Design
Post Ischaemic Stroke Cardiovascular Exercise Study - Zoom Delivered Intervention Against Cognitive Decline (PISCES-ZODIAC) is a multicentre Phase IIb assessor-blind randomised controlled trial testing the preliminary efficacy and feasibility of an eight-week home-based exercise intervention modelled on cardiac rehabilitation, delivered at two-months after stroke. Participants are assessed before (at two-months post-stroke: t1) and after (at four-months post-stroke: t2) the exercise intervention and again at 12-months post-stroke (t3). We will report the trial according to the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines (Table 1) (23) and the Template for Intervention Description and Replication (TIDieR) (24). It is registered with the Australian and New Zealand Clinical Trials Registry (12616000942459).
Study setting
Participants are recruited from four metropolitan university teaching healthcare networks in Melbourne, Victoria, Australia: Austin Health, Eastern Health, Epworth Health, and Western Health. Central ethics approval has been granted by the Austin Health Human Research Ethics Committee (lead site, HREC/16/Austin/45), with site-specific governance approval from others (most recent approval 07 September 2023, Protocol_v14). Austin, Eastern and Western Health sites are public hospital networks, while Epworth Health is a private hospital network with sites throughout the state of Victoria.
Participants
Eligibility criteria
The patient population remains the same in the current protocol: adult ischaemic stroke survivors without severe prior functional or any reported cognitive impairment. The inclusion and exclusion criteria are summarised in Table 2. We placed a one-hour travel radius around Melbourne Central Business District for our participant residence catchment (~ 100 km) for logistical reasons (courier truck delivery costs for exercise equipment; attendance at 3 in-person study visits).
Screening and consenting
Information flyers are handed to potential participants by clinical staff where appropriate. Screening is done remotely via accessing hospital discharge summaries or direct notification from clinical staff. Participants are contacted for their interest. Initial consent can be completed via video-call or phone call. Further information is then obtained regarding MRI eligibility (e.g., medical notes for prior procedures, etc.) after initial consent. Written consent is provided at first study visit and their study schedule is arranged.
Assessments
Outcome measures for each time point are listed in Table 1 and Supplementary Table 1. To accommodate a range of participant needs, and to adapt to COVID-19 restrictions and risk of infection, participants are given the option to complete the questionnaires and neuropsychological test battery either face-to-face or remotely, via the Zoom (www.zoom.us) videoconferencing platform. Participants make this decision prior to their baseline assessments and the delivery mode remains consistent across their subsequent time points. Participants that choose to complete the assessments remotely receive a hard-copy remote cognitive assessment package in the post, including detailed instructions. Some of the assessments are enclosed in a separate envelope which is opened on request while in the testing session. Supplementary Table 1 details summarises mode of delivery of study procedures.
Randomisation and blinding
Randomisation and blinding protocols have been described in Johnson, Werden (25), and remain the same in the current protocol, except for added precautions for delivery of exercise equipment (see below). In brief, after Study Visit 1, randomisation occurs following MRI. The randomisation stratification includes total intracranial volume and baseline modified Rankin Scale (26). Once intervention arm is allocated, the home safety visit is arranged to allow delivery of and familiarisation with the exercise equipment.
Home Safety Assessment
A Trained Exercise Professional (TEP) performs a home visit and safety assessment for every participant, prior to commencing either arm of the exercise program. A comprehensive home safety assessment checklist was developed using clinical home-visit standards and protocols used by Austin Health clinicians (see acknowledgements) and used to guide each safety assessment. The participant’s level of function determines if they can complete the exercise protocol safely on their own, or if they will require a ‘study partner’ (e.g., family member, neighbour, friend) to be present. The TEP and the participant also establish a safe exercise space and an appropriate location for videoconferencing. All TEPs strictly abide by current Australian Department of Health and Human Services requirements and local hospital governance guidelines, including the use of Personal Protective Equipment as required.
Appropriate equipment delivery
All equipment required to complete the intervention is delivered either by a professional courier company (larger items e.g., bike) or the TEP (smaller items e.g., mats, weights) (see Table 3 for full equipment list). The TEP familiarises the participant with the training protocol, exercise equipment, and the technological aspects (i.e., iPad, data plan and Zoom). Equipment for the aerobic exercise arm (Steelflex PB10 Upright Bike or Steelflex CBSG Artiso Commercial Upright Bike) is stored offsite in a secure storage facility nearby our professional couriers. Once treatment arm is allocated, couriers are notified to collect and deliver large equipment (e.g., exercise bike) to the participant. This is to prevent unblinding of study staff members should they witness the arrival of a courier vehicle to our offices.
Safe-to-Exercise Measures and Monitoring
Safe-to-exercise measures including blood pressure (BP), resting heart rate, and oxygen saturation (SpO2) are taken before and after each exercise session. The participant is trained by the TEP during the home-safety visit to complete these measures on their own or with assistance (study partner) under the observation (via Zoom) of the TEP.
