Patients on haemodialysis (HD) have extremely high rates of cardiovascular disease (CVD) related mortality [
1]. US renal data system (USRDS) data suggests it is the leading cause of mortality in prevalent HD patients, accounting for 42.3 % of all deaths [
2]. There is good evidence that these excessive rates of CVD are driven by a different set of processes than in the general population and attempts to modify traditional cardiac risk factors have not improved outcomes in HD patients [
3]. According to the USRDS database, up to 64 % of all cardiac mortality among HD patients is due to sudden cardiac death (SCD) or arrhythmias [
4] of an order of around 100 times higher than the background population [
5]. Classical atherosclerotic disease is the leading cause of myocardial ischaemia in the general population [
6], but this is not the case in HD patients, who are subject to a unique set of factors that change the cardiac environment and lead to changes that alter cardiac structure and function [
7]. These changes are associated with SCD among HD patients and include: abnormalities in myocardial structure and function such as left ventricular hypertrophy (LVH) which is present in 75 % of dialysis patients; interstitial fibrosis and microvascular disease; with chronic volume overload and large volume ultrafiltration during dialysis treatments also contributing to the excess burden of CVD observed [
8‐
10]. The recurrent and frequent stresses on the heart from ultrafiltration are associated with an increase in ventricular arrhythmias [
11] that associate with SCD and raised biomarkers of cardiac myocyte damage as well as being independent predictors of HD related cardiac injury [
12] and ultimately myocardial fibrosis [
13]. To date, efforts to address these changes and improve outcomes have concentrated on medical therapies (e.g. pharmaceutical agents and implantable cardiac defibrillators) which are yet to show positive benefits [
14] and studies have also demonstrated significantly increased mortality rates in HD patients after coronary revascularisation compared to the general population confirming the differences in pathogenesis [
15], and again highlighting the limitations of current treatment options for HD patients.
The benefits of exercise for patients on haemodialysis
Exercise is not as commonly-used a therapeutic intervention in HD patients as it is in other chronic diseases, e.g. cardiac and respiratory patients, and although it is clear there are a number of potential benefits from exercise in this patient population the quality of evidence is variable and there are large gaps in the evidence base. There are several systematic reviews that summarize the potential cardiovascular benefits of exercise in HD patients, as well as the likely benefits to dialysis quality, quality of life and other health related benefits [
16,
17]. Exercise interventions have been largely divided into those that occur between dialysis sessions (interdialytic exercise) and those that occur during dialysis (intradialytic exercise). Whilst there is evidence that interdialytic training may yield superior cardio-respiratory adaptations, there is also a much a higher drop-out rate from such programmes [
18,
19]. Intra-dialytic exercise programmes are associated with significant improvements in cardio-respiratory reserve compared to control patients and have very good adherence rates [
18].
Cardiovascular disease, exercise and HD patients
In the general population, lifestyle changes that result in increased physical exercise lower mortality [
20]. Unfortunately, HD patients are less active than even sedentary healthy people with <50 % of HD patients reporting exercising once a week and unsurprisingly, higher mortality rates have been shown in such patients [
21]. Exercise training during or outside of dialysis has been shown in a number of uncontrolled and non-randomised trials to lead to significant improvements in a number of cardiovascular risk factors that predispose to SCD, both traditional and those unique to patients with end stage renal disease on HD [
22‐
24]. These studies, however, are all limited by either small sample size, non-randomised or uncontrolled design and there are no large studies that have used cardiac MRI (CMR) to assess changes in myocardial structure and function in HD patients who undergo a structured programme of exercise.
CMR in HD patients
The term uraemic cardiomyopathy has traditionally been given to a constellation of changes in cardiac structure and function seen in patients with end stage renal disease (ESRD) that include: left ventricular hypertrophy, left ventricular dilatation and left ventricular systolic dysfunction. All of these structural and functional changes have been shown to associate with poor cardiovascular outcomes [
25,
26]. It is acknowledged that studies which have used echocardiography have limited accuracy and reproducibility in defining geometric parameters and indices of systolic and diastolic function. This may be especially true in HD patients who are subject to significant changes in cardiac filling from fluid status [
27]; indeed LV mass and cavity size may be overestimated in up to 50 % of dialysis patients [
28]. CMR has been shown in HD patients to be a reliable technique for LV mass measurement, with excellent intra- and inter-observer variability for end-diastolic volume, end-systolic volume and LV mass [
29]. LV mass is a proven continuous variable in a graded relationship with cardiovascular risk [
30], and cardiovascular outcomes improve as LV mass regresses, underlying the importance of being able to accurately quantify LV mass change in intervention studies.
Echocardiography gives only limited information about tissue characterisation, compared to CMR. The risk of nephrogenic systemic fibrosis currently precludes the administration of gadolium-based contrast agents to HD patients [
31]. New native T1 mapping techniques have been shown to correlate very well with histological collagen percentage in patients with severe aortic stenosis [
32] and native T1 mapping has been shown to be reproducible in patients with Fabry’s disease and amyloidosis [
33,
34]. Native T1 mapping holds great promise in further defining pathogenesis and tissue characterisation in HD patients and patients with ESRD and CKD.
Whilst atheroma related arterial disease remains an important factor in patients on HD, arteriosclerosis is of at least equal importance. Arteriosclerosis is a process characterised by hypertrophy and increased collagen deposition in the medial layer of the arterial wall, with circumferential calcification, and commonly occurs in patients with ESRD and CKD [
35]. This causes arterial stiffness and has been shown to independently predict cardiovascular morbidity and mortality in ESRD [
36]. There is a significant amount of research examining the relationship between arterial stiffness and cardiovascular disease in patients with ESRD and CKD, with much of it derived from applanation tonometry techniques that measure aortic pulse wave velocity and measures of aortic/arterial distensibility [
36‐
38]. More recently CMR has been used to assess aortic distensibilty and there is gathering evidence of both the importance of aortic distensibility and its relationship with the development of uraemic cardiomyopathy, as well as the validity of using CMR for its assessment [
39‐
41].
Myocardial strain and strain rate have been shown to be early markers of contractile dysfunction in many conditions that precede declines in ejection fraction. Systolic strain and strain rates have traditionally been assessed with CMR using tissue tagging techniques [
42]. Increased circumferential basal strain and strain rates and reduced longitudinal function may occur in non-diabetic CKD patients, before any other structural or functional changes are apparent; suggesting it may be a very early indicator of uraemic cardiomyopathy [
43]. Whilst tissue tagging is proven to be reproducible, scan acquisition requires additional long breath-holds and analysis can be cumbersome. Newer methods of strain and strain rate analysis are now available that can assess LV strain and strain rates directly from cine images. Our group has shown that Feature Tracking has excellent reproducibility and maybe more robust than tissue tagging in acute myocardial infarction patients [
44,
45]. There are currently no studies that have used CMR to assess the effects of exercise on cardiac structure and function in HD patients.
The primary aim of this study is to investigate the effects of a six month programme of intra-dialytic exercise on LV mass as assessed by CMR. We will assess other cardiac structural and functional end-points, as well as biochemical markers of acute and chronic cardiac dysfunction, anthropometric measurements and measures of physical function and quality of life. An important secondary aim is to establish whether intradialytic exercise training is associated with an increase in cardiac arrhythmia, thereby addressing one of the major safety concerns.