individual patient factors and the degree of improvement seen within an individual. The present project aims to provide information on efficacy in the whole population, and in groups of individuals, whilst also allowing some description of the mechanisms of the effect through examination of changes in cardiac function at rest and the force-frequency relationship as a surrogate of exercise.

This study will answer several of these questions, and provide key pilot data for a larger study that will explore the benefits of optimizing patients' pacemaker programming to their individual situation.

PLAN OF INVESTIGATION Workpackage 1 Part A Introduction The present proposal will use existing CRT pacemakers or defibrillators in patients with CHF to explore the clinical effects of increased pacing output from the left ventricular lead on cardiac function and exercise time on a treadmill and also to establish whether the changes in response to the intervention relate to baseline clinical such as etiology, co-morbidities and severity.

Methods: 105 patients with stable CHF due to left ventricular systolic dysfunction with a left ventricular ejection fraction <50%and a CRT device will be invited to attend the clinical research facility at Leeds General Infirmary, aiming for 90 full datasets. Each patient will undergo and a resting echocardiogram pacemaker interrogation to check for phrenic nerve stimulation and to establish options for pacing vectors. Demographic and clinical data (including duration of heart failure symptoms), medical therapy, current symptomatic status, resting HR and blood pressure will be recorded. Patients without a cardiopulmonary exercise test from <6 months ago in their clinical record will then undergo an exercise test to describe peak oxygen consumption and exercise time.

Medical therapy Medical therapy will be continued throughout. Although beta-blockers might blunt the force frequency induced contractile response in the normal heart in CHF, HR reduction might improve FFR counteracting the negative inotropic properties of ß-blockers. Digitalis also has effects on the FFR by increasing intracellular sodium levels, which enhances calcium influx, restoring the FFR. , which may explain the positive inotropic effect of cardiac glycosides aside from their heart rate limiting properties. However, these agents are essential in CHF and information gained in their absence would not reflect the usual situation for most CHF patients.

Atrial rhythm Atrial fibrillation (AF) is a common dysrhythmia in CHF. Whether patients with AF have an abnormal FFR is unknown. Patients with AF will not be excluded unless heart rate is poorly controlled (>80bts/min).

Pacing protocol Images will be collected at rest as described above, following which randomisation will be undertaken by the Leeds Clinical Trials Research Unit who will provide a telephone service to determine the order of the programming. The unblinded cardiac physiologist will then program the pacemaker device to a 'high' output setting (aiming to extend the pulse width and increase the amplitude to the maximum tolerated) or a 'standard' output setting (with no change to baseline settings). The patient and the echocardiographer will be unaware of the allocation. Atrial pacing will be initiated in the AAI-mode at 45 beats/min (or the next highest 'round figure' above the baseline heart rate). After four minutes, images will be recorded, and the pacing rate will then be increased in a stepwise 15-beat interval with images recorded after every four minutes. This step-wise increase will be repeated until the maximum predicted heart rate as per the calculation by Astrand (220-age) is reached. At this point peak data will be collected and pacing will then return to baseline settings. Five minutes after the end of atrial pacing, a final set of images and a blood pressure will be recorded. Angina pectoris will also stop the test and the heart rate will be allowed to return to normal. After 10 minutes, this procedure will be repeated with left ventricular lead output programmed to the other setting (either high output or standard output).

BP measurement Systolic pressure (SBP) measured using a manual blood pressure cuff and a standard stethoscope will be used as a surrogate for end-systolic LV pressure. The SBP will be recorded as the point where the first tapping sound occurs for 2 consecutive beats.

Image recording and image analysis Full baseline echocardiography will be carried out with grey-scale and tissue Doppler images recorded in the two and four chamber views using harmonics to improve border definition if necessary. Further images will be recorded at each 15 beat frequency increase during the protocol. Images will be stored in the 'echopac' digital imaging system and analysed offline in an anonymous and randomised fashion (with HR data removed). All echocardiographic analyses will be performed offline with the observer blinded to the clinical status of the patient. This analysis will include a calculation of LV end diastolic and end systolic volumes using the biplane discs (modified Simpson's) method by tracing the endocardial border excluding the papillary muscles. An average of three measurements will be used in the final analysis. The frame at the R-wave will be taken as end diastole, and the frame with the smallest LV cavity, as end systole.61 The slope of the exercise ejection fraction will be calculated with the linear best fit from the stress ejection fraction values. The LV end-systolic volume index (LVESVi) will be calculated at each stage as LVESV/body surface area.

FFR calculation The contractility at each HR will be calculated as previously described using: [SBP/LVESVi] and a smoothed graph plotted for each patient to define peak contractility, the slope of the FFR, and the optimal HR for contractility. The slope of the FFR will then be calculated as the ratio between SBP/LVESVi increase (from baseline to optimal heart rate)/HR increase (from baseline to optimal heart rate). FFR will be defined as up-sloping when peak exercise SBP/LVESVi is higher than baseline at intermediate stress values (Figure 2), and biphasic when an initial up-sloping is followed by a down-sloping trend). In a biphasic pattern, the optimal heart rate will be the heart rate beyond which SBP/LVESVi declines by 5%. The FFR slope will be classed as negative if the optimal heart rate is the starting heart rate i.e. the slope is downgoing. The investigators expect reproducibility to be satisfactory since this method of left ventricular volume calculation during echo is widely used and accepted. In addition, the evaluation of end-systolic volume has a higher reproducibility than end-diastolic volume from echo images, and only the former will be used in the calculation. 10 randomly selected patients will be invited for a second visit and image analysis will also be repeated in 10 further randomly selected patients to document reproducibility of data collection and analysis.

