Methods

Trial design

We conducted this prospective, multicentre RCT in five regions of the UK.

Ethics approval and research governance

The trial protocol was approved by the relevant health authorities and institutional review boards, and all patients provided written informed consent. London Fulham Research Ethics Committee approved the protocol (reference 14/LO/1851). The study was prospectively registered as ISRCTN83508514. An independent data and safety monitoring committee performed regular safety surveillance. Data were entered into an electronic case report form by the investigators and were analysed at the Robertson Centre for Biostatistics, University of Glasgow.

Patient and public involvement

The planning and delivery of the PEDAL trial was facilitated by close engagement with people living with CKD and utilising HD therapy. Patients contributed directly to both the development of trial materials (e.g. participant information leaflets and consent forms) and the way the trial was conducted. INVOLVE good practice guidelines23 were followed to ensure service-user leadership in the trial delivery and dissemination of the findings. INVOLVE is a national advisory group established and funded by the NIHR to support active public involvement in NHS, public health and social care research.23

Participants

Inclusion criteria

• Prevalent patients with stage 5 CKD receiving maintenance HD therapy for > 3 months who were male or female, aged > 18 years and able to provide written informed consent.

Exclusion criteria

• Patients unlikely to be receiving HD for > 6 months (this includes cachectic patients and those with severe heart failure (New York Heart Association Functional Classification ≥ 3)].
• Patients for whom dialysis withdrawal was being considered.
• Patients likely to receive a live donor transplant or be transferred to peritoneal dialysis in the period of time.
• All patients within 3 months of initiation of HD. (Patients in this time frame are generally less clinically stable, many having vascular access procedures performed, and have much higher rates of intercurrent events, including death and hospitalisation).
• Patients deemed to be clinically unstable by their treating physician.
• Patients with bilateral lower-limb amputations.
• Patients with dementia or severe cognitive impairment, and other patients unable to give informed consent.
• Patients with psychiatric disorders (who are not treated and stable).
• Patients who were pregnant.

Screening and eligibility

The majority of potential participants for the PEDAL trial were identified during routine HD clinical consultations and concurrent evaluation of clinical records to confirm eligibility for participation. Patients already established on HD for > 3 months were eligible and easily identified from hospital databases and dialysis logs. If considered eligible for the study, they were approached by a member of the renal health-care team who discussed the study and provided them with a participant information sheet with further details. After allowing the patient a minimum of 24 hours to read and consider the information in the participant information sheet, and to consult with family members, the research team approached the patient (usually during the next dialysis session) to answer any questions. If the patient agreed to proceed, an appointment was made for familiarisation and baseline outcome assessment sessions. Written informed consent was obtained by a member of the research team prior to any study assessments.

Randomisation and blinding

Randomisation was conducted using a centrally controlled, web-based randomisation system run by the Glasgow Clinical Trials Unit (GCTU). To ensure balanced assignment across critical variables, a minimisation algorithm was employed, taking into account baseline age, sex and diabetes status. It was impossible to blind the ‘treating’ physiotherapy assistants or the participants and, thus, the study implemented a blinded outcome assessment and analysis.

Intervention

Intradialytic exercise training

The intradialytic exercise prescription was based on current physical activity guidelines for the elderly24 and for people with diabetes25 and cardiovascular disease.26 These recommend that a minimum target amount of 1000 kcal per week is expended in physical activity for health benefits, with optimal benefits associated with a weekly target physical activity accumulation of 1500–2000 kcal or at least 150 minutes per week at moderate exercise intensity and at least 2 days per week of resistance-based training for muscular endurance and strength gains. Because opportunity for structured, prescribed exercise was largely restricted to patients’ 3 days of HD, the training aim was for patients to accumulate a level as close as possible to or exceeding this minimum threshold level of 1500 kcal per week through intradialytic exercise. A target overall volume of exercise was calculated in energy expenditure units (kcal per week) using the frequency, intensity, time, type (FITT) principle and provided for all patients, with an individualised progression plan towards this goal.