For safety purposes and to measure intervention fidelity, during each exercise session the participant will self-monitor and report their heart rate (HR) and the Borg Rating of Perceived Exertion (27) back to the TEP in real time. If a participant experiences a dose-limiting event (e.g., a participant does not pass safe-to-exercise measures or experiences an adverse event during a session), rules are in place to manage and adapt the exercise training within safety limits, to encourage continued participation.
Telerehabilitation Delivery
A TEP delivers and monitors the participant’s exercise sessions via Zoom three times a week for eight weeks. Each training session consists of up to 60 minutes of exercise, beginning and ending with a five-minute warm-up and cool down, unchanged from the previous protocol.
Cardiorespiratory Exercise Intervention and the Balance and Stretching Groups
The intervention and active control groups have been previously described (25). We emphasise that the key prescription parameters of the intervention (i.e., cardiorespiratory exercise intensity, duration, and frequency) remains unchanged except for the mode of delivery. Participants randomised to the intervention group complete up to 32 minutes of cardiorespiratory exercise on a stationary bike three times per week (not treadmill alternative in view of falls risk at home). This includes two moderate-to-high intensity interval sessions and one continuous steady-state session per week. Participants also complete twice-weekly 15-minute bouts of moderate intensity strength training during the cardiorespiratory interval training sessions. Strength training progression is based on participant tolerance, with the aim for participants to maintain an RPE of 12 to 15 (i.e., light to hard)
Participants in the active control group complete 30 minutes of light, static stretching, and 20 minutes of static and dynamic balance-based exercises three times per week. The protocol for this group was designed to match the attention and dose (i.e., session duration and frequency) the participants in the intervention group receive, however limits exercise intensity that could potentially lead to improvements in cardiorespiratory fitness. Supplementary Table 2 outlines an example of a weekly schedule and exercises for both intervention and active control groups.
Outcome Measures
Primary outcome is unchanged: between-group (intervention versus control) difference in hippocampal and total brain volume change between two- and four-months post-stroke (i.e., pre- and post-intervention).
Secondary outcome (changed)
Differing from the original protocol, the main cognitive outcome measure (secondary outcome) is global cognitive ability as measured by the Alzheimer’s Disease Assessment Scale-cognition sub-scale (ADAS-Cog) (28) at 12 months/Study Visit 3.
This change was made to capture global cognition with a validated scale that can also be delivered remotely. The ADAS-Cog measures several cognitive domains: memory, language, praxis, and orientation. It is administered at two- and 12-months post-stroke, in addition to neuropsychological testing (see Supplementary Table 1), which is delivered at each study visit. Our cognitive outcome previously was executive function as measured by Trail Making-B test performed in the neuropsychological testing battery at each study visit. This executive score is now an exploratory outcome.
Exploratory outcomes
The relationship between cardiorespiratory fitness, cognitive domains, physical activity, 24-hour ambulatory BP, recurrent stroke, pre-specified growth factors and APOE ε4 allele status continue to be investigated with the following additions:
-
Intervention fidelity for each session for cardiorespiratory participants is achieved (yes/no) if exercise intensity (mean % of heart rate reserve) and time (minutes+/-5%) meet the prescription. Attendance is met if participants attend ≥ 20 sessions of 24 scheduled.
-
Measures related to mood, sleep, cardiorespiratory fitness, cardiovascular health, and daily physical activity, are described in the original protocol (25) and listed in Supplementary Table 1. Additional to these measures are the following:
-
inflammatory markers: venous blood is drawn to analyse inflammatory biomarkers (IL-6, IL-1β, TNF-α, IL-8, IL-10, and IL-1ra) and neurodegenerative markers (neurofilament light chain, NfL)
-
dietary intake: participants complete a three-day food diary using a smartphone application (Research Food Diary, Xyris Software, Australia xyris.com.au). This is shared with the study nutritionist. Nutrient analysis is undertaken using an Australian food composition database, the FoodWorks Professional (v10) software.
-
gut microbiome: participants are provided with a take-home stool specimen collection kit, including storage tubes with a DNA preserving agent, and instructions. The sample must be from the first bowel movement of the day and not collected on a day that participants are also completing the food diary. The stool sample is stored in a thermo insulated bag in the freezer until it can be returned to the research team within 24 hours. Samples are stored in -80°C freezers until microbial sequencing takes place to examine the gut microbiome composition.
Specific pandemic-driven changes to the original protocol
Cardiorespiratory Fitness – removal of VO 2peak .
