High-intensity interval training after stroke: A mixed-methods systematic review and meta-analysis of safety,feasibility and acceptability

Abstract

Objective

To synthesise evidence regarding the safety, feasibility and acceptability of high-intensity interval training in adults post-stroke.

Data sources

A systematic literature search using eight major databases from inception to January 2025.

Review methods

Studies of any design, involving adults post-stroke, reporting on safety, feasibility and/or acceptability of any type of high-intensity interval training, were eligible. Methodological quality was assessed using the Effective Public Health Practice Project or the Mixed Methods Appraisal Tool, as appropriate. Meta-analyses of adverse events and dropouts were undertaken on data from randomised controlled trials. GRADE methodology was used to categorise the level of certainty of the evidence from each meta-analysis.

Results

Twenty studies plus one follow-up and a cost analysis, involving 658 participants, were included. Methodological quality varied. High-intensity interval training was mainly conducted individually and in supervised settings. There were no fatalities related to high-intensity interval training. There was low–moderate certainty of no difference in the risk of adverse events, whether these were intervention-related, non-intervention related or unclear. Average attendance at high-intensity interval training sessions was 94.4%. Reporting on acceptability of high-intensity interval training was scarce, but where reported participants’ experiences were generally favourable.

Conclusion

This systematic review found low–moderate certainty evidence that high-intensity interval training can be safe and feasible for carefully selected, generally younger, mildly affected, male stroke survivors in the chronic stage, supervised by trained professionals in controlled settings. Future studies should investigate experiences of high-intensity interval training and explore its use in a wider stroke population and range of settings.

Keywords

High-intensity interval training stroke fitness training rehabilitation safety feasibility acceptability

Introduction

Current clinical guidelines for stroke in the UK and Ireland recommend moderate to high-intensity aerobic training and resistance training to improve outcomes.¹ For individuals with stroke-related motor impairments, exercise intensity should reach 40%–60% of heart rate reserve or at least 70% peak heart rate during mixed or high-intensity training, respectively.¹ A contemporary form of exercise that has shown significant motor recovery and cardiorespiratory benefits is high-intensity interval training.² This consists of alternating short intervals of vigorous intensity exercise with recovery periods of rest or low activity at light intensity.^3–5 Typically, sessions last from 10‒40 minutes and consist of exercise intervals of between 15 seconds and 5 minutes at high intensity (at least 70% of maximum heart rate), followed by slightly longer periods of recovery at lower intensities (40%‒50% of maximum heart rate).

High-intensity interval training has shown improvements in health- and skill-related fitness outcomes as well as quality of life after stroke.^6–8 A systematic review by Wiener et al.⁹ found improvements in various cardiorespiratory (e.g., peak oxygen uptake and walking economy) and mobility (e.g., Berg Balance Test and 6-Minute Walk Test) outcomes following high-intensity interval training after stroke. Other recent findings showed that high-intensity interval training enhanced neuroplasticity and cerebrovascular plasticity^10,11 and showed greater improvements than lower intensity in health/skill related fitness outcomes after stroke.^6,7,12

While previous research has mostly focused on the clinical effects of high-intensity interval training after stroke,^8,9,13 a search of the major databases as well as the PROSPERO database returned no evidence synthesis on the safety, feasibility or acceptability of high-intensity interval training after stroke. The evidence on the effectiveness of high-intensity interval training after stroke is essential to inform clinical practice; however, it is not sufficient for its implementation, as evidence for the safety, feasibility and acceptability of an intervention – especially in high-risk populations – is also essential.¹⁴ The purpose of this systematic review and meta-analysis was to synthesise evidence on the safety, feasibility and acceptability of high-intensity interval training after stroke.

Materials and methods

Study design

This mixed-methods systematic review and meta-analysis was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses¹⁵ and PRISMA for Searching.¹⁶ Its protocol was registered prospectively in PROSPERO (CRD42022386835).

Eligibility criteria

Quantitative, qualitative and/or mixed methods studies of any design were eligible. Review articles were excluded; however, their reference lists were searched to ensure all relevant studies were included in the final selection.

To be eligible, studies had to include adult participants with a clinical diagnosis of stroke, of any severity, and at any time since diagnosis. Where the study included a mixed population, at least 75% had to have the index diagnosis of stroke and data from ‘eligible participants’ had to be retrievable.¹⁷

High-intensity interval training interventions could include low/high volume and short high-intensity interval training, targeting a minimum interval intensity criterion of at least 60% of heart rate reserve or maximal oxygen uptake, at least 70% of maximum heart rate, or at least 14 on the Borg rating of perceived exertion [6–20 scale, suggesting a ‘somewhat hard’ level of intensity], with short bursts of concentrated effort alternating with planned recovery periods of low activity or rest.¹⁸ The use of performance variables (e.g., speed and cadence) to monitor intensity was only accepted if studies showed in their results section that participants exercised at appropriate high-intensity interval training intensities according to physiological variables (e.g., heart rate and oxygen uptake).⁵ Additionally, the mode of exercise could comprise equipment such as treadmills, cycle ergometers and recumbent steppers, which allowed participants to use mobility aids including walkers, sticks and splints, as required. Studies were excluded if: the effects of high-intensity interval training could not be isolated for data synthesis; only one single session of high-intensity interval training was reported; supramaximal (e.g., sprint interval training), which comprises all-out, supramaximal efforts equal to, or greater than, a pace that results in ≥100% peak oxygen uptake,¹⁹ was implemented as part of the intervention; and the recovery component of high-intensity interval training was not clearly pre-determined.

Studies were not required to have a comparator or control group (e.g., cohort studies); however, comparators could include no treatment, usual care or exercise programmes including low-to-moderate intensity training, education sessions, social support or sham interventions – provided the impact of high-intensity interval training could be identified.

The main outcomes for this review were safety, feasibility and acceptability of high-intensity interval training.

Safety during the intervention period was operationalised as follows:

case fatality

other adverse events: number, cause (categorised according to descriptions extracted from each of the included studies: (1) ‘intervention-related’, i.e., adverse events, reported during the intervention period, described by the original study authors as being related to the intervention (e.g., syncope during exercise); (2) ‘non-intervention-related’, i.e., those adverse events that, according to the original study authors, were not related to the intervention, (e.g., a non-injurious fall in the follow-up period); and (3) ‘unclear’, in cases where study authors did not describe an adverse event as either intervention-related or not), severity (grade 1: mild, to grade 5: death related to an adverse event, as per the Common Terminology Criteria for Adverse Events²⁰), type (intervention-related or non-intervention-related adverse events or unclear) and reasons

Feasibility was operationalised as follows:

participant recruitment and retention:

number of participants assessed for eligibility and those randomised (or allocated otherwise) to the intervention

number of dropouts during the intervention period (i.e., ‘intervention-related dropouts’) and reasons

intervention fidelity:

attendance – whether the participant was present or not for each exercise session (reported objectively or subjectively)

adherence to: ▪

the planned content of the exercise protocol

▪

the appropriate/planned exercise intensity as prescribed for high-intensity interval training¹⁸

cost of equipment, space, personnel time and personnel training

Acceptability was operationalised as follows:

perception of access to/cost of equipment, space and personnel time (e.g., ease of availability and equipment malfunction)

acceptability of the intervention reported by study participants and/or healthcare professionals delivering high-intensity interval training

Information sources and search strategy

The search was conducted using the following electronic databases: MEDLINE, AMED, and CINAHL via EBSCOhost, PEDro, Cochrane Library, ProQuest, Scopus and ScienceDirect from inception to January 2025. Search strategies are provided in Supplementary Material A. The reference lists of all included studies and reviews were searched for additional studies. Only articles published in English were included. Aligned with the research question's PICO (Population, Intervention, Comparison, and Outcomes) elements, variations of the keywords related to ‘stroke’ and ‘high-intensity interval training’ were used for each database search strategy and a combination of controlled medical subject headings and free text terms was adapted for each database (Supplementary Material A).

Study selection process

After removal of duplicates, studies were screened for eligibility based on title and abstract, according to predetermined criteria listed above. One reviewer (HB) retained those that were definitely or possibly relevant, and a second independent reviewer (FvW) cross-checked a random 10% of the titles and abstracts. Two independent reviewers (HB and FvW) were involved in the full-text selection. Any discrepancies that arose between the reviewers at any stage were resolved through discussion with a third independent reviewer (LP) where required. A PRISMA flow diagram¹⁵ was used to trace the overall review process using Covidence software.²¹

Data collection process and data items

The data extracted included: study characteristics (i.e., author[s], year of publication, country of origin, study design, method of recruitment, inclusion and exclusion criteria, sample size and study setting) and participant characteristics (i.e., total number, age, sex, stroke type, lesion site, stroke severity, time post stroke and ambulatory status). Data were extracted from intervention protocols using the Consensus on Exercise Reporting Template²² comprising seven categories: materials, provider, delivery (including intervention frequency, intensity, type and timing), location, dosage, tailoring and compliance. Additionally, data on safety, feasibility and acceptability outcomes were extracted. Where reported, the Functional Ambulation Category by Holden et al.²³ was used for collecting data regarding participants’ ambulatory status. Dropouts and adverse events during the training period and at follow-up were extracted.

