Effects of high-intensity interval training versus moderate-intensity continuous training on cardiorespiratory function in patients after stroke: a systematic review and meta-analysis of randomized trials

TL;DR

This study compares high-intensity interval training and moderate-intensity continuous training in stroke patients, finding that high-intensity training improves cardiorespiratory fitness more effectively.

Contribution

The study provides the first meta-analysis comparing HIIT and MICT effects on cardiorespiratory outcomes in post-stroke patients.

Findings

HIIT improved V̇O2-peak and VO2-VT more than MICT in stroke patients.

HIIT did not enhance mobility outcomes like walking endurance or balance more than MICT.

HIIT resulted in higher heart rate during training sessions compared to MICT.

Abstract

Whether high-intensity interval training (HIIT) is more effective than moderate-intensity continuous training (MICT) in improving cardiorespiratory fitness (CRF) among patients after stroke remains unclear. We conducted this systematic review and meta-analysis to investigate the effects of HIIT versus MICT on CRF.

We performed a literature search in the PubMed, Embase, and Cochrane Library from their earliest publication record to February 2025. Randomized trials comparing the outcomes of HIIT and MICT in patients after stroke were included. The mean difference (MD) and standardized mean difference (SMD) were determined by pooling the means and standard deviations of pretreatment–posttreatment changes for the CRF outcomes [i.e., oxygen consumption at peak (V̇O2-peak) and at ventilation threshold (VO2-VT)], mobility outcomes (i.e., walk endurance, speed, and postural balance) and training fidelity parameters (i.e., peak and mean heart rate during training sessions).

Nine articles, encompassing eight trials and a total of 371 patients, were included in the analysis. The meta-analysis revealed that HIIT was more effective in improving V̇O2-peak (MD = 1.88 mL/kg/min, 95% CI: 1.20 to 2.55, p < 0.05) and VO2-VT (MD = 2.20 mL/kg/min, 95% CI: 0.46 to 3.95, p < 0.05). However, HIIT did not show greater effectiveness in improving the 6-min walk test, 10-meter gait speed, or Berg Balance Score. Regarding training fidelity, a significantly higher mean heart rate [measured as a percentage of heart rate reserve (HRR, %)] was observed in HIIT sessions (MD = 19.36% HRR, 95% CI: 13.83 to 24.90, p < 0.05).

HIIT is more effective than MICT in improving V̇O2-peak and VO2-VT in patients after stroke, supporting HIIT may serve as an alternative for aerobic training in this population.

https://www.crd.york.ac.uk/PROSPERO/view/CRD42025645342, CRD42025645342.

Figures

Flowchart of study identification and selection.

Forest plot presenting a comparison of changes in (A) VO2-peak and (B) VO2-VT.

Forest plot presenting a comparison of changes in motor function: (A) 6-min walk test; (B) 10-meter gait speed; (C) Berg Balance Score.

Forest plot presenting a comparison of changes in training fidelity: (A) Mean heart rate and (B) Peak heart rate.

Tables

Table 1. Characteristics of included studies.