The following progression plan was devised as a general guide:

• Introductory/adoption phase: a minimum cycling duration of 21 minutes per dialysis session was used as a starting point target. By the end of the first 4 weeks, all patients were expected to cycle for 21 minutes either in two 10-minute bouts or continuously at the low end of the moderate exercise intensity range. The progression requirements were that, by week 8, all participants should be able to cycle for ≥ 21 minutes continuously on the cycle ergometer (63 minutes per week) at the low end of the moderate exercise intensity range. The overall minimum target volume of exercise was set at 140 kcal per cycling session.
• Progression/adoption phase: expected to last 12–14 weeks. During these weeks, emphasis was placed on progressing exercise stimulus mainly through duration. A reasonable and achievable exercise target for this group would be a minimum cycling duration of 21–30 minutes per dialysis session at moderate intensity. The overall minimum target volume of exercise was set at 1000–1200 kcal per week or ≈ 170 kcal per dialysis exercise session.
• Behaviour development phase: expected to last up to 24 weeks. All patients should be able to achieve a target duration ranging from 30–40 minutes per dialysis session (90–120 minutes per week) at moderate to vigorous exercise intensity (55–70% VO_2 reserve). The overall minimum target volume of exercise was set at ≈ 1500 kcal per week (or ≈ 214 kcal per dialysis session) and this would be achieved through adjustments to duration and intensity.

The prescribed individualised training intensity was derived from a peak aerobic capacity (VO_2 peak) assessment, using a 1-minute ramp incremental protocol on a cycle ergometer. New exercise intensity ranges were established at the 3-month follow-up assessment time point. Exercise prescription was set at a workload corresponding to 40–75% of VO_2 reserve. We also made a note of rate of perceived exertion, heart rate and blood pressure responses corresponding to these ranges of exercise intensity during the incremental cycle testing protocol and used these indices to guide and monitor progression until the next planned assessment point (3 months), at which time the exercise prescription was renewed.

Using a modified and custom-made Monark cycle ergometer (HaB International Ltd., Southam, UK), aerobic exercise was performed, in a semi-recumbent position, three times per week during the first 2 hours of HD. Exercise duration and intensity were recorded and monitored for each exercise session in exercise diaries. In addition, ratings of perceived exertion, BP and HR were recorded during the training sessions. The energy expenditure goals were deliverable through a progressive increase in intradialytic cycling from short bouts of 8–10 minutes to bouts of 21 minutes and bouts of 40 minutes or more, at the prescribed exercise intensity, resulting in 55%, 69% and 75%, respectively, of the target weekly minimum physical activity volume of 1000 kcal being achieved. Twice per week participants also completed lower-extremity muscular conditioning exercise, using ankle weights, after aerobic cycling exercise. Physiotherapy assistants (band 4 technical instructors) were employed in each region to deliver the intradialytic intervention. This role, supervised and quality assured by a regional co-ordinator, involved the technical implementation of the exercise prescription produced by the regional research assistant, who was blinded to treatment allocation.

Usual care: haemodialysis therapy

Usual care was based on UK Renal Association guidelines for HD2 and included management of blood pressure, treatment of anaemia, phosphate control and cardiovascular risk mitigation strategies. For the purposes of the trial, we specified that usual care, in both arms of the trial, should allow all of these treatments to continue unchanged so that we were investigating any additional benefit of the intradialytic exercise training intervention compared with usual care.

Adherence

We attempted to minimise the loss to follow-up in this study by (1) emphasising to participants the importance of their attendance at follow-up assessments even if they were no longer compliant with the intervention, (2) reducing outcome assessment appointments to a maximum of two non-dialysis day visits, (3) using a reminder protocol for non-dialysis day assessment appointments that utilised prompts by the dialysis unit staff, letters and telephone contact, (4) providing travel remuneration (including, where necessary, taxi costs), and (5) the provision of training in issues related to compliance for all study staff who came into contact with the participants.

Primary and secondary outcomes

Primary outcome

The primary end point for this study was change in KDQoL-SF PCS score between baseline and 6 months. The KDQoL-SF is a disease-specific QoL measure that includes the SF-36 as a generic core plus symptoms/problems of kidney disease scales. The SF-36 has 36 items compiled into eight scales: physical functioning, role functioning/physical, bodily pain, general health, vitality, social functioning, role functioning/emotional and mental health. These scales are scored from 0 to 100 AUs; a higher score is more positive (i.e. less pain or less limitation). Normalised scores representing overall physical functioning and mental functioning are calculated from the individual scales and are presented as PCS and mental component summary (MCS) scores. The PCS includes the dimensions of physical functioning, role functioning/physical, bodily pain, general health, vitality and social functioning.

