Submaximal Exercise Testing to Dose High-Intensity Interval Training After Stroke: The FAST Randomized Clinical Trial

Abstract

Importance:

High-intensity interval training (HIIT) improves peak oxygen uptake (VO₂peak) and walking post-stroke. However, previous HIIT trials have primarily implemented maximal exercise testing, limiting clinical implementation.

Objective:

Evaluate the preliminary efficacy of HIIT, compared to moderate-intensity continuous training (MICT) using a submaximal exercise test. Hypothesis: HIIT will produce greater improvements than MICT in VO₂peak, vascular measures, and walking outcomes.

Design:

This was a randomized preliminary efficacy trial conducted between July 2023 and December 2025.

Setting:

University of Kansas Medical Center.

Participants:

Participants with chronic stroke, 20–85 years of age, were randomized to HIIT or MICT.

Intervention:

HIIT and MICT were performed on a total-body recumbent stepper 3 times per week for 4 weeks, with intensity prescribed using peak power output (PPO) to achieve target heart rate zones derived from a submaximal exercise test. HIIT was performed for 25 minutes with 1-minute vigorous-intensity intervals (65–95% PPO) interspersed with 1-minute active recovery intervals. MICT was performed continuously at 45–65% PPO for 25 minutes.

Main Outcomes:

The primary outcome was change in predicted VO₂peak. Secondary outcomes included middle cerebral artery velocity, peripheral vascular function, and arterial stiffness with gait speed and walking endurance as tertiary outcomes.

Results:

Forty-nine participants (HIIT: n=25, MICT: n=24) were randomized (62.4(12.5) years, 42.9% female), attended 99.5(2.0)% of sessions, and achieved target intensity zones. No study-related serious adverse events occurred. Our results showed no significant between-group differences (p=0.54) for study outcomes. Both groups significantly improved VO₂peak (HIIT: +1.13 mL•kg⁻¹•min⁻¹ (95% CI: 0.05–2.21), p=0.04; MICT: +1.58 mL•kg⁻¹•min⁻¹ (95% CI: 0.18–2.97), p=0.03) and with fast gait speed and walking endurance. Peripheral vascular function significantly improved following HIIT.

Conclusions and Relevance:

HIIT can be safely implemented in individuals with chronic stroke using a submaximal exercise test. Both HIIT and MICT elicited clinically meaningful gains in VO₂peak and walking. However, only HIIT led to a significant improvement in peripheral vascular function, suggesting a biologic signal for intensity-dependent vascular adaptation.

Trial Registration:

ClinicalTrials.gov identifier: NCT05936008.

Introduction

Exercise intensity, specifically vigorous intensity, appears to function as a critical dosing ingredient for improving both peak oxygen consumption (VO₂peak)^1–3 and walking outcomes^4–6 after stroke. A recent systematic review and meta-analysis demonstrated high-intensity interval training (HIIT) elicits greater gains in cardiorespiratory fitness than moderate-intensity continuous training (MICT) in people with stroke.² Consequently, HIIT is increasingly viewed as a promising rehabilitation strategy.

Despite this growing evidence, most prior HIIT trials share key design features that limit clinical translation. Protocols have largely relied on treadmill-based training with harness for safety^4,7 and maximal graded exercise testing with electrocardiographic (ECG) monitoring to establish eligibility, ensure safety, and determine exercise intensity targets.^3,4,7–10 While methodologically rigorous, this approach creates a paradox: interventions demonstrating benefit require resources unavailable in the majority of stroke rehabilitation settings and community settings.^11,12 Maximal exercise testing is infrequently performed in routine stroke care, and reliance on it may impede adoption for safety concerns even if HIIT proves superior to MICT. Therefore, an essential next step is determining whether HIIT can be accurately prescribed using methods feasible for real-world implementation.

To address this gap, our group previously developed and validated a total-body recumbent stepper (TBRS) submaximal exercise test in healthy adults¹³ and older adults.¹⁴ The test demonstrates strong agreement between predicted and measured VO₂peak¹⁵ and high reliability¹⁶ in people with stroke. We have subsequently applied the TBRS submaximal test to accurately prescribe an acute bout of high-intensity interval exercise in healthy young adults, older adults, and individuals post-stroke while characterizing cerebrovascular and autonomic physiological responses.^17–21 Submaximal testing allows individualized dosing without requiring maximal exertion, specialized equipment, or cardiopulmonary exercise testing expertise¹³ and we showed in people with chronic stroke, target heart rates reached thresholds for vigorous intensity.¹⁹

The Fitness After Stroke (FAST) randomized clinical trial evaluated the preliminary efficacy of short-interval, high-volume HIIT compared with MICT using a recumbent stepper modality that allows for inclusion of people with stroke who have a wide range of physical abilities.^8,22,23 We hypothesized that HIIT would produce greater improvements in VO₂peak than MICT, with corresponding improvements in vascular measures and walking outcomes.