Study quality assessment

Quantitative studies were assessed using the Effective Public Health Practice Project tool,²⁴ which is designed for randomised and non-randomised studies. This tool comprises six components and provides an overall quality rating of ‘weak’, ‘moderate’ or ‘strong’. Mixed-methods studies were evaluated using the Mixed Methods Appraisal Tool.²⁵ This tool uses a set of criteria and screening questions to provide an overall quality rating from one to five stars. Each study was assessed by two independent reviewers (HB + FvW), and the findings were discussed subsequently. If necessary, any disagreements over study quality were resolved through consultation with a third reviewer (LP).

Synthesis methods

A convergent mixed-methods synthesis design was planned, where quantitative and qualitative findings were synthesised in a parallel or complementary manner.²⁶ For the analysis of safety outcomes, numerical data were analysed from all included studies on the cause, severity (scored according to the Common Terminology Criteria for Adverse Event²⁰), type and reasons. To compare the intervention and control groups, the number of case fatalities and participants with adverse events (i.e., intervention-related, non-intervention-related and unclear) were extracted from RCTs only, to calculate risk ratios (with 95% confidence intervals) using random-effects meta-analyses using RevMan Web, according to Cochrane guidance. This included RCTs only but excluding those with non-reported or zero events in both study arms.^27,28 For the analysis of feasibility outcomes, all numerical data were analysed and tabulated, and a narrative summary was presented in text and tabular form. To enable comparison between the intervention and control groups, the number of dropouts at the end of the intervention period was used to calculate risk ratios (with 95% confidence intervals) for random-effects meta-analyses, as described for adverse events. For the analysis of acceptability outcomes, content analysis was planned to analyse relevant themes, concepts and thematic patterns, as appropriate.²⁹ If sufficient data were available, subgroup analyses were planned to explore any heterogeneity related to different types of stroke, time post stroke and ambulatory status. GRADE methodology was used to categorise the level of certainty of the evidence from each meta-analysis as ‘high’, ‘moderate’, ‘low’ or ‘very low’, as per GRADE guidance.³⁰

Results

Study selection

A PRISMA flow diagram shows the process of record screening during this systematic review. The electronic search of eight databases retrieved 4833 potential records. Four records were excluded due to duplication of study data.^31–34 Twenty-two records were included in the final review (Figure 1).

Study characteristics

Of the 20 studies (22 records) included, 18 were quantitative and two studies used a mixed-methods design. There were 12 RCTs,^12,35–45 including one follow-up report⁴² presenting 6- and 12-months post intervention data from their 2019 study,³⁹ and a cost analysis report⁴⁵ of the HIT stroke trial by Boyne et al.³⁷ There were no qualitative studies. There were insufficient data for undertaking the planned subgroup analyses.

Quality assessment

Of the 18 quantitative studies, nine (50%) were classified as ‘strong’, five (28%) as ‘moderate’ and four (22%) as ‘weak’ (Table 1). The two mixed-methods studies^35,36 rated poorly on the qualitative and mixed-methods components but met four or more criteria on the quantitative component (Table 2).

Table 1.

Methodological quality assessment of the included quantitative studies according to the Effective Public Health Practice Project tool²⁴ (n = 18 studies).

Study	Selection bias	Study design	Confounders	Blinding	Data collection method	Withdrawals and dropouts	Global rating
Amanzonwé (2024)⁴⁶	M	M	N/A	M	S	S	S
Askim (2014)⁴⁷	M	M	N/A	M	S	S	S
Bosch (2021)⁴⁸	M	S	S	M	S	S	S
Boyne (2015)⁴⁹	W	M	N/A	M	S	S	M
Boyne (2019)⁵⁰	W	M	N/A	M	S	S	M
Boyne (2020)⁵¹	W	M	N/A	M	S	S	M
Boyne (2022)⁵²	W	M	S	M	S	S	M
Boyne (2023)³⁷	S	S	S	M	S	S	S
Calmels (2011)⁵³	M	M	N/A	W	S	S	M
Do (2024)⁴⁴	M	S	S	M	S	S	S
Ekechukwu (2017)³⁸	W	S	W	M	W	W	W
Gjellesvik (2012)⁵⁴	M	M	N/A	M	S	S	S
Gjellesvik (2020)⁴⁰	M	S	S	M	S	S	S
Hsu (2021)¹²	W	S	S	S	W	S	W
Lazarieva (2024)⁵⁵	W	M	W	M	W	W	W
Marzolini (2023)⁴¹	S	S	S	M	S	S	S
Moncion (2024)⁴³	M	S	S	W	S	W	W
Steen Krawcyk (2019)³⁹	S	S	S	M	S	S	S

W: weak; M: moderate; S: strong; N/A: not applicable to studies with only one group.

Table 2.

Quality assessment of mixed-methods studies according to the Mixed-Methods Appraisal Tool²⁵ (n = 2 studies).

References	Screening		Qualitative component					Quantitative component (related to RCTs)					Quantitative component (related to non-RCTs)	Quantitative descriptive component	Mixed-methods design					Total qualitative rating	Total quantitative rating	Total mixed methods rating
References	S1	S2	Q1.1	Q1.2	Q1.3	Q1.4	Q1.5	Q2.1	Q2.2	Q2.3	Q2.4	Q2.5	Quantitative component (related to non-RCTs)	Quantitative descriptive component	Q5.1	Q5.2	Q5.3	Q5.4	Q5.5	Total qualitative rating	Total quantitative rating	Total mixed methods rating
Boyne (2016)³⁵	Y	Y	N	?	?	N	N	Y	Y	Y	Y	Y	N/A	N/A	Y	N	N	?	N	0	*****	*
Valkenborghs (2019)³⁶	Y	Y	Y	Y	?	N	?	Y	N	Y	Y	Y	N/A	N/A	Y	Y	?	?	N	**	****	**

?: can’t tell; Y: yes; N: no, N/A = not applicable.

Scores range from 0 (no criteria met), to * (one criterion met) to ***** (all criteria met).

S1. Are there clear research questions? S2. Do the collected data allow to address the research questions? Q1.1. Is the qualitative approach appropriate to answer the research question? Q1.2. Are the qualitative data collection methods adequate to address the research question? Q1.3. Are the findings adequately derived from the data? Q1.4. Is the interpretation of results sufficiently substantiated by data? Q1.5. Is there coherence between qualitative data sources, collection, analysis and interpretation? Q2.1. Is randomisation appropriately performed? Q2.2. Are the groups comparable at baseline? Q2.3. Are there complete outcome data? Q2.4. Are outcome assessors blinded to the intervention provided? Q2.5. Did the participants adhere to the assigned intervention? Q5.1. Is there an adequate rationale for using a mixed methods design to address the research question? Q5.2. Are the different components of the study effectively integrated to answer the research question? Q5.3. Are the outputs of the integration of qualitative and quantitative components adequately interpreted? Q5.4. Are divergences and inconsistencies between quantitative and qualitative results adequately addressed? Q5.5. Do the different components of the study adhere to the quality criteria of each tradition of the methods involved?

Participant characteristics

All participant characteristics have been fully documented in Supplementary Material B. There was a total of 658 participants with stroke in the 20 included studies, with sample sizes ranging from 8⁵⁴ to 71.³⁹ The study by Boyne et al.⁵² included a control group without stroke, but these data were not included in this review. The mean age of participants ranged from 49 ± 10 years⁵⁴ to 70 ± 8 years.⁴⁷ Fifteen studies reported sex with overall 66% male and 34% female participants included.

Mean time since stroke ranged from 16 ± 7 days⁴⁶ to 102 ± 108 months.³⁶ Only Amanzonwé et al.⁴⁶ involved participants in the early subacute phase of stroke (1 week to 3 months). Askim et al.⁴⁷ involved participants in the late subacute stage of stroke (3–6 months), and the rest included participants in the chronic stage (>6 months) with six studies including participants more than 5-years post stroke.^{35,36,44,49,50,54} Fifteen studies reported the type of stroke and lesion site; 78% were ischaemic, 22% were haemorrhagic and 1% were unknown; 48% had right sided lesions, 48% left sided lesions and 4% had both sides affected.

Stroke severity was indicated in only six studies; participants were categorised as having mild or no significant disability despite symptoms.^{12,39,40,43,46,47} Participants’ ambulatory status or use of gait aids was reported in 12 studies (356 participants).^{35,37,39–41,46–49,52–54} Among these, 149 (42%) used assistive devices (e.g., walking aids and orthoses). Five studies^{35,37,44,49,52} reported the Functional Ambulation Category⁵⁶ of their 126 participants, of which 23 (18%) were fully ambulatory. Only one study³⁶ included non-ambulatory participants.

Intervention protocols

All Consensus on Exercise Reporting Template items have been fully reported in Supplementary Material B. The type of exercise equipment used in the intervention protocols comprised treadmills (with safety harness), bicycle ergometers, seated steppers, a rehabilitation robot and other equipment to provide aerobic exercise. The interventions were mostly provided by physical therapists and ‘study coordinators’. All study interventions were performed individually and within healthcare or university settings and all except one were supervised. The study by Steen Krawcyk et al.³⁹ was the only one to include unsupervised high-intensity interval training interventions at home. The measurement and reporting of adherence to exercise intensity was collected via heart rate monitors in 13 studies. The rest used a mix of rating of perceived exertion, treadmill speed and oxygen uptake to record intensity. Motivational strategies were reported in five studies only^{39,46,48,50,54} and mostly consisted of verbal encouragement.

Intervention parameters varied considerably (Table 3). All exercise protocols were tailored to the individual in terms of intensity and/or modality. The decision rules for the exercise starting level (i.e., speed, resistance and workload) consisted of individually calculated intensity training zones (e.g., peak heart rate, heart rate reserve and ventilatory threshold heart rate), exercise tests (e.g., steep ramp test and incremental cycle ergometer test) or starting rapid treadmill accelerations.