Secondary outcomes

Quality of life

A documented change in EQ-5D-5L score from baseline to 6 months is a key QoL secondary outcome.

Peak aerobic capacity

VO_2 peak was determined by an incremental cycling exercise tolerance protocol. Breath-by-breath gas exchange was measured using cardiopulmonary exercise testing equipment calibrated prior to each patient assessment. The exercise testing protocol started with a 3-minute unloaded cycle, followed by ramp increases in resistance of 15 watts per min until one of the following occurred: (1) plateau in oxygen uptake, (2) attainment of respiratory exchange ratio ≥ 1.15 or (3) patient request to stop. Average oxygen uptake during the final 20 seconds of the test was recorded as the VO_2 peak. Electrocardiogram and heart rate were continuously monitored, and blood pressure was recorded every 1 or 2 minutes throughout the ramp incremental test. Rate of perceived exertion and angina scale were recorded every minute for safety.

Physical performance tests

Patient physical function was assessed by the sit-to-stand 60 (STS60) test27 and the 10-metre Timed Up and Go (10mTUG) test,28 both of which have been used as accurate and valid measures of lower-leg strength, balance, co-ordination, gait speed and physical function in HD patients.

Anthropometric measures

Measures of height, body mass, BMI and waist circumference were performed.

Cardiovascular risk

Carotid–femoral PWV is considered the gold standard for non-invasive arterial stiffness assessment in clinical practice29 and has been suggested as a surrogate cardiovascular end point. This was assessed with the Vicorder^® system (Skidmore Medical Limited, Bristol, UK). The Vicorder system is small, portable, non-invasive and non-operator-dependant. In addition, it was available in all centres to ensure comparability of the data. Conditions for assessment, as described in the expert consensus statement by Laurent et al.,29 were adhered to for all measurements. The measurement protocol by Hickson et al.30 was used, mathematically removing the additional femoral segment from the Vicorder standard protocol to correct for any inherent bias at high arterial PWV. An average of three measurements (of 20 consecutive signals) was recorded at each time point. Resting predialysis blood pressure was assessed.

Physical function, physical activity and fear of falling questionnaires

Patients completed the following questionnaires to capture data about physical activity, activities of daily living and falls: the International Physical Activity Questionnaire Long Form,31 the Duke Activity Status Index (a self-reported, 12-item questionnaire that assesses activities of daily living)32 and the Tinetti Falls Efficacy Scale (measuring fear of falling).33

Blood tests

Clinical blood tests were collected pre dialysis and included haemoglobin, serum phosphate and parathyroid hormone.

Medication

Dosages of erythropoiesis-stimulating agents were recorded.

Safety and monitoring of the intervention

Safety and monitoring data included discontinuation from the exercise intervention and permanent study withdrawals with reasons, compliance with the exercise programme and adherence to the exercise prescription and SAEs.

Engagement and fidelity to exercise prescription

General engagement with the exercise intervention was defined as the percentage of exercise sessions completed of the total prescribed. Compliance was calculated as the number of sessions completed at week 13 and 26 of the exercise programme divided by the number of expected sessions completed (39 and 78 for weeks 13 and 26, respectively) and multiplied by 100. Acceptable compliance was defined as > 70% compliance and poor compliance was defined as < 50% compliance. To monitor adherence (fidelity) to exercise prescription in terms of intensity and volume of exercise achieved, data on duration of cycling (excluding warm up and cool down), power output, number of sets and repetitions for muscle conditioning exercises were recorded in individual participant exercise diaries for each exercise session. Drop-out/temporary cessation of exercise of > 2 weeks with reasons were also recorded in the individual exercise diaries.

Sample size

Primary outcome

A sample size calculation, based on an assumed difference of 4 AU in KDQoL-SF PCS score AU and a standard deviation (SD) of 10 AU as seen in the study by Painter et al.,7 and conservatively comparing 6-month KDQoL-SF PCS scores between groups by two sample t-test (two-sided 5% significance level) resulted in a sample size of 133 completers per group, with 83% power to detect a mean difference of 4 AU in KDQoL-SF PCS score. To allow for 30% loss to follow-up over 6 months, as seen in phase II trials,25 190 participants per group were suggested to be randomised. This sample size calculation did not take into account adjustment for baseline PCS scores. Subsequent analysis by the GCTU of within-trial change in PCS score from baseline, adjusted for baseline levels and randomisation minimisation variables, suggested that the study will have 80% power to detect a 4-point difference with only 87 participants per group with complete data at baseline and 6-month follow-up. Likewise, 115 participants per group would be needed if a zero score is imputed for deaths prior to 6 months.