METHODS

Study Design and Oversight

The FAST trial was a single-site, randomized, parallel-group preliminary efficacy trial comparing HIIT with MICT in individuals with chronic stroke. The trial was registered at ClinicalTrials.gov (NCT05936008) before enrollment. The University of Kansas Medical Center Institutional Review Board approved all procedures, and written informed consent was obtained from all participants prior to study procedures. This randomized clinical trial is reported in accordance with the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. The full study protocol including inclusion and exclusion criteria is provided in Supplement 1 and are described in the published methods paper.²⁰

Participants

Community-dwelling adults aged 20 to 85 years with ischemic or hemorrhagic stroke ≥6 months prior to enrollment were recruited between July 2023 and December 2025 through the University of Kansas Stroke Recovery Registry, physician clinics, outpatient physical therapy clinics, and community outreach events.

Randomization and Blinding

Participants were allocated 1:1 to HIIT or MICT using minimization²⁴ stratified by lower extremity motor function, defined by the Fugl-Meyer Assessment–Lower Extremity²⁵ score. The first participant was randomly assigned with subsequent participants allocated using weighted randomization with 80% probability to whichever group would improve balance. Group allocation was performed by an unblinded team member not involved in study assessments.

Outcome assessments were conducted by blinded assessors, and study statistician (RNM) was masked to group allocation. The principal investigator was blinded to primary and secondary outcomes identified in ClinicalTrials.gov. Participants were not informed of intervention labels (HIIT or MICT), specific exercise structure (interval vs continuous) or study hypotheses to minimize expectancy bias.²⁰

Exercise Interventions

Participants trained on the TBRS 3 times per week for 4 weeks with one-on-one supervision and continuous heart rate monitoring. Exercise intensity was prescribed using peak power output (PPO) derived from the TBRS submaximal exercise test. Each session included a standardized warm-up and cool-down. Rating of perceived exertion²⁶ was recorded immediately after the intervention and again after cooldown. Blood pressure was taken using sphygmomanometer and stethoscope between minutes 16–17 of the intervention. Capillary blood lactate was obtained immediately post-exercise at sessions 2, 5, 8, and 11 as a marker of training intensity.⁴

The HIIT protocol consisted of repeated 1-minute high-intensity intervals alternated with 1-minute active recovery for 25 minutes. High-intensity intervals targeted 65%–95% of PPO at 90–100 steps per minute (spm) and recovery intervals were performed at 10% PPO and ~50 spm. The MICT protocol consisted of continuous exercise for 25 minutes at 45%–65% of PPO at 90–100 spm.

Outcome Measures

Assessments were performed before and within one week after completion of the intervention.

Primary Outcome

The primary outcome was change in predicted VO₂peak, assessed during the TBRS submaximal exercise test.

Secondary Outcomes

Cerebrovascular hemodynamics were assessed using transcranial Doppler ultrasound to measure middle cerebral artery velocity (MCAv) at rest. Peripheral vascular function was assessed using brachial artery flow-mediated dilation, and arterial stiffness was measured using carotid-femoral pulse wave velocity (cfPWV).

Tertiary Outcomes

Gait speed and walking endurance were assessed using the 10-meter walk test and 6-minute walk test, respectively.

Intervention Acceptability

Exercise enjoyment was assessed using the 8-item Physical Activity Enjoyment Scale (PACES-8) during the post-intervention assessment visit.

Adverse Event Monitoring

Prior to each study visit, we inquired about any medical changes that occurred since the prior visit. The safety of the intervention was evaluated by monitoring adverse events (AEs), defined as “any physical or psychological sign, symptom, or disease experienced during the study period that is temporally, but not necessarily causally, related to the intervention”.²⁷ Consistent with standardized guidance for adverse event and serious adverse event reporting”,²⁸ adverse events were classified using the National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0.²⁹ An independent physician blinded to group assignment served as the independent adjudicator for relatedness, expectedness and severity.