Table 3.

Overview of high-intensity interval training intervention parameters from the included studies.

Intervention parameters	Duration and time (range)	Intensity (range) for different variables
Exercise frequency	1–5 days per week 3–24 weeks 3–120 sessions in total	N/A
Warm-up	3–15 minutes	Light to moderate intensity 30–50% peak oxygen uptake/heart rate reserve Self-selected walking speed
Main training phase (high-intensity interval training)	2–10 intervals ranging from 30 seconds to 5 minutes	70–85% of maximum heart rate 85–95% peak heart rate 80% peak oxygen uptake 40–80% heart rate reserve 77–93% age-predicted maximal heart rate 110% to 120% of ventilatory threshold heart rate 14–16 rating of perceived exertion 70–80% maximum workload or performance-based criterion Maximum possible cadence Maximum tolerated/safe speed ‘Not able to speak comfortably’ on the Talk Test
Recovery periods	30 seconds to 4 minutes of walking/cycling	30–70% peak heart rate 40% of peak oxygen uptake 64–70% of age-predicted maximal heart rate 20–30% heart rate reserve 30–40% maximum workload 10–12 rating of perceived exertion Exercise device was stopped to allow participants to sit, stand still or march in place
Cool-down	2–10 min	Light to moderate intensity 30–70% peak heart rate 30–50% peak oxygen uptake

Twelve out of 20 studies reported a control group.^{12,35–41,43,44,48,55} Three consisted of usual/standard care.^39,40,48 Six studies included continuous aerobic exercise at moderate intensity with various equipment^{12,35,37,38,41,43} and the remaining studies included a repetitive task-specific upper limb training program,³⁶ active physiotherapy⁵⁵ and home exercise education.⁴⁴

Findings

The following sections present findings on safety, feasibility and acceptability. Judgements about the certainty of the evidence emerging from the GRADE analysis of each of the meta-analyses are presented, together with explanatory notes in Supplementary Material C.

Safety

Overall, only one death was reported in the high-intensity intervention groups (377 participants randomised), but this was unrelated to the study⁴⁰ (Table 4). There were no documented deaths in the control groups, and therefore, a meta-analysis could not be undertaken to compare case fatality between groups.²⁷ Nineteen studies reported adverse events and two did not^38,55 (Table 4). A total of 233 adverse events were reported in 21% of participants randomised across all intervention groups, of which 77 adverse events (33%) were intervention-related and mostly of the lowest severity (57 at Grade 1, 19 at Grade 2 and 1 at Grade 3). A total of 107 adverse events (46%) were reported as unrelated to the intervention and were also mostly of the lowest severity, while the cause of 49 adverse events (21%) was unclear. The most common exercise-related adverse events were general pain/soreness (35%), light-headedness (24%), fatigue (12%) and joint/muscle pain (10%).

Table 4.

Summary overview of the demographic data, intervention parameters and safety, feasibility and acceptability outcomes in the included records. Full details can be found in Supplementary materials B and D.