PISCES fitness testing via the graded exercise test conducted on a total body recumbent stepper (NuStep, T5XR, NuStep, Inc., Ann Arbor, MI) estimated peak volume of oxygen consumption (VO2peak) via a gas analysis system. To minimise the risk of aerosol transmission whilst COVID-19 lockdown restrictions were in place, VO2peak was predicted via an equation utilising heart rate (29). COVID-19 hospital infection control protocols were followed until mid-2022 when government COVID-19 restrictions were fully eased. Upon easing of restrictions, we were permitted to return to a breath-by-breath pneumotach gas analysis system (Jaeger® Oxycon Mobile) to measure VO2peak at one site only and remained prohibited from its use at our other sites. In view of this, we have removed the VO2peak as an outcome measure and replaced it with a prediction equation developed in older adults to estimate VO2peak using a total body recumbent stepper (30). We note that this equation will require optimisation for our participants. It was developed on community volunteers, potentially biasing the results toward fitter, more motivated participants. Submaximal effort due to health anxiety and inability to exercise arising from musculoskeletal issues will impact fitness assessments. We note that the energy cost of steady-state activity is greater in stroke survivors (31), which could also potentially affect our estimation.
Physical activity measure changed to Actiwatch
The technology used to measure daily physical activity has been changed to the Actiwatch Spectrum Plus (Koninklijke Philips, Amsterdam, Netherlands). This change was because they are relatively light weight, travel well via postage, and could be given with a return padded envelope to be mailed back by the participant. Additionally, the Physical Activity Scale for the Elderly (PASE) questionnaire (32) is utilised weekly throughout the eight-week intervention to monitor participants’ physical activity outside their scheduled sessions as well as phone-delivered 2-monthly between Study Visit 2 and 3. We included this to maintain participant interest in the study, given that study visits and in-person access to our TEPs had reduced with our new study design. It also allowed us to monitor self-report of their activity between the 4- and 12-month visits.
Sample Size Estimation and Statistical Analysis
The database for the 34 PISCES participants was locked at their final study visit (10th December 2020). Recruitment for PISCES-ZODIAC commenced in November 2020. We aim to pool data in our analyses as well reporting individual datasets.
Sample size
Sample size estimation and statistical analysis plan was reconsidered from the original PISCES protocol (25). Originally, a total sample size of 100 participants (50 per group) was calculated to yield 80% power to detect a difference in hippocampal volume between groups corresponding to a medium-to-large effect size (d = 0.6), assuming the standard settings of two-tailed significance and alpha = 0.05. In CANVAS (5), we found hippocampal volume loss of − 3% over four months after stroke, in contrast to the − 0.4% change in the stroke-free control group. This gave us a total 45 patients per group. We increased the sample size by 10% originally to account for attrition over the 12-month period and the possibility of non-evaluable scans. We estimated a potential increase in attrition rates to 20% as a result of the pandemia as there were rolling lockdowns including curfews and limitations of travel (5-kilometre radius), and potential COVID-19 infection affecting ability to participate in the intervention or attend assessments. We were reassured by multiple studies where remote exercise interventions were not inferior to in-person on measures of adherence, safety, feasibility, and health quality of life (20). Hence, our PISCES-ZODIAC sample size increased to 110 (45/group + 20) participants, still requiring a total of 90 completed participants overall.
Statistical analysis. The primary outcome will be analysed on an intention-to-treat basis. To account for variance between PISCES and PISCES-ZODIAC, we will use mixed-effects regression models with trial and trial-by-treatment interaction terms incorporated as random effects in all models. The difference in hippocampal volume from baseline to four months will be compared using a linear mixed-effect regression model with group as a factor (intervention versus control) and baseline total brain volume and functional status (mRS 0–1 versus mRS 2–3) as co-variates. Effect estimates for the pooled dataset as well as for individual components (PISCES + PISCES-ZODIAC) will be reported with 95% confidence intervals.
Separate mixed-effects regression models, with groups as a factor and baseline total brain volume and functional status as co-variates, will be used to analyse the difference in hippocampal volume, total brain volume and executive function from baseline to one year. A repeated measures random effect regression model will be used to determine the relationship between hippocampal and total brain volume and executive function. The tertiary and exploratory outcomes will be analysed according to standard statistical principles for comparison of change between groups and associations between variables. As one of these exploratory analyses, we anticipate investigating the magnitude of the difference in various outcomes of interest (e.g., brain volume) between any PISCES/PISCES-ZODIAC exercise intervention (active or control) and historical controls (no intervention, observational cohort) from the CANVAS study.
Study Organisation
All study procedures follow Good Clinical Practice (GCP) guidelines, and all investigators with participant-facing roles have valid GCP certification and are placed on our delegation log. A Data Safety Committee monitors trial performance, protocol violations, data quality and unblinded serious adverse events (SAEs) for trial safety bi-annually. SAEs are reported to our Data Safety Medical Monitor within 24 hours of notification. An Operations Committee and a Steering Committee consisting of members of the investigation team and other trial staff convene at regular intervals. Monthly investigator meetings are held to update study staff.