Statistical Analysis

This preliminary efficacy trial was designed to estimate effect sizes rather than test definitive hypotheses. Descriptive statistics are reported as mean (standard deviation) or median (interquartile range), as appropriate. Between-group differences in change scores were calculated with corresponding 95% confidence intervals. Effect sizes were computed as the between-group difference divided by the pooled standard deviation. Between group differences were assessed using analysis of covariance (ANCOVA) models. For the primary outcome, VO₂peak at 4 weeks was the response variable and the model was adjusted for baseline VO₂peak, group, and Fugl-Meyer score. Model fits were assessed using residual diagnostic plots and observed versus predicted plots. Analyses followed an intention-to-treat approach. Within-group changes were assessed post-hoc using paired t-tests, with p-values adjusted for multiple comparisons using the Benjamini-Hochberg procedure. Statistical analyses were performed using R Studio (versions 4.3.1 and 4.5.1).

RESULTS

Recruitment

Fifty-nine participants were consented, of whom 50 were eligible for randomization. One participant withdrew prior to randomization, resulting in 24 participants allocated to MICT, and 25 to HIIT (Figure 1). Participants had a mean age of 62.4 (SD=12.5, range=35–82) years. Demographic characteristics are presented in Table 1.

Figure 1.

CONSORT Flow Diagram

Table 1.

Baseline Participant Characteristics

	Participantsa
Characteristic	MICT (n=24)	HIIT (n=25)	p-value
Age, mean (SD), y	62.42 ± 12.94	62.44 ± 12.33	0.99
Sex, n(%)
Female	12 (50)	9 (36)	0.32
Male	12 (50)	16 (64)
Race, n(%)
Black/African American	4 (17)	7 (28)	0.34
White	20 (83)	18 (72)
BMI, mean (SD)	29.13 ± 5.92	31.40 ± 7.97	0.26
Education, n(%)
High school	2 (8)	5 (20)	0.31
Some college	8 (33)	6 (24)
Associate’s	1 (4)	1 (4)
Bachelor’s	5 (21)	9 (36)
Master’s	8 (33)	3 (12)
Professional	0 (0)	1 (4)
Fugl-Meyer lower limb motor score, mean (SD)b	29.5 ± 5.2	26.8 ± 5.9	0.10
Handedness, n(%)
Right	21 (87)	24 (96)	0.277
Left	3 (12)	1 (4)
Stroke characteristics
Side of paresis, n(%)
Left	7 (29)	14 (56)	0.06
Right	17 (70)	11 (44)
Stroke Type, n(%)
Ischemic	21 (87)	15 (60)	0.06
Hemorrhagic	4 (16)	10 (40)
Stroke Location, n(%)
Left Hemisphere	11 (46)	7 (28)	0.90
Right Hemisphere	6 (25)	9 (36)
Left Subcortical	3 (13)	2 (8)
Right Subcortical	1 (4)	2 (8)
Brainstem	2 (8)	3 (12)
Cerebellar	1 (4)	1 (4)
Bilateral	1 (4)	1 (4)
Stroke chronicity, mean (SD), months	30.80 ± 34.64	28.44 ± 24.11	0.78
Cardiovascular Risk
Known Disease, n(%)			0.63
Cardiac disease	2 (8)	5 (20)
Peripheral vascular disease	0 (0)	1 (4)
Cerebrovascular disease	24 (100)	25 (100)
Type I diabetes	0 (0)	0 (0)
Type II diabetes	6 (25)	4 (16)
Renal disease	2 (8)	2 (8)
Smoking history			0.52
Current	0 (0)	1 (4)
Former	12 (50)	10 (40)
Never	12 (50)	14 (56)
CAD risk factors, n(%)			0.67
Age	19 (79)	24 (96)
Family History	11 (46)	5 (20)
Cigarette smoking	2 (8)	1 (4)
Physical inactivity	22 (92)	24 (96)
BMI	6 (25)	11 (44)
Blood pressure	18 (75)	22 (88)
Lipids	22 (92)	21 (84)
Blood glucose	7 (29)	8 (32)
HDL-C	0 (0)	1 (4)
Prescribed medication, n(%)			0.27
Antihypertensive	17 (71)	21 (84)
Statin	20 (83)	20 (80)
Anticoagulant	19 (79)	13 (52)
Antispasmodic	4 (17)	8 (32)
Antidepressant	11 (46)	8 (32)
Pain reliever	13 (54)	5 (20)
Stroke impact scale, mean(SD)
Raw score	250.3 ± 20.3	235.4 ± 31.8	0.06
Strength domain	65.9 ± 20.7	60.2 ± 27.1	0.42
Memory and thinking domain	87.1 ± 7.4	79.9 ± 18.4	0.08
Emotion domain	88.5 ± 10.1	81.6 ± 15.1	0.06
Communication domain	90.9 ± 13.2	86.7 ± 15.5	0.31
ADL/IADL domain	88.3 ± 11.0	84.4 ± 13.1	0.26
Mobility domain	86.1 ± 12.8	83.3 ± 15.3	0.49
Hand function domain	71.0 ± 28.8	53.6 ± 38.8	0.08
Participation/role function domain	82.4 ± 18.3	74.1 ± 21	0.15