Author (year); study designNumber of study participants per group and sex (M/F)#; age (mean ± SD unless stated otherwise); time since stroke(mean ± SD unless stated otherwise)	FITT (Frequency, Intensity, Time, Type)intervention parameters	Number of dropouts (N) during the intervention and follow-up period (where included)^b	Case fatality (number of deaths recorded during the study period)	Number of participants with AEs (n) and total number of AEs (N) during the study period (grade severity according to the CTCAE^a, if reported)
				Total^c	Related to the intervention
Amanzonwé (2024)⁴⁶; Cohort study INT 10 (6/4); Median (IQR): 63.5 (56.7–71.2); 16 ± 7 days	INT F: 3 days per week, 6-week regimen I & T: Each session included 3 minutes warm up of unloaded cycling and 3 minutes cool-down stretching. HIIT: 1 minute at 70% peak workload interspersed with 4 minutes at 30% for weeks 1 and 2 (4 repetitions of 5 minutes each) at 50 rpm; as tolerated, workload increased by approximately 5 minutes every 2 weeks to 30 minutes from week 5 (six repetitions). T: recumbent bike	INT N = 0	INT N = 0	INT n = N/R N = 2	INT N = 0
Askim (2014)⁴⁷; Cohort study INT 14 (9/5); 70.0 ± 7.7; 5.8 ± 1.7 months	INT F: 2 days per week, 6-week regimen; I & T: 10–15 minutes walking at a moderate intensity (warm-up), 25 minutes high-intensity AIT: 4 × 4-minute intervals at 85–95% of HR peak, 3-minute active breaks at 70% of HR peak, 5–10 minutes cool-down walking at light to moderate intensity. T: BWST	INT N = 1 at 6 weeks follow-up 1 diagnosed with cancer	INT N = 0	INT n = 2 N = 2 (called ‘happenings’ during the study intervention)	INT n = 2 N = 2; 2(1) 1 discontinued the last session, feeling unwell (normal blood pressure and HR; recovered but chose to stop), 1felt unwell after training.
Bosch (2021)⁴⁸; Quasi experimental study INT 10 (7/3); 59.5 ± 16.4; 24.5 ± 14.6 months CON 7 (4/3); 55.4 ± 11.9; 29.8 ± 5.6 months	INT F: 3 days per week, 10-week regimen including 2 HIIT sessions and 1 independent, moderate intensity overground walking session at VT HR for 30 minutes/week on a separate day. I & T: HIIT: 5-minute warm-up at a self-selected walking speed on a level treadmill, 4 × 4-minute high-intensity (fast) intervals targeted at ≤90–95% AP-HR_max interspersed with 3 × 3-minute recovery bouts at VT HR, equivalent to 64% of AP-HR_max, 5-minute cool-down. Home walking: 30-minutes of continuous walking overground at VT HR, if not feasible: 3 × 10-minute bouts to achieve 30 minutes of walking in the same day. T: BWST CON Usual routine for the 10-week interim: 2 participants received physical therapy and 1 reported using a gym.	INT N = 0 CON N = 2 at follow-up 1 unrelated medical procedure precluded exercise, 1 schedule conflict precluded return	INT N = 0 CON N = 0	INT N = 0 CON n = 1 N = 1	INT N = 0 CON N = 0
Boyne (2015)⁴⁹; within-subject repeated measures design INT 18 (10/8); 61.9 ± 8.3; 5.8 ± 4.2 years	INT F: single sessions of three different treadmill HIT protocols with variable recovery periods (30 seconds (P30), 60 seconds (P60), or 120 seconds (P120)) with a 1-week washout period between sessions. I & T: HIT: 5-minute warm-up at 30%–50% VO₂ peak, up to 20 minutes of HIT with repeated 30-second bursts at maximum tolerated speed (0% incline), alternated with pre-specified recovery periods, where the treadmill was stopped (option to stand still, march in place or sit), 5-minute cool-down at 30%–50% VO₂ peak T: treadmill with safety harness (not BWST)	INT N = 1 at follow-up N/S	INT N = 0	INT n = 1 N = 1	INT n = 1 N = 1; 1(1) 1 reported mild transient light-headedness during 1 HIT session
Boyne (2016)³⁵; Mixed methods RCT INT 11 (7/4); 59 ± 9; 3.8 ± 2.9 years CON 5 (2/3); 57 ± 12; 6.3 ± 2.0 years	INT F: 3 days per week, 4-week regimen I & T: For both HIT and MICT: 3-minute warm-up at 30%–50% HRR, 20 minutes of training, and 2-minute cool-down at 30%–50% HRR. HIT: 30-second bursts at maximum safe speed, alternated with 30–60-second recovery periods, where the treadmill was stopped (recovery periods lasted 60 seconds during the first 3 sessions; for the remaining nine sessions, recovery duration began at 60 seconds and changed to 30 seconds after the first 3 bursts. T: treadmill with safety harness (not BWST) CON MICT: Moderate intensity continuous treadmill walking with speed adjusted to 45% ± 5% HRR – target HR was progressed to 50% ± 5% HRR after 2 weeks of training.	INT N = 2 during the intervention. 1 completed 3 sessions and then withdrew due to osteoarthritis-related knee pain, 1 completed 6 sessions and then was withdrawn after colliding with a cabinet at home and injuring nonparetic leg CON N = 0	INT N = 0 CON N = 0	INT n = 9 N = 21 CON n = 4 N = 8	INT n = 6 N = 13; 10(1), 3(2) 8 joint/muscle pain, 3 fatigue, 2 light-headedness CON n = 1 N = 4; 3(1), 1(2) 4 joint/muscle pain
Boyne (2019)⁵⁰; within-participant crossover INT 16 (9/7); 57.4 ± 9; 6.5 ± 4.1 years	INT F: single sessions of three different exercise protocols 1 week apart I & T: Each protocol included a 3-minute warm up at 30–50% HRR, 20 minutes of exercise and a 2-minute cool down. Treadmill HIIT: repeated 30-second bursts at maximum tolerated speed (0% incline), alternated with recovery periods where the treadmill was stopped initial burst speed determined by a rapid acceleration test where speed was increased by 0.1 mph every 5 seconds until reaching the limit where the participant drifted backward or exhibited gait instability recovery duration was decreased from 60 to 30 seconds after the first 5 minutes, duration was extended if fastest safe burst speed decreased below the initial speed, if the participant requested to sit and was not able to stand back up in time or if the participant was too short of breath to talk when a burst was scheduled to start Seated stepper HIIT: same burst and recovery durations as HIT-treadmill but on a seated stepper bursts were performed at maximum possible cadence against 50% of maximal resistance if end-burst cadence was <80 or >200 steps/minute, resistance was decreased or increased for the next burst, respectively recovery duration was extended if the participant was too short of breath to talk when a burst was scheduled to start Treadmill MICT: continuous treadmill walking at 45 ± 5% HRR T: treadmill with safety harness (not BWST) or seated stepper	INT N = 0	INT N = 0	INT n = N/R N = 7	INT HITS n = 2 N = 2; 1(2), 1(3) 2 exhibited symptomatic hypotension and near syncope during recovery, which led to temporary hospitalisation in n = 1
Boyne (2020)⁵¹; within-participant repeated measures design INT 10 (6/4); 59.8 ± 6.8; 2.4 ± 1.7 years	INT F: 3 days per week, 4-week regimen, alternating between short and long-interval HIIT sessions I & T: Each session included a 3-minute warm up (overground walking at ∼40% HRR) and 10 minutes of overground HIIT, and then 20 minutes of treadmill HIIT, followed by another 10 minutes of overground HIIT and a 2-minute cool down (overground walking at ∼40% HRR). SIT: 30-second bursts at maximum safe speed alternated with 30–60 second resting recovery periods each overground, and treadmill bout started with 60-second recovery for the first 3 bursts, and then progressed to 30-second recovery for the rest of the bout (if burst speed substantially decreased or the participant requested a seated rest break, recovery duration was temporarily increased again). LIT: On treadmill: 3 × 4-minute bursts with recovery durations of 3 minutes, 3 minutes and 2 minutes. Overground: 2 × 3-minute bursts with a 2-minute recovery period. Target HR was ∼90% HR peak for bursts and ∼70% HR peak for recovery periods – the goal was to reach the target HR zone within 1–2 minutes each burst recovery periods started with rest breaks until the HR trajectory was determined. T: treadmill (except for overground condition)	INT N = 0	INT N = 0	INT N = 0 serious AE	INT N = 0 serious AE
Boyne (2022)⁵²; cohort of stroke participants and a cohort of healthy controls INT 10 (6/4); 59.8 ± 6.8; 2.4 ± 1.7 years	INT As in Boyne (2020)⁵¹	INT N = 0	INT N = 0	INT n = 8 N = 55	INT n = 5 N = 18; 15(1), 3(2) 10 soreness/pain, 3 fatigue and 5 light-headedness
Boyne (2023)³⁷; RCT INT 27 (16/11); 63.8 ± 9.9; 32.4 ± 12 months CON 28 (20/8); 61.5 ± 9.9; 24 ± 12 months	INT F: 3 days, 12-week regimen I & T: 3-min warm-up of overground walking at 30%–40% HRR, 10 minutes of overground HIIT or MAT, 20 minutes of treadmill HIIT or MAT, another 10 minutes of overground HIIT or MAT and 2-minute cool down at 30%–40% HRR. HIIT protocol: repeated 30-second bursts walking at maximum safe speed, alternated with 30–60 seconds passive recovery periods (standing or seated rest as tolerated), targeting a mean aerobic intensity above 60% HRR T: treadmill with safety harness (not BWST) + overground walking CON MAT: same training protocol as HIIT but continuous walking practice involved speed adjusted to maintain an initial target HR of 40% ± 5% HRR, progressing by 5% HRR every 2 weeks up to 60% of the HRR as tolerated.	INT N = 8 discontinued intervention. 4 self-withdrew: 3 due to AEs likely unrelated to treatment and 1 due to treatment-related AE; 4 withdrawn by the study team: 2 due to AEs likely unrelated to treatment, 1due to COVID-19 shutdown and 1 due to treatment-related AE CON N = 5 discontinued intervention. 1 self-withdrew due to AEs likely unrelated to treatment; 4withdrawn by the study team: 1due to AEs likely unrelated to treatment and 3due to COVID-19 shutdown	INT N = 0 CON N = 0	INT n = 27 N = 109 CON n = 23 N = 84	INT n = 13 N = 39; 26(1), 13(2) 15 pain/soreness, 2 fatigue, 10 light-headedness, 4 fall CON n = 10 N = 29; 24(1), 5(2) 18 pain/soreness, 2 fatigue, 8 light-headedness
Calmels (2011)⁵³; cohort study INT 16 (12/2)#; 53.7 ± 8.6; 12.1 ± 7.52 months	INT F: 3 days per week, 8-week regimen I & T: Cycling performed at 60 rpm: 6 sessions × 4 minutes at 40% of the maximal workload and 1 minute at 80% (30 minutes total). T: adapted cycle ergometer (with largest pedals and straps to fix the hemiparetic foot)	INT N = 2 1 fracture following a fall at home (unrelated to the training program) and 1 due to development of a comorbidity requiring surgical intervention	INT N = 0	INT n = 2 N = 2	INT N = 0
Do (2024)⁴⁴; RCT INT 12 (7/4)#; 61.8 ± 7.3; 88.0 ± 81.5 months CON 12 (8/3)#; 63.5 ± 8.1; 81.5 ± 52.4 months	INT F: 3 days per week, 8-week regimen I & T: Each session included 5-minute stretching exercises for warm-up and cool-down. HIIT: four sets of 3-minute bursts of RAGT on the Morning Walk at RPE >14, alternating with 2-minute rest periods at RPE 10–11. T: Morning Walk^®, an end-effector-type lower limb rehabilitation robot CON Home exercise programme (aerobic, balance, stretching and strengthening exercises) + Outpatient physical therapy (twice per week, content N/R).	INT N/R CON N/R	INT N = 0 CON N = 0	INT n = 0 N = 0 CON n = 0 N = 0	INT N = 0 CON N = 0
Ekechukwu (2017)³⁸; RCT INT + CON 1 + CON 2 69 (N/R); 59.33 ± 8.80; N/R INT (IAT) 25 (N/R); N/R; N/R CON 1 (CT) 21 (N/R); N/R; N/R CON 2 (CBAT) 23 (N/R); N/R; N/R	INT F: 3-day per week, 8-week regimen; 20 minutes/day during the first 2 weeks, increased by 5 minutes every 2 weeks until the 8th week I & T: All protocols included up a 5-minute warm-up, a main training phase, and a 5-minute cool down. IAT at 40–70% HRR (moderate/vigorous intensity) and rest/recovery at 20–29% HRR (low intensity) with a work-recovery ratio of 1 : 2 minutes. T: stationary cycle ergometer CON 1 CT at 40–70% HRR CON 2 CBAT: 1/2 IAT + 1/2 CT at 40–70% HRR.	INT N/R CON 1 + CON 2 N/R	INT N/R CON 1 + CON 2 N/R	INT N/R CON 1 + CON 2 N/R	INT N/A CON 1 + CON 2 N/A
Gjellesvik (2012)⁵⁴; cohort study INT 8 (4/4); 48.9 ± 10.6; 7.2 ± 7.5 years	INT F: 5 days per week, 4-week regimen I & T: 10-minute warm-up on treadmill at self-selected speed and inclination, 4 × 4 minute intervals at 85–95% HR peak of uphill treadmill walking, between intervals 3-minute active walking breaks at 50% HR peak, 3-minute cool-down at 50–70% HR peak. T: treadmill	INT N = 0	INT N = 0	INT N = 0	INT N = 0
Gjellesvik (2020)⁴⁰; RCT INT 36 (21/15); 57.6 ± 9.2; 25.4 ± 14.5 months CON 34 (20/14); 58.7 ± 9.2; 27.4 ± 14.7 months	INT F: 3 days per week, 8-week regimen I & T: 10-minute warm-up, 4 × 4 minute intervals at 85%–95% of HR peak, 3 minutes of active walking breaks at 50%–70% of HR peak. T: treadmill CON Standard care (the Norwegian guidelines recommend patients with chronic stroke be physically active and engage in activities with moderate to high intensity 3–5 days per week)	INT N = 3 at post-test 2 withdrew, 1 underwent surgery N = 5 at 12 months follow-up 1 died, 1 withdrew, 1 not available, 1 admitted to hospital, 1 low back pain CON N = 3 at post-test 2 withdrew, 1 did not tolerate face mask N = 3 at 12 months follow-up 2 withdrew, 1 not available	INT N = 1 CON N = 0	INT n = 4 (at follow-up) N = 4 CON n = 4 (at follow-up) N = 4	INT N = 0 CON N = 0
Hazen (2024)⁴⁵; cost analysis of RCT	See Boyne (2023)³⁷	See Boyne (2023)³⁷	See Boyne (2023)³⁷	See Boyne (2023)³⁷	See Boyne (2023)³⁷
Hsu (2021)¹²; RCT INT 10 (8/2); 58.5; 38.5 months CON 13 (12/1); 53.1; 28.8 months	INT F: 2–3 sessions/week, total 36 sessions I & T: 3-minute warm-up at 30% of VO₂ peak, 5 × 3-minute intervals at 80% VO₂ peak, 3-minute recovery at 40% of VO₂ peak, 3-minute cool-down at 30% VO₂ peak T: bicycle ergometer CON MICT: 2 to 3 sessions/week: warm-up at 30% of VO₂ peak for 3 minutes, 60% VO₂ peak for 30 minutes, cool-down at 30% VO₂ peak for 3 minutes	INT N = 3 at follow-up 1 hernia surgery, 1 recurrent stroke, 1 unstable BP CON N = 2 at follow-up 1 recurrent stroke, 1 incomplete ex.	INT N = 0 CON N = 0	INT n = 3 N = 3 CON n = 2 N = 2	INT N = 0 CON N = 0
Lazarieva (2024)⁵⁵; cohort analytic study INT 27 (27/0); ‘mature age’ (37–50 years); from 2nd week post stroke CON 26 (26/0); ‘mature age’; from 2nd week post stroke	INT F: 2 sessions of 50–60 minutes per day, 5-day per week, 8-week regimen I & T: Warm up and cool down N/R. HIIT program for a 5-day week: 5 × 1 min at 70–85% of HRmax with 5-minute rest between intervals (HR N/R), increasing interval time each day by 1 minute up to 5 minutes on day 5. T: N/R CON Active physiotherapy: 2 sessions of 50–60 minutes per day, 5-day per week, 8-week regimen: walking, functional skills, bed mobility, lower limb strength training.	INT N/R CON N/R	INT N/R CON N/R	INT N/R CON N/R	INT N/Av CON N/A
Marzolini (2023)⁴¹; RCT INT 24 (20/4); 62.8 ± 13.2; 10.8 ± 7.3 months CON 23 (18/5); 60.9 ± 8.4; 10.6 ± 9.5 months	INT F: 5 days per week, 24-week regimen. Week 1–4: both HIIT and MICT groups participated in overground MICT (5x/week). Weeks 5–8: HIIT workout #1 3x/week + MICT 2x/week. Weeks 8–20: workout #1 was reduced to 1x/week, and workout #2 was carried out 2x/week I & T: HIIT workout #1: 5-minute warm-up, 20–22 minutes of 30-second intervals of maximal safe intensity alternated with 60 seconds of active recovery (∼30%–40% HR peak attained on the CPET or RPE of 10–12 [‘light’]), 5-min cool-down. Starting with 3 HIIT intervals, an increasing number of intervals was prescribed so that after 4 weeks, all 15 HIIT intervals were completed. HIIT workout # 2: 5-minute warm-up, 20–22 minutes of 2-minute intervals of maximal safe intensity alternated with 3 minutes of active recovery, 5-minute cool-down. Supervised exercise overground: 1–2 times/month. T: treadmill with safety harness (not BWST) CON MICT overground target: 60 minutes of exercise at the treadmill workload or HR corresponding to the V˙O₂ that occurred at the ventilatory anaerobic threshold. The initial walking prescription was set at a distance of up to 1.6 km depending on exercise tolerance, with a 3-minute warm-up and cool-down at 1.2 minutes/km slower than the prescription pace. Prescriptions progressed every two weeks.	INT N = 1 discontinued intervention 1 gall bladder surgery CON N = 1 discontinued intervention 1 bowel bleed unrelated to the intervention	INT N = 0 CON N = 0	INT n = 10 N = 16 CON n = 13 N = 17	INT n = 2 N = 2; 2(1) CON n = 5 N = 6; 6(1)
Moncion (2024)⁴³; RCT INT 42 (27/15); 65.4 ± 8.9; 1.9 ± 1.3 years CON 40 (23/17); 64.4 ± 9.7; 1.7 ± 1.3 years	INT F: 3 days per week, 12-week regimen I & T: Each session included a 3-minute warm-up and 2-minute cool-down (HR N/R). Short-interval HIIT: 10 × 1 minute intervals of high-intensity exercise at 80% HRR (RPE 14–17/20), interspersed with 9 × 1 minute low-intensity intervals at 30% HRR, for 19 minutes. T: adaptive recumbent steppers CON As for short HIIT but intensity differed: target 40% HRR (RPE, 9–11/20) for 20 minutes and progressed by 10% HRR and 5 minutes every 4 weeks up to 60% HRR (RPE, 13–14/20) for 30 min.	INT N = 9 during intervention 5 medical condition or injury unrelated to the intervention, 3 Covid-19 lockdown and 1 returned to work N = 9 at follow-up 4 not available for re-resting, 2 unable to be re-contacted, 1 Covid-19 lockdown, 1 MSK injury unrelated to the intervention, 1 unknown CON N = 13 during intervention 4 medical condition unrelated to the intervention, 3 Covid-19 lockdown, 2 transportation issues, 1 returned to work, 1 other rehabilitation priority, one did not enjoy intervention, 1 unknown N = 3 at follow-up 2 not available for re-resting, 1 unable to be re-contacted	INT N = 0 CON N = 0	INT n = 7 N = 7 CON n = 5 N = 5	INT N = 0 CON N = 0
Steen Krawcyk (2019)³⁹; RCT INT 31 (23/8); 63.7 ± 8.9; N/R CON 32 (26/6); 63.7 ± 9.2; N/R	INT F: 5 days per week, 12-week regimen: usual care + home-based HIIT daily I & T: 3 × 3 minutes at 77–93% HRmax or 14–16 RPE or ‘not able to speak comfortably’ on the Talk Test, 2 minutes of active recovery in between. T: different modes of aerobic exercise (e.g., brisk walking stair stepping, stationary bicycling, outdoor cycling, running, indoor rowing, high knee exercises and swimming) CON Usual care: secondary preventive medication and advice on self-managed lifestyle changes and habitual level of physical activity	INT N = 2 at allocation 2 reduced mental surplus N = 2 at follow-up 1 lumbar herniated disc allegedly induced by heavy lifting during a house renovation project, 1 work obligations CON N = 4 at follow-up 1 chronic pain in her knee from osteoarthritis and awaiting surgery, 1 reduced mental surplus, 2 no reason	INT N = 0 CON N = 0	INT n = 1 N = 1 CON n = 4 N = 4	INT N = 0 CON N = 0
Steen Krawcyk (2023)⁴²; follow-up of Steen Krawcyk (2019)³⁹	See Steen Krawcyk (2019)³⁹	INT N = 1 at 6 months follow-up 1 withdrew consent without reason N = 2 at 12 months follow-up 1 work obligations, 1 hospitalised due to TBI (not related to the intervention) CON N = 0 at 6 months follow-up N = 1 at 12 months follow-up 1 reduced mental surplus	INT N = 0 CON N = 0	INT n = 3 N = 3 CON N = 0	INT N = 0 CON N = 0
Valkenborghs (2019)³⁶; mixed methods RCT INT 9 (5/4); 62.1 ± 11.7; 34.6 ± 46.3 months CON 11 (6/5); 49.8 ± 17.4; 102.2 ± 108.8 months	INT F: AEX + TST = 30 minutes of AEX + 1 hour of TST over 10 weeks. TST: 60 hours of TST over 10 weeks (3 × 1 hour sessions with a therapist per week and 3 × 1 hour of home-based self-practice per week). AEX: 3-day, 10-week regimen. I & T: AEX: 4 × 4-minute intervals of high-intensity exercise (85% of HR_max) with a 3-minute active recovery (70% of HR_max) period between each interval per 30-minute session. T: low entry level upright or semi-recumbent cycle ergometer depending on individual ability and impairment CON TST for the upper limb: 60 hours of TST over 10 weeks (3 × 1-hour sessions with a therapist per week and 3 × 1 hour of home-based self-practice per week): TST: 100–300 repetitions of tasks per hour (e.g., extend elbow and flex shoulder to reach to cup, open fingers and thumb to grasp cup, and flex elbow and shoulder to transfer cup to mouth).	N = 1 1 bronchitis N/R	N = 0 N = 0	N = 0 N = 0	N = 0 N = 0