Statistical considerations Although the data collection is straightforward, the later stages of the analysis will require complex statistical modelling. Pre-specified subgroups will be patients with and without ischaemic heart disease, patients with and without type II diabetes mellitus, patients with and without an echocardiographic response since implant (improvement in left ventricular ejection fraction >5% OR improvement in left ventricular end systolic volume index >15%) and patients who do or do not feel themselves to have had a relevant improvement in symptoms following their CRT implant. In this way it might be possible to identify subgroups in whom the intervention might be most useful.

Data analysis plan As described, a measure of contractility will be collected at each heart rate interval. These can be plotted against heart rate for each patient to achieve three novel variables per patient (peak contractility, heart rate for peak contractility and the slope of the FFR). A curve will be created for each individual with standard LV output and then compare this with the same curve created from data collected during high output pacing. The difference in the three key variables between the two curves will then be compared.

A further key secondary aim is to explore the relationship between key baseline clinical variables: 1) heart failure aetiology (ischaemic/non-ischaemic), diabetes mellitus (Y/N), and baseline ejection fraction (as a continuous variable) and the influence of high output left ventricular pacing on cardiac contractility over the critical heart rate range that is optimal for that individual.

Part B:

Study design: This will be a randomised placebo-controlled cross-over pilot study designed to establish the relevance to patients of exercising with high output programming. The randomisation order provided for the echocardiogram assessment (Part A) for this participant will be used a second time.

Methods: 40 consecutive attendees will be enrolled to provide 25 evaluable paired data sets accepting up to a possible 35% drop-out rate from this part of the study due to intolerance of the high-output setting. Patients will undergo two cardiopulmonary exercise tests one week apart (but at the same time of the day) on the same protocol, during which their device will be programmed in a random order to provide either high output pacing on the left ventricular lead or normal settings. During the test, patients will score their symptoms. During each exercise test the unblinded cardiac physiologist will monitor the electrocardiograph. Neither the blinded observer nor the patient will have sight of the electrocardiograph. This arrangement has worked well previously.

Data analysis plan: Data from the cardiopulmonary exercise tests regarding symptoms (from the Borg scale) can be related to objective measures of ventilation, oxygen consumption and workload. The difference in the force frequency relationship through high output pacing found during echocardiography can then be related to the differences seen during the exercise tests.

Work package 2:

Introduction: The primary aim of this work package is to determine the longer term effect of high output pacing while also providing information on safety, tolerability and battery longevity.

Study design: This will be a randomised, controlled parallel pilot study where the comparator will be standard output programming.

Methods: 70 consecutive attendees will be enrolled to provide 50 evaluable paired data sets (25 per group) accepting a possible 30% drop-out rate from this part of the study due to intolerance of the high-output setting and longer follow-up period - lower because participants in Workpackage 1 Part B that did not tolerate high-output pacing are less likely to be asked to participate in Workpackage 2. Randomisation will be undertaken by the Leeds CTRU after the baseline assessment, (echocardiography, exercise test, blood tests including B-type natriuretic peptide, and pacemaker battery longevity), who will provide a telephone service to determine allocation and balance critical baseline variables using minimization: diabetes mellitus (Y/N), aetiology (ischaemic/non-ischaemic). The unblinded cardiac physiologist will program the pacemaker device to a 'high' output setting (aiming to extend the pulse width and increase the amplitude to the maximum tolerated) or a 'low' output setting (with no change to baseline settings). Patients will receive a telephone call at one week to ask about angina and phrenic nerve stimulation and then again at 6 months when the baseline assessments will be repeated.

Statisical considerations: While accounting for a drop out from the second test due to withdrawal of consent of 30%, to achieve 25 paired evaluable datasets in each group, 70 patients will be recruited.

Data analysis plan: The primary analysis will focus on change in exercise time, with key secondary endpoints of change in 1) quality of life 2) modified Packer score and 3) left ventricular structure and function at 6 months. More complex analysis will assess whether patients' response specifically to exercise time or left ventricular remodeling vary and whether these relate to the degree of change of the force frequency relationship or contractility measures. This will help to identify people who may benefit more from the programming change. In this way additional information for the personalization of pacemaker programming will be provided, since high output pacing might have an adverse effect on battery longevity such that perhaps only patients with a clinically relevant improvement or with persistent symptoms are put forward for such reprogramming.

Inclusion Criteria:

1. Symptomatic heart failure due to left ventricular systolic dysfunction (LVEF<50%)
2. Willing and able to give informed consent
3. CRT-D or CRT-P device in situ (>6 months)

Exclusion Criteria:

1. Angina pectoris symptoms limiting exercise tolerance
2. Unstable heart failure symptoms (medical therapy changes in last three months)
3. Poor imaging quality (details of patients excluded for this reason will be recorded)
4. Calcium channel blocker use (these depress the force of contraction at all heart rates)
5. Uncontrolled heart rate (either atrial fibrillation or sinus rhythm (>80bts/min))
6. Co-morbidities significantly affecting exercise or symptoms such as severe chronic airways disease or arthritis
7. Haemodynamically significant aortic stenosis