Abbreviations: BMI, body mass index; mo, month; HDL, high-density lipoprotein; CAD, coronary artery disease; ADL, activities of daily living; IADL, instrumental activities of daily living

^aData are presented as the number (percentage) of participants unless otherwise indicated

^bRange, 0 to 34; higher scores indicate less motor impairment.

Treatment Fidelity

Overall session attendance averaged 99.5% (SD=2.0%; range=91.7%–100%) with no between-group differences (HIIT: 99.0% [2.7%]; MICT: 100% [0%]; p = 0.08). In the HIIT group, participants achieved heart rate values for vigorous intensity during 80.6% (17.7%) of prescribed high-intensity intervals, reaching a mean peak heart rate of 78.8% (3.8%) of HRmax (Figure 2) and a mean postexercise lactate concentration of 3.78 (SD=1.58, range=1.4–8.6) mmol/L. In the MICT group, 67.5% (25.3%) of training time was performed within the moderate-intensity zone, with a mean heart rate of 67.3% (6.9%) of HRmax and lactate of 3.04 (SD=0.96, range=1.4–5.8) mmol/L. The lower proportion of time spent within the target HR zone in the MICT group may reflect transient decreases in HR during blood pressure measurement (Figure 2). Postexercise lactate concentration was higher in the HIIT group (mean difference=0.735, 95% CI [0.353, 1.117], p<0.001).

Figure 2.

Heart Rate Responses During HIIT and MICT Sessions.

Mean heart rate (black line) and SD (gray shading) across all training sessions for participants in the HIIT (A) and MICT (B) groups. Dotted lines indicate prespecified target intensity zones. %HRmax indicates percentage of age-predicted maximum heart rate. The transient decrease in heart rate in the MICT group corresponds to brief pauses for blood pressure measurement.

Exercise enjoyment did not differ between groups (PACES-8: HIIT, 47.4 [6.2]; MICT, 46.9 [7.7]; p = 0.81). However, participant comments supported high acceptability across both interventions, with descriptions of enjoyment, perceived physical benefits, improved exercise confidence, and plans to continue regular exercise after the intervention (see Supplement 1 for details). Notably, participants in the HIIT group provided more frequent and detailed feedback, particularly regarding enjoyment of the interval structure and perceived benefits of training.

Adverse Events

Seventeen participants experienced a total of 24 adverse events (AEs), with 10 occurring in the HIIT group and 14 in the MICT group. Six AEs were classified as possibly or definitely related to the study intervention. In the HIIT group, three AEs were considered possibly related and one definitely related to the intervention. In the MICT group, two AEs were considered possibly related, one of mild severity and one of moderate severity. One serious adverse event requiring hospitalization occurred but was determined to be unrelated to the study intervention. Studyrelated AEs resulted in intervention modifications, including one dose reduction and three temporary pauses in the intervention. Missed visits were made up within the allotted 4-week exercise period.

Outcome Measures

Descriptive statistics are presented in Table 2. The assumptions of the models were not violated, and overall model fit was good. There was no significant between-group effect for the primary outcome, predicted VO₂peak (β = −0.58; SE, 0.93; p = 0.54), although both groups demonstrated significant within-group improvements.

Table 2.