Note: Number (#) of participants analysed, if different from the number of participants randomised or allocated otherwise to the intervention. Adverse events were categorised according to descriptions extracted from each of the included studies: (1) ‘intervention-related’, i.e., adverse events, reported during the intervention period, described by the original study authors as being related to the intervention (e.g., syncope during exercise); (2) ‘non-intervention-related’, i.e., those adverse events that, according to the original study authors, were not related to the intervention, (e.g., a non-injurious fall in the follow-up period); (3) ‘unclear’, in cases where study authors did not describe an adverse event as either intervention-related or not.

Grade 1: mild; asymptomatic or mild symptoms; clinical or diagnostic observations only; intervention not indicated. Grade 2: moderate; minimal, local or non-invasive intervention indicated; limiting age-appropriate instrumental ADL. Grade 3: severe or medically significant but not immediately life-threatening; hospitalisation or prolongation of hospitalisation indicated; disabling; limiting self-care ADL. Grade 4: life-threatening consequences; urgent intervention indicated. Grade 5: Death related to AE.

For RCTs only. Intervention groups (236 participants): a total of 48 dropouts (20% of participants) including 22 during the intervention, 25 at follow-up and 1 N/S. Control groups (248 participants): a total of 35 dropouts (14% of participants) including 19 during the intervention and 16 at follow-up.

For RCTs only. Intervention groups (236 participants including 64 with AEs): a total of 164 AEs including 54 (38(1), 16(2)) related to the intervention, 106 (69(1), 17(2), 9(3), 1(5), 10 (N/S)) unrelated to the intervention and 4 unclear (4(N/S)). Control groups (248 participants including 55 with AEs): a total of 124 AEs including 39 (33(1), 6(2)) related to the intervention, 82 (48(1), 18(2), 7(3), 1(4), 8(N/S)) unrelated to the intervention and 3 unclear (3(N/S)).