Key secondary outcome

Analysis of blinded within-trial data adjusting for baseline levels and randomisation minimisation variables suggest that a study of 44 participants per group would have 90% power to detect the minimum clinically important difference in peak aerobic capacity of 3.5 ml/kg/min (1 MET).

Statistical analyses

A full statistical analysis plan was signed off prior to database lock and study unblinding. Descriptive statistics of clinical and sociodemographic variables at baseline are presented split by treatment group.

Primary outcome

The primary outcome measure (change from baseline to 6 months in KDQoL-SF PCS score) was compared between the control and intervention groups using a normal linear model, adjusting for baseline KDQoL-SF PCS score and the randomisation minimisation variables. The findings are presented as the adjusted mean difference, with 95% confidence interval (CI), between the treatment groups. The main analysis was carried out on research participants with PCS assessments at baseline and at 6 months. Two sensitivity analyses were also carried out, first imputing a score of zero for those who died prior to 6 months and, second, based on all participants with a baseline PCS score using the method of multiple imputation.

Secondary outcomes

Other continuous outcomes were analysed as for the primary outcome.

Binary outcomes were compared between treatment groups using logistic regression models, adjusting for baseline value. The results are reported as the adjusted odds ratio with 95% CI.

Time-to-event outcomes were calculated as time from randomisation and compared between treatment groups using Cox proportional hazards regression models. The results are reported as the adjusted hazard ratio for intervention versus control with 95% CI.

Data involving counts of events were compared between treatment groups using negative binomial regression models, adjusting for length of follow-up. The results are reported as the adjusted treatment effect with 95% CI.

Subgroup analyses

In the intervention group, the primary outcome was compared between those who had completed < 30% of the expected exercise sessions, those who had completed ≥ 30% but < 50% of the expected exercise sessions, those who had completed ≥ 50% but < 70% of the expected exercise sessions and those who had completed ≥ 70% of the expected exercise sessions. This was done using a linear regression model, predicting the primary outcome from the above categories for the expected exercise sessions completed, adjusting for minimisation variables.

The primary outcome was also assessed, comparing the control group with the three completer groups defined above. In addition, the relation of the percentage of sessions completed at 6 months follow-up to the primary outcome were analysed in two ways:

1. as a linear regression predicting the primary outcome from the percentage of sessions completed at 6-month follow-up, adjusting for baseline PCS score and minimisation variables, in the intervention group only
2. as a linear regression predicting the primary outcome from the percentage of sessions completed at 6-month follow-up, adjusting for baseline PCS score and minimisation variables, in all patients, setting the percentage of sessions completed to 0 in the control group.

Safety analyses

Discontinuations from the intervention and permanent study withdrawals and their reasons were tabulated, as was adherence to the exercise prescription. SAEs were tabulated by system organ class and body system.

Data monitoring and quality assurance

The trial was coordinated by a Trial Management Group (TMG), consisting of the chief investigator, GCTU Assistant Director and senior clinical trial manager and a statistician. The trial manager co-ordinated the study and was accountable to the chief investigator. The central trial office (King’s College Hospital and GCTU) provided support to each site. The central trial office was responsible for randomisation, collection of data in collaboration with the research co-ordinator, data processing and analysis. Publication and dissemination of the study results was co-ordinated by the trial office in collaboration with the chief investigator and investigators. A Trial Steering Committee was established to oversee the conduct and progress of the trial. The study’s funder – the NIHR Health Technology Assessment (HTA) programme – formally appointed the chairperson and members after the nominations from the TMG. A charter was drawn up to describe membership, roles and responsibilities of the Trial Steering Committee. The Independent Data Monitoring Committee was an independent group of experts, consisting of a nephrologist, physiotherapist, lay member and a statistician, who monitored patient safety and treatment efficacy data while the clinical trial was ongoing; the primary mandate of this committee was to protect patient safety.