Outcome Measures at Baseline and Post-Intervention

	MICT					HIIT
Outcome	n	Baselinea	Post-interventiona	Change (95% CI)b	p value	n	Baselinea	Post-interventiona	Change (95% CI)b	p-value
Primary outcome
Predicted VO2 peak, mL•kg−1• min−1	22	22.31 ± 8.69	24.34 ± 9.39	1.58 (0.18, 2.97)	0.03	25	25.14 ± 11.70	26.27 ± 12.06	1.13 (0.05, 2.21)	0.04
Secondary outcomes
Contralesional resting MCAv, cm•sec−1	22	49.01 ± 13.02	50.69 ± 10.79	1.20 (1.95, 4.35)	0.44	17	43.93 ± 12.47	44.87 ± 11.16	−0.23 (2.82, 2.35)	0.85
Ipsilesional resting MCAv, cm•sec−1	19	48.48 ± 12.42	49.01 ± 12.97	0.54 (1.87, 2.95)	0.64	16	41.31 ± 12.53	42.42 ± 11.65	−0.08 (2.35, 2.18)	0.94
Stroke-affected side flow mediated dilation, %	21	5.63 ± 3.31	6.08 ± 3.91	0.41 (0.85, 1.68)	0.50	23	4.09 ± 2.74	5.41 ± 3.95	1.43 (0.26, 2.60)	0.02
Non-affected side flow mediated dilation, %	22	5.37 ± 4.13	6.57 ± 4.84	0.89 (0.73, 2.50)	0.27	23	4.33 ± 2.64	5.85 ± 3.47	1.60 (0.07, 3.13)	0.04
cfPWV, m•sec−1	20	9.52 ± 1.88	9.57 ± 1.75	−0.05 (0.54, 0.44)	0.83	22	9.36 ± 1.96	9.40 ± 2.09	0.06 (0.40, 0.52)	0.78
Additional outcomes
Comfortable gait speed, m•sec−1	22	0.96 ± 0.36	1.09 ± 0.36	0.11 (0.04, 0.18)	0.004	25	0.90 ± 0.44	0.94 ± 0.39	0.04 (0.02, 0.10)	0.15
Fast gait speed, m•sec−1	22	1.39 ± 0.55	1.54 ± 0.57	0.11 (0.04, 0.18)	0.004	25	1.26 ± 0.61	1.38 ± 0.66	0.12 (0.07, 0.17)	<0.001
6-minute walk test, m	22	366.25 ± 163.36	406.36 ± 160.28	28.64 (10.99, 46.28)	0.003	25	306.00 ± 163.17	342.00 ± 172.12	36.00 (20.83, 51.17)	<0.001

Abbreviations: VO₂peak, peak oxygen consumption; mL•kg⁻¹•min⁻¹, milliliters per kilogram per minute; MCAv, middle cerebral artery blood velocity; cm•sec⁻¹, centimeters per second; m•sec⁻¹, meters per second; m, meters

^aBaseline and post-intervention columns show the mean (SD) for each group.

^bChange column shows mean difference (95% CI) from baseline to post-intervention.

No significant between-group effects were observed for resting MCAv (ipsilesional: β = −3.46; SE, 1.87; p = 0.07; contralesional: β = −0.47; SE, 1.78; p = 0.79), peripheral vascular function (flow-mediated dilation: stroke-affected arm β = 0.48; SE, 1.09; P = .66; non-affected arm β = 0.42; SE, 0.85; p = 0.62), or arterial stiffness (pulse wave velocity: β = 0.14; SE, 0.34; p = 0.68). Within-group analyses demonstrated significant improvements in flow-mediated dilation following HIIT (Table 2).

No significant between-group differences were observed for walking outcomes, including comfortable gait speed (β = −0.06; SE, 0.04; p = 0.19), fast gait speed (β = 0.03; SE, 0.04; p = 0.53), or walking endurance (β = 9.04; SE, 10.32; p = 0.39). Within-group analyses demonstrated improvements in fast gait speed and walking endurance in both groups, whereas comfortable gait speed improved only following MICT.

DISCUSSION

This randomized clinical trial demonstrates that HIIT can be accurately prescribed and delivered after stroke without maximal cardiopulmonary exercise testing. Using a submaximal testing approach, we achieved clear separation of exercise intensity. This intensity-guided strategy enabled vigorous training with high adherence and low adverse event rates in a clinically complex cohort. Despite verified differences in training intensity and objective confirmation of workload through heart rate and lactate responses, HIIT did not produce greater improvements in predicted VO₂peak compared to MICT over 4 weeks. These findings align with the initial 4-week timepoint of the multicenter HIT-Stroke trial, in which early between-group differences were not observed and divergence emerged only with longer training durations.⁴ Collectively, these data suggest that the early aerobic response is driven by adequate dosing, whereas the incremental benefits of HIIT may require longer exposure^1,4 and should be considered for future trials.