For the meta-analysis of intervention-related adverse events, nine RCTs were excluded as one did not report intervention-related adverse events,³⁸ six RCTs had zero intervention-related adverse events in both study arms,^{12,36,39,40,43,44} one reported follow-up data⁴² and one focused on financial data only.⁴⁵ Meta-analysis of the remaining three RCTs^35,37,41 with a total of 120 participants showed no difference between the high-intensity interval training and the control group (which comprised moderate intensity exercise) in terms of the proportion of participants experiencing intervention-related adverse events (Risk Ratio (RR): 1.12, 95% confidence interval (CI): 0.51 to 2.49, I² = 27%, P =0.78, moderate certainty, Figure 2(a)). High-intensity interval training-related adverse events were mostly of severity Grade 1 and none were more serious than Grade 2 (Table 4).

Similarly, for the meta-analysis for non-intervention-related adverse events, six RCTs could be included,^{34,36,38,39,40,42} with a total of 343 participants. This also showed no between-group difference in the proportion of participants having a non-intervention-related adverse event (RR: 1.10, 95% CI: 0.85 to 1.43, I² = 0%, P =0.46, moderate certainty, Figure 2(b)).

Finally, for the meta-analysis for adverse events of which the cause was unclear, two RCTs could be included^12,42 with a total of 110 participants. Again, this showed no difference between the high intensity interval training and the control groups in the proportion of participants having an unclear adverse event (RR: 1.48, 95% CI: 0.37 to 6.00, I² = 0%, P =0.58, low certainty, Figure 2(c)).

Feasibility

Twenty studies reported the number of people assessed for eligibility, which represented 4007 participants. From these, the total number of people randomised or allocated otherwise to the intervention was 506 (on average 13% of individuals assessed). The study by Steen Krawcyk et al.³⁹ excluded 3027/3098 (98%) of the screened participants because they did not have a specific diagnosis of lacunar stroke. Omitting this study with this particular inclusion criterion resulted in an average of 59% of participants being eligible. All included participants then underwent careful pre-exercise screening, using a treadmill electrocardiogram or stress test prior to study intervention and stringent exclusion criteria were applied. The most common exclusion criteria were: unstable cardiopulmonary conditions (e.g., serious rhythm disorder and valve malfunction), evidence of electrocardiogram abnormalities, severe hypertonia and severe cognitive impairments.

All the Common Terminology Criteria for Adverse Event, dropouts and adverse events outside of the intervention periods have been fully documented in Supplementary Material D. Nineteen studies reported dropouts and two did not.^38,55 A total of 22 dropouts (5.8% of participants) were reported in the intervention groups during the intervention period. Common reasons included: medical conditions/withdrawals (due to factors mostly unrelated to study intervention) and external factors (e.g., COVID-19 and work) (Table 4).

For the meta-analysis of risk of dropouts during the intervention period, seven RCTs were excluded, as two did not report intervention-related dropouts in both study arms^36,38 and three had zero dropouts in both study arms^12,40,44; one study reported a follow-up⁴² and one reported financial data only.⁴⁵ Meta-analysis of the remaining five RCTs^{35,37,39,41,43} showed no difference in the risk of dropout between the HIIT and control groups (RR: 1.00, 95% CI: 0.58 to 1.74, I² = 0%, P = .99, moderate certainty, Figure 3).

Sixteen studies reported attendance to the high-intensity interval training interventions. The mean attendance was 94.4%, ranging from 82.3%³⁷ to 100%.^{12,33,35,46,50,52}

Only six studies reported on adherence to the planned content of the exercise intervention (Supplementary Material D). Two studies reportedly required some adaptation in their protocols; Valkenborghs et al.³⁶ modified their aerobic exercise protocol to make it feasible for more severely affected and non-ambulatory participants, and Boyne et al.⁵⁰ adapted their high-intensity interval training-stepper intervention after two participants experienced hypotension and near syncope during the recovery periods. Three studies^35,49,50 documented participants unable to complete some sessions due to exhaustion/fatigue. Lastly, Do et al.⁴⁴ reported ‘good adherence’ but did not quantify this.

There was a lack of clarity in the reporting of adherence to appropriate high-intensity interval training intensity. Out of 13 studies reporting intensity adherence, ten met the corresponding intensity criteria for ‘high-intensity’, set in their own exercise protocol.^{37,40,41,43,47–52} The remaining three studies^35,36,39 did not meet the required high-intensity interval training intensity or reported incomplete information.

Within the studies included, only the study by Hazen et al.⁴⁵ reported cost. For each participant who completed the intervention, baseline assessments costed (including graded exercise testing) US$1002, the intervention (including equipment) costed US$5130, medical translation and interpreter services costed US$1725.

Acceptability

The acceptability of high-intensity interval training was reported in five studies, but data were scarce and often anecdotal. There were no qualitative studies exploring acceptability. Of the two mixed methods studies appraised in Table 2, neither reported on the acceptability of high-intensity interval training. Gjellesvik et al.⁴⁰ mentioned that the intervention was ‘well tolerated’ by study participants, while Askim et al.⁴⁷ received ‘very good’ feedback from the participants. The study by Boyne et al.³⁵ revealed that the intervention was mostly favourable with initial feelings of apprehension, followed by increased confidence and enjoyment of high-intensity interval training. Ekechukwu, Omotosho and Hamzat³⁸ reported that in the first two weeks, participants were reluctant to exercise due to fear of a recurrent stroke. The study by Amanzonwé et al.⁴⁶ was the only study to utilise a published tool to explore participants’ perceptions of high intensity interval training. They reported high scores on both the credibility and expectancy subscales of the Credibility and Expectancy Questionnaire at admission and follow-up, with no significant changes over time. Together, these reports signal largely positive feedback on high-intensity interval training – but the nature and scarcity of data precluded synthesis as originally planned. The perception of study coordinators regarding the acceptability of high-intensity interval training was reported in five studies.^{35,37,39,40,48} The authors commonly mentioned the clinically feasible training volume, the modest time commitment and potential for future home-based options for ease and low-cost practice; however, there was a lack of data from providers of high-intensity interval training directly.

Differences between protocol and review

Risk ratios were calculated instead of risk differences, in accordance with the recommendations from the most recent Cochrane Handbook.²⁸ Given the paucity of data, planned subgroup or sensitivity analyses were not feasible. The analysis of the certainty of the evidence, using GRADE methodology, was undertaken subsequently.

Discussion

This is the first systematic review on the safety, feasibility and acceptability of high-intensity interval training post-stroke. The main findings indicate low to moderate GRADE level certainty that this intervention is safe and feasible for carefully screened, monitored and supervised, mostly male, relatively young, ambulant chronic stroke survivors with mild impairments, delivered in controlled environments by trained professionals. Given the demanding nature of high-intensity interval training, it is prudent to control risk factors; however, this affects the generalisability of the findings from this review.

The quantity of the body of evidence included was limited, comprising 20 primary studies with 658 participants, of which only 10 studies were RCTs.^{12,35–41,43,44} There were no qualitative studies, hence there were very limited data on intervention acceptability. The quality of the evidence varied considerably; only half of the quantitative studies were rated as ‘strong’, and the mixed-methods studies lacked coherence between qualitative data collection, reporting, analysis and interpretation of the findings.

Of the 18 studies that reported deaths, there was only one death, unrelated to the intervention. Synthesising adverse events was complicated due to the absence of a standardised method of reporting, lack of clarity in attribution and missing data. Eleven studies reported zero high-intensity interval training-related adverse events. There was no difference in the risk of intervention-related adverse events between the high-intensity interval training and the moderate intensity exercise control groups, but the certainty of the evidence was moderate. Most events were classified as ‘mild; asymptomatic or mild symptoms; clinical or diagnostic observations only; intervention not indicated’²⁰ and mainly related to general pain and light-headedness. It is likely that some participants were deconditioned pre-intervention, which could be prevented in future by starting exercise interventions at lower intensities and slowly progressing. Similarly, there were no differences between groups in the proportion of participants having an adverse event reported as being unrelated to the intervention (moderate level of certainty), or an adverse event of which the cause was unclear (low level of certainty). Overall, the absence of intervention-related fatalities and the low rate of mainly mild adverse events may be due to the rigorous eligibility criteria, the specialist environment, close supervision and monitoring by trained staff. These findings align with those from another systematic review on the safety of high-intensity interval training for people with cardiovascular disease,⁵⁷ which found an overall cardiovascular event rate (including one major and non-fatal) in the high intensity interval training groups of 1 per 5667 training hours compared to no cardiovascular events in the moderate intensity training control groups, and no significant difference in risk difference for all adverse events between the high intensity and moderate intensity groups.

With regard to feasibility, this review found that across all studies – with one excluded due to specific inclusion criteria related to lesion location³⁹ – an average of 59% of participants identified were deemed eligible. There was moderate certainty evidence of no difference in the risk of dropout between the intervention and control groups in RCTs. Furthermore, dropout rates were relatively low and acceptable,⁵⁸ and occurred mostly at follow-up. The main reasons for dropout were medical conditions other than stroke, unspecified withdrawals and external factors. The rigorous inclusion criteria and pre-exercise screening may reduce eligibility; however, findings suggest that, once participants were allocated to the intervention, retention was high. The high average attendance rate (94.4%) supports the feasibility of high-intensity interval training, which may be attributed to the careful screening, individualised nature and close supervision provided. Recording adherence to intervention intensity is critical as it indicates how well participants are able to meet the targeted intensity⁵⁹ – but this was patchy in the studies included. Most studies that did report on adherence to exercise intensity showed that stroke survivors who met rigorous selection criteria exercised at planned intensities, while some indicated that interventions needed to be adapted. There were no qualitative studies about the acceptability of high intensity interval training according to stroke survivors or professionals. Hence, data on acceptability were scarce but where reported, stroke survivors indicated a mostly favourable experience – despite some initial feelings of apprehension. Only one study⁴⁵ reported costs of high-intensity interval training, provided as part of an RCT conducted in USA. Further economic analyses of high-intensity interval training are needed to strengthen the evidence about the financial feasibility of this type of intervention.