Prior HIIT trials have typically relied on maximal cardiopulmonary exercise testing with ECG monitoring to determine eligibility and prescribe exercise intensity. As highlighted in contemporary commentary, these requirements represent a substantial barrier to implementation in routine clinical practice.^30,31 Further, surveys of physical therapists indicate that maximal exercise testing is rarely used in stroke rehabilitation, with only approximately 2% reporting that the majority of their patients undergo stress testing prior to exercise prescription, highlighting the limited use of maximal exercise testing in clinical practice.¹¹ Therefore, by prescribing and delivering vigorous-intensity training using a submaximal testing model that does not require maximal exertion or ECG monitoring, the present study addresses this translational barrier directly. Importantly, vigorous-intensity intervals were achieved without exceeding 85% of age-predicted HRmax, demonstrating that clinically meaningful training intensity can be delivered safely using this approach. These findings support the feasibility of intensity-guided aerobic training outside specialized laboratory environments.

The magnitude of change observed in the HIIT group (+1.1 mL•kg⁻¹•min⁻¹) was nearly identical to that reported at the 4-week time point in the multisite treadmill-based HIT-Stroke trial (+1.3 mL•kg⁻¹•min⁻¹).⁴ In contrast, the improvement observed in our MICT group (+1.6 mL•kg⁻¹•min⁻¹) exceeded that reported in the treadmill trial at 4 weeks (+0.4 mL•kg⁻¹•min⁻¹), potentially reflecting higher achieved training intensity. Notably, both groups in the present study achieved improvements exceeding 1 mL•kg⁻¹•min⁻¹ within 4 weeks, a magnitude considered clinically meaningful for peak VO₂.⁸ However, evidence from longer duration trials suggests that differences between HIIT and MICT may emerge over time.^4,8 For example, Moncion et al. reported a 3.52 mL/kg/min improvement in VO₂peak following HIIT compared with 1.76 mL/kg/min following MICT at 12 weeks.⁸ These findings underscore the need for longer-duration trials to better define the optimal time course of aerobic adaptation and determine training duration required to differentiate exercise modalities after stroke.

Within the HIIT group, flow-mediated dilation improved significantly over 4 weeks, whereas no change was observed following MICT. Endothelial adaptation is a well-recognized early response to aerobic training, often preceding detectable changes in arterial stiffness or vascular structure.^32,33 The oscillatory shear stress characteristic of interval training may provide a potent endothelial stimulus, consistent with observations in other cardiovascular populations.^34,35 The magnitude of change observed (1.4–1.6%) is associated with reduced cardiovascular risk, underscoring the potential relevance of vigorous-intensity training for secondary prevention after stroke.³⁶ However, between-group differences were not detected for other cerebrovascular or peripheral vascular outcomes. Pulse wave velocity remained unchanged, consistent with evidence that structural arterial remodeling requires longer training exposure. Resting MCAv similarly did not change, consistent with prior reports suggesting that cerebrovascular adaptations may manifest preferentially in vasomotor reactivity rather than resting velocity.³⁷

Both training groups demonstrated clinically meaningful improvements in walking performance over the 4-week intervention despite not engaging in task specific treadmill training. Fast gait speed improved following HIIT and MICT. Comfortable gait speed improved following MICT, with changes meeting established minimal clinically important difference thresholds (≥0.1 m/s).^38,39 Walking endurance also improved in both groups, with the HIIT group exceeding the minimal clinically important difference for the 6-minute walk test (≥34.4 m). Although task-specific treadmill training is often emphasized in post-stroke gait rehabilitation, the present findings support the concept of training transference, whereby repetitive lower-extremity activation during cyclic exercise translates to improved overground walking performance.⁴⁰ Our data suggests that intensity-guided aerobic exercise using a total-body recumbent stepper can meaningfully influence walking outcomes. Importantly, the magnitude of improvement in walking endurance and fast gait speed observed over 4 weeks was comparable to, and in some cases exceeded, that reported in longer-duration treadmill-based⁴ and total-body recumbent stepper protocols,¹ which may reflect the high level of treatment fidelity and achievement of target training intensities in the present study.

Strengths and Limitations

This trial has several notable strengths. First, the FAST trial enrolled participants with both ischemic and hemorrhagic stroke, and women comprised 43% of the cohort, a proportion higher than that reported in many prior HIIT studies.^1,3,4,41,42 Second, randomized allocation with concealed group assignment and blinded outcome assessment minimized risk of bias. Participants were unaware of intervention allocation, reducing expectation effects that may influence performance in trials comparing HIIT and MICT. Third, treatment fidelity was exceptionally high, with near-complete attendance and clear physiologic separation of intensity confirmed through continuous heart rate monitoring (Figure 2). Fourth, systematic adverse event collection demonstrated a low rate of intervention-related events and no study-related serious adverse events despite inclusion of a clinically complex cohort spanning a broad age range.

The intervention model further strengthens the translational relevance of the findings. By using a submaximal testing paradigm, this study operationalized vigorous-intensity training during HIIT without reliance on maximal cardiopulmonary exercise testing or ECG monitoring. Both interventions exceeded typical intensity levels previously reported in stroke rehabilitation,^43–45 enhancing confidence that observed effects were attributable to adequately dosed aerobic stimulus rather than minimal clinical exposure. Participant-reported feedback supports the acceptability of this approach. Exercise enjoyment did not differ between groups, and qualitative responses (Supplemental data) indicated high engagement, increased exercise confidence, and intent to continue aerobic training after study completion. Several participants described improvements in physical steadiness and overall well-being, suggesting potential enhancement of exercise self-efficacy, although not directly assessed. Together, these behavioral signals strengthen the external validity and sustainability implications of intensity-guided aerobic training after stroke.

A primary limitation of the present study was the modest sample size, which may have limited our ability to detect between-group differences. As a preliminary efficacy trial, this study was designed to estimate effects and inform future trials, not definitive comparisons. Accordingly, the absence of statistically significant differences does not exclude the possibility that clinically meaningful divergence may emerge with larger sample sizes or longer intervention durations, as reported in other trials comparing HIIT and MICT. Additionally, we cannot fully exclude the potential influence of expectancy bias and that both groups exercised in their target HR zones with high adherence, which may contribute to the lack of significant findings.

Additional limitations include a single-site design and a short intervention duration of 4 weeks. The primary outcome relied on predicted rather than directly measured VO₂peak, although the submaximal testing approach was intentionally selected to enhance clinical scalability and has demonstrated validity.^13,15 Physical activity outside supervised sessions was not objectively monitored, and long-term maintenance of training effects was not assessed. Finally, findings are limited to individuals with chronic stroke and may not generalize to earlier recovery phases.

Conclusion

In this randomized clinical trial involving individuals with chronic stroke, vigorous-intensity HIIT delivered using a submaximal prescription model did not produce greater improvements in predicted VO₂peak than MICT over 4 weeks. However, both HIIT and MICT resulted in clinically meaningful improvements in cardiorespiratory fitness and walking performance. HIIT additionally improved endothelial function, supporting the biologic plausibility of intensity-dependent vascular adaptation. Importantly, the ability to prescribe and deliver vigorous-intensity training without maximal cardiopulmonary exercise testing supports the potential for broader integration of intensity-guided aerobic training into routine stroke rehabilitation.

Key Points.

Question:

In individuals with chronic stroke, does high-intensity interval training (HIIT) improve predicted VO₂peak more than moderate-intensity continuous training (MICT)?

Findings:

In this randomized clinical trial of 49 participants with chronic stroke, both HIIT and MICT achieved prescribed intensity targets with high adherence and resulted in clinically meaningful improvements in predicted VO₂peak and walking outcomes after 4 weeks, with no significant between-group difference in our primary outcome of VO₂peak.

Meaning:

These findings suggest that when aerobic exercise is prescribed to achieve target intensity, both HIIT and MICT produce meaningful improvements in fitness and walking after stroke, supporting the importance of appropriate exercise dosing.

Funding Statement

SAB, TB, AE, and EDV were supported in part by P30 AG072973. BLB was supported in part by F31HL182212 and T32HD057850. REDCap at the University of Kansas Medical Center was supported by the National Center for Research Resources UL1TR002366

Data access, responsibility, and analysis:

SAB had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.