There were a number of limitations to this review, despite having applied rigorous frameworks, including for the review itself,^15,16 extraction of data related to the intervention^22,60 and adverse events.²⁰ It would have been useful to compare the risk of adverse events between the high-intensity interval training and the different control interventions (e.g., no treatment, usual care, low intensity or moderate intensity). However, as the control groups in the majority of RCTs involved moderate intensity exercise, subgroup analyses were not considered to be meaningful. Due to resource constraints, studies in languages other than English had to be excluded, and relevant literature may have been overlooked. Additionally, resource constraints meant that only published studies could be included, and this may have contributed to publication and non-reporting biases. A funnel plot analysis was not considered appropriate however, as the number of studies included in each analysis was well below 10.⁶¹ Sensitivity analyses were also not feasible, given the paucity of data. Missing data were fully reported; however, the associated risk of bias was not analysed formally.

Several implications for research and practice arise from this review. Considering the nature of high-intensity interval training, it is prudent that this was tested with carefully selected participants, often in the chronic stage, and in specialist settings, under professional supervision. Practitioners should be encouraged by the emerging evidence of the safety and feasibility of high-intensity interval training in stroke populations, however, with the caveats above in mind. Future studies should explore the safety, feasibility and acceptability of high-intensity interval training with a population of stroke survivors with a wider range of characteristics (e.g., female stroke survivors and individuals in the subacute stage).

To strengthen the evidence base, consistent reporting of levels of stroke severity, including ambulatory status, will be necessary, based on recommended measures.⁶² It will also be essential to develop agreed, stroke-specific high-intensity interval training protocols and standardise the reporting of adverse events (e.g., using the Common Terminology Criteria for Adverse Event template).²⁰ To facilitate implementation, further research is also recommended to investigate the safety, feasibility and acceptability of high-intensity interval training in community environments with trained personnel. Furthermore, high quality qualitative studies are needed to gain a deeper understanding of the experiences, barriers and facilitators for this type of training among both service users and providers.

In conclusion, findings from this systematic review and meta-analysis indicate low-moderate certainty evidence that high-intensity interval training can be safe and feasible for stroke survivors who are generally younger, mildly affected, male and in the chronic stage, who have been carefully selected and supervised by trained professionals in controlled settings. There was very limited evidence about the acceptability of high-intensity interval training; however, where documented, participants’ experiences were mostly favourable.

Clinical messages

High-intensity interval training after stroke may be safe and feasible for carefully screened stroke survivors, when supervised by trained staff in controlled environments.

In these contexts, there were no deaths related to high-intensity interval training, and there was no difference in the risk of adverse events between training and control groups (low to moderate certainty evidence) – where reported, adverse events were generally mild and transient.

Attendance at high-intensity interval training sessions approached 95% on average. There was no difference in dropouts between training and control groups (moderate certainty evidence).

There were very limited data about the acceptability of high-intensity interval training, but where reported, stroke survivors' experiences were generally favourable – despite some expressing initial feelings of apprehension.

Supplemental Material

sj-docx-1-cre-10.1177_02692155251385222 - Supplemental material for High-intensity interval training after stroke: A mixed-methods systematic review and meta-analysis of safety, feasibility and acceptability

Supplemental material, sj-docx-1-cre-10.1177_02692155251385222 for High-intensity interval training after stroke: A mixed-methods systematic review and meta-analysis of safety, feasibility and acceptability by Hugo Blatgé, Lorna Paul and Frederike van Wijck in Clinical Rehabilitation

Supplemental Material

sj-docx-2-cre-10.1177_02692155251385222 - Supplemental material for High-intensity interval training after stroke: A mixed-methods systematic review and meta-analysis of safety, feasibility and acceptability

Supplemental material, sj-docx-2-cre-10.1177_02692155251385222 for High-intensity interval training after stroke: A mixed-methods systematic review and meta-analysis of safety, feasibility and acceptability by Hugo Blatgé, Lorna Paul and Frederike van Wijck in Clinical Rehabilitation

Supplemental Material

sj-docx-3-cre-10.1177_02692155251385222 - Supplemental material for High-intensity interval training after stroke: A mixed-methods systematic review and meta-analysis of safety, feasibility and acceptability

Supplemental material, sj-docx-3-cre-10.1177_02692155251385222 for High-intensity interval training after stroke: A mixed-methods systematic review and meta-analysis of safety, feasibility and acceptability by Hugo Blatgé, Lorna Paul and Frederike van Wijck in Clinical Rehabilitation

Supplemental Material

sj-docx-4-cre-10.1177_02692155251385222 - Supplemental material for High-intensity interval training after stroke: A mixed-methods systematic review and meta-analysis of safety, feasibility and acceptability

Supplemental material, sj-docx-4-cre-10.1177_02692155251385222 for High-intensity interval training after stroke: A mixed-methods systematic review and meta-analysis of safety, feasibility and acceptability by Hugo Blatgé, Lorna Paul and Frederike van Wijck in Clinical Rehabilitation

Footnotes

Acknowledgements

The authors would like to thank Kirsten McCormick (academic librarian) for helping with the review search strategy.

ORCID iDs

Hugo Blatgé

Lorna Paul

Frederike van Wijck

Author contributions

Hugo Blatgé: study design, literature searching, title and abstract screening, full text screening, risk of bias assessment, data extraction, data analysis, data interpretation, preparation of the first manuscript draft and review of manuscript and editing.

Frederike van Wijck: study design, title and abstract screening, full text screening, risk of bias assessment, data extraction verification, data analysis, data interpretation, review of manuscript and editing.

Lorna Paul: study design, data interpretation, review of manuscript and editing.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Frederike van Wijck received royalties from Churchill Livingstone Elsevier as a co-editor and co-author of a book on fitness training after stroke.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Data availability statement

All data reported in this review are included in the publication (including supplemental materials).

Supplemental material

Supplemental material for this article is available online.

References

Intercollegiate Stroke Working Party . National clinical guideline for stroke for the UK and Ireland. London: Intercollegiate Stroke Working Party, 2023. www.strokeguideline.org .

Crozier

Roig

Eng

, et al. High-intensity interval training after stroke: an opportunity to promote functional recovery, cardiovascular health, and neuroplasticity. Neurorehabil Neural Repair 2018; 32: 543–556.

Roy

. High-intensity interval training: efficient, effective, and a fun way to exercise: brought to you by the American college of sports medicine www.acsm.org. ACSMs Health Fit J 2013; 17: 3.

Weston

Wisløff

Coombes

. High-intensity interval training in patients with lifestyle-induced cardiometabolic disease: a systematic review and meta-analysis. Br J Sports Med 2014; 48: 1227–1234.

American College of Sports Medicine . ACSM’s guidelines for exercise testing and prescription (11th edition). Philadelphia: Wolters Kluwer, 2021.

van Duijnhoven

Heeren

Peters

, et al. Effects of exercise therapy on balance capacity in chronic stroke: systematic review and meta-analysis. Stroke 2016; 47: 2603–2610.

Luo

Meng

Wang

, et al. Effect of high-intensity exercise on cardiorespiratory fitness in stroke survivors: a systematic review and meta-analysis. Ann Phys Rehabil Med 2020; 63: 59–68.

Anjos

Neto

Dos Santos

, et al. The impact of high-intensity interval training on functioning and health-related quality of life in post-stroke patients: a systematic review with meta-analysis. Clin Rehabil 2022; 36: 726–739.

Wiener

McIntyre

Janssen

, et al. Effectiveness of high-intensity interval training for fitness and mobility post stroke: a systematic review. PM&R 2019; 11: 868–878.

10.

Huang

Zhang

Zhou

, et al. Lactate as potential mediators for exercise-induced positive effects on neuroplasticity and cerebrovascular plasticity. Front Physiol 2021; 12: 656455.

11.

Hugues

Pellegrino

Rivera

, et al.

Is high-intensity interval training suitable to promote neuroplasticity and cognitive functions after stroke?

Int J Mol Sci 2021; 22: 3003.

12.

Hsu

C-C

T-C

Huang

S-C

, et al. Increased serum brain-derived neurotrophic factor with high-intensity interval training in stroke patients: a randomized controlled trial. Ann Phys Rehabil Med 2021; 64: 101385.

13.

Saunders

Sanderson

Hayes

, et al. Physical fitness training for stroke patients. Cochrane Database Syst Rev 2020; (3): CD003316.

14.

Gaskins

Bray

Hill

, et al. Factors influencing implementation of aerobic exercise after stroke: a systematic review. Disabil Rehabil 2021; 43: 2382–2396.

15.

Page

McKenzie

Bossuyt

, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Int J Surg 2021; 88: 105906.

16.

Rethlefsen

Kirtley

Waffenschmidt

, et al. PRISMA-S: an extension to the PRISMA statement for reporting literature searches in systematic reviews. Syst Rev 2021; 10: 1–19.

17.

Higgins

Thomas

Chandler

, et al. Defining the criteria for including studies and how they will be grouped for the synthesis. In: Cochrane handbook for systematic reviews of interventions. Cochrane, https://training.cochrane.org/handbook/archive/v6 (2019, accessed 18 November 2024).

18.

Mezzani

Hamm

Jones

, et al. Aerobic exercise intensity assessment and prescription in cardiac rehabilitation: a joint position statement of the European Association for Cardiovascular Prevention and Rehabilitation, the American Association of Cardiovascular and Pulmonary Rehabilitation and the Canadian Association of Cardiac Rehabilitation. Eur J Prev Cardiol 2013; 20: 442–467.

19.

MacInnis

Gibala

. Physiological adaptations to interval training and the role of exercise intensity. J Physiol 2017; 595: 2915–2930.

20.

National Institutes of Health. Common Terminology Criteria for Adverse Events (CTCAE) v5.0. Cancer Therapy Evaluation Program, 2023. https://ctep.cancer.gov/protocoldevelopment/electronic_applications/docs/CTCAE_v5_Quick_Reference_5x7.pdf (2017, accessed 12 November 2024).

21.

Veritas Health Innovation. Covidence Systematic Review Software, www.covidence.org (2024).

22.

Slade

Dionne

Underwood

, et al. Consensus on exercise reporting template (CERT): explanation and elaboration statement. Br J Sports Med 2016; 50: 1428–1437.

23.

Holden

Gill

Magliozzi

, et al. Clinical gait assessment in the neurologically impaired: reliability and meaningfulness. Phys Ther 1984; 64: 35–40.

24.

Thomas

Ciliska

Dobbins

, et al. A process for systematically reviewing the literature: providing the research evidence for public health nursing interventions. Worldviews Evidence-Based Nurs 2004; 1: 176–184.

25.

Hong

Fàbregues

Bartlett

, et al. The mixed methods appraisal tool (MMAT) version 2018 for information professionals and researchers. Educ Inf 2018; 34: 285–291.

26.

Noyes

Booth

Moore

, et al. Synthesising quantitative and qualitative evidence to inform guidelines on complex interventions: clarifying the purposes, designs and outlining some methods. BMJ Glob Health 2019; 4: e000893.

27.

Deeks

Higgins

Altman

, et al. Analysing data and undertaking meta-analyses. In: Cochrane handbook for systematic reviews of interventions version 6.5. Cochrane. https://training.cochrane.org/handbook/archive/v6.4 (2024, accessed 4 April 2025).

28.

Higgins

Thomas

Chandler

, et al. Cochrane handbook for systematic reviews of interventions. Cochrane, 2024. www.training.cochrane.org/handbook .

29.

Finfgeld-Connett

. Use of content analysis to conduct knowledge-building and theory-generating qualitative systematic reviews. Qual Res 2014; 14: 341–352.

30.

Schünemann

Higgins

Vist

, et al. Completing ‘summary of findings’ tables and grading the certainty of the evidence. In: Cochrane handbook for systematic reviews of interventions. Cochrane, 2024. cochrane.org/handbook.

31.

Carl

Boyne

Rockwell

, et al. Preliminary safety analysis of high-intensity interval training (HIIT) in persons with chronic stroke. Appl Physiol Nutr Metab 2017; 42: 311–318.

32.

Valkenborghs

Erickson

Nilsson

, et al. Feasibility of aerobic interval training in nonambulant persons after stroke. J Clin Exerc Physiol 2019; 8: 97–101.

33.

Boyne

Meyrose

Westover

, et al. Effects of exercise intensity on acute circulating molecular responses poststroke. Neurorehabil Neural Repair 2020; 34: 222–234.

34.

Gjellesvik

Becker

Tjønna

, et al. Effects of high-intensity interval training after stroke (the HIIT stroke study) on physical and cognitive function: a multicenter randomized controlled trial. Arch Phys Med Rehabil 2021; 102: 1683–1691.

35.

Boyne

Dunning

Carl

, et al. High-intensity interval training and moderate-intensity continuous training in ambulatory chronic stroke: feasibility study. Phys Ther 2016; 96: 1533–1544.

36.

Valkenborghs

van Vliet

Nilsson

, et al. Aerobic exercise and consecutive task-specific training (AExaCTT) for upper limb recovery after stroke: a randomized controlled pilot study. Physiother Res Int 2019; 24: e1775.

37.

Boyne

Billinger

Reisman

, et al. Optimal intensity and duration of walking rehabilitation in patients with chronic stroke: a randomized clinical trial. JAMA Neurol 2023; 80: 342.

38.

Ekechukwu

Omotosho

Hamzat

. Comparative effects of interval and continuous aerobic training on haematological variables post-stroke–a randomized clinical trial. Afr J Physiother Rehabil Sci 2017; 9: 1–8.

39.

Steen Krawcyk

Vinther

Petersen

, et al. Effect of home-based high-intensity interval training in patients with lacunar stroke: a randomized controlled trial. Front Neurol 2019; 10: 664.

40.

Gjellesvik

Becker

Tjønna

, et al. Effects of high-intensity interval training after stroke (the HIIT-stroke study): a multicenter randomized controlled trial. Arch Phys Med Rehabil 2020; 101: 939–947.

41.

Marzolini

Robertson

MacIntosh

, et al. Effect of high-intensity interval training and moderate-intensity continuous training in people with poststroke gait dysfunction: a randomized clinical trial. J Am Heart Assoc 2023; 12: e031532.

42.

Steen Krawcyk

Vinther

Petersen

, et al. High-intensity training in patients with lacunar stroke: a one-year follow-up. J Stroke Cerebrovasc Dis 2023: 106973.

43.

Moncion

Rodrigues

De Las Heras

, et al. Cardiorespiratory fitness benefits of high-intensity interval training after stroke: a randomized controlled trial. Stroke 2024; 55: 2202–2211.

44.

Lim

W-T

Kim

, et al. Effects of high-intensity interval robot-assisted gait training on cardiopulmonary function and walking ability in chronic stroke survivors: a multicenter single-blind randomized controlled trial. J Back Musculoskelet Rehabil 2024; 37: 1309–1319.

45.

Hazen

Bartsch

Billinger

. Financial analysis of a locomotor exercise trial for post-stroke recovery: insights from the HIT Stroke Trial. Front Stroke 2024; 3: 1425385.

46.

Amanzonwé

Kossi

Noukpo

, et al. High-intensity interval training is feasible, credible and clinically effective in the early subacute stroke stage in the low-income country of Benin. J Sports Sci 2024; 42: 1030–1040.

47.

Askim

Dahl

Aamot

, et al. High-intensity aerobic interval training for patients 3–9 months after stroke: a feasibility study. Physiother Res Int 2014; 19: 129–139.

48.

Bosch

Holzapfel

Nordin

, et al. High-intensity interval training for adults with chronic stroke: a pilot feasibility study. Cardiopulm Phys Ther J 2021; 32: 20–29.

49.

Boyne

Dunning

Carl

, et al. Within-session responses to high-intensity interval training in chronic stroke. Med Sci Sports Exerc 2015; 47: 476–484.

50.

Boyne

Meyrose

Westover

, et al. Exercise intensity affects acute neurotrophic and neurophysiological responses poststroke. J Appl Physiol 2019; 126: 431–443.

51.

Boyne

Scholl

Doren

, et al. Locomotor training intensity after stroke: effects of interval type and mode. Top Stroke Rehabil 2020; 27: 483–493.

52.

Boyne

Doren

Scholl

, et al. Preliminary outcomes of combined treadmill and overground high-intensity interval training in ambulatory chronic stroke. Front Neurol 2022; 75.

53.

Calmels

Degache

Courbon

, et al. The feasibility and the effects of cycloergometer interval-training on aerobic capacity and walking performance after stroke. Preliminary study. Ann Phys Rehabil Med 2011; 54: 3–15.

54.

Gjellesvik

Brurok

Hoff

, et al. Effect of high aerobic intensity interval treadmill walking in people with chronic stroke: a pilot study with one year follow-up. Top Stroke Rehabil 2012; 19: 353–360.

55.

Lazarieva

Vinohradov

. The effect of high-intensity interval training on post-stroke patients. Zdr Listy 2024: 6–11.

56.

Mehrholz

Wagner

Rutte

, et al. Predictive validity and responsiveness of the functional ambulation category in hemiparetic patients after stroke. Arch Phys Med Rehabil 2007; 88: 1314–1319.

57.

Wewege

Ahn

, et al. High-intensity interval training for patients with cardiovascular disease—is it safe? A systematic review. J Am Heart Assoc 2018; 7: e009305.

58.

National Institutes of Health. Study quality assessment tools. National Heart, Lung, and Blood Institute. https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools (2021, accessed 12 November 2024).

59.

Wood

Olive

LaValle

, et al. Dissimilar physiological and perceptual responses between sprint interval training and high-intensity interval training. J Strength Cond Res 2016; 30: 244–250.

60.

Hoffmann

Glasziou

Boutron

, et al. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide. Br Med J 2014; 348: g1687–g1687.

61.

Sterne

JAC

Sutton

Ioannidis

JPA

, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. Br Med J 2011; 343: d4002–d4002.

62.

Van Criekinge

Heremans

Burridge

, et al. Standardized measurement of balance and mobility post-stroke: consensus-based core recommendations from the third stroke recovery and rehabilitation roundtable. Int J Stroke 2024; 19: 158–168.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB

0.16 MB

0.02 MB

0.15 MB