Author year; country	Selection criteria	Number of patients(% male)	Age, year,mean ± SD	Number of patients with ischemic stroke (%)	Intervention					PEDro scale
					Modality	Intensity	Duration(min/sess)	Frequency(sess/wk)	Intervention period(wks)
Boyne et al. (2016) (35); USA	Ambulatory, poststroke duration > 6 months	H: 11 (63.6)M: 5 (40)	H: 59 ± 9M: 57 ± 12	H: 9 (81.8)M: 2 (40)	H: Treadmill exerciseM: Treadmill exercise	H: 30-s bursts of HI walking, alternated with 30- to 60-s LI periods (treadmill stopped). Average session HR progressed from a mean 53% HRR in 1st week to 72% HRR in 4th week.M: Walking with speed adjusted to maintain 45% HRR, then progressed to 50% HRR after 2 wks of training.	H: 25M: 25	H: 3M: 3	H: 4M: 4	8/10
Boyne et al. (2023) (34); USA	Patients after a single stroke	H: 27 (59.3)M: 28 (71.4)	H: 63.8 ± 9.9M: 61.5 ± 9.9	H: 15 (55.6)M: 19 (67.9)	H: Walking practice and treadmill exerciseM: Walking practice and treadmill exercise	H: 30-s bursts of HI walking, alternated with 30- to 60-s LI periods; targeting a mean aerobic intensity above 60% of the HRR.M: walking with speed adjusted to maintain a target HR of 40% of the HRR, progressing by 5% of the HRR every 2 weeks up to 60% of the HRR.	H: 45M: 45	H: 3M: 3	H:12M: 12	8/10
Hsu et al. (2021) (36); Taiwan	Poststroke duration > 3 months	H: 10 (80)M: 13 (92.3)	H: 58.5 ± 8.7M: 53.1 ± 6.9	H: 6 (60)M: 9 (69.2)	H: Bicycle ergometer M: Bicycle ergometer	H: Five 3-min intervals at 80% VO2 peak, each interval separated by 3 min of exercise at 40% of VO2 peak, for 30 minM: 60% VO2 peak for 30 min	H: 30M: 30	H: 2–3M: 2–3	Totally 36 sessions	6/10
Lapointe et al. (2023) (37);Canada	Ischemic stroke, poststroke duration > 3 months	H: 19 (68.4)M: 16 (62.5)C: 17 (58.8)	H: 71.8 ± 9.9M: 65.6 ± 11.3\u00B0C: 69.6 ± 10.7	H: 19 (100)M: 16 (100)C: 17 (100)	H: Ergocycle training + home exercise.M: Ergocycle training + home exercise.C: Usual care without any additional physical activity.	H: Bouts at 95% of peak power output interspersed with a 60-s recoveryM: 50% of peak power output(Supervised training)	H: 20–40M: 20–40	H: 3 sess/wk. of HIIT for 2 months, then 2 sess/wk. of HIIT + 1 sess/wk. of unsupervised MICT for the next 2 months, and then 1 sess/wk. of HIIT + 2 sess/wk. of unsupervised MICT for the next 2 months.M: 1 sess/wk. supervised MICT + 2 sess/wk. of unsupervised MICT for 6 months.		4/10
Mahrukh et al. (2023) (38); Pakistan	Poststroke duration > 6 months	H: 30 (NI)M: 30 (NI)	Patients aged 50–80 years old were included.	NI	NI	NI	NI	NI	H: 12M: 12	4/10
Marzolini et al. (2023) (39); Canada	Poststroke duration > 10 weeks	H: 24 (83.3)M: 23 (78.3)	H: 62.8 ± 13.2M: 60.9 ± 8.4	H: 16 (66.7)M: 20 (87.0)	H: Treadmill exerciseM: Treadmill exercise	H: Alternative HI and LI intervals. HI: RPE of ≥17, LI: RPE of 10–12.M: HR-VT, RPE of 12–16, or 60–80% of HRR	H: 20–22M: ≤60	H: 3 HIIT + 2 MICTM: 5 MICT	H: 24M: 24	8/10
Moncion et al. (2024) (25) and Rodrigues et al. (41); Canada	Poststroke duration between 6 to 60 months	H: 42 (64.3)M: 40 (57.5)	H: 65.4 ± 8.9M: 64.4 ± 9.7	H: 30 (71.4)M: 33 (82.5)	H: Adaptive recumbent steppersM: Adaptive recumbent steppers	H: 10 × 1-min bouts of HI exercise, interspersed with 9 × 1-min LI intervals, for 19 min. HI intervals targeted 80% HRR and progressed by 10% every 4 weeks up to 100% HRR. LI intervals targeted 30% HRR.M: Targeted 40% HRR for 20 min and progressed by 10% HRR and 5 min every 4 weeks, up to 60% HRR for 30 min	H: 19M: 20–30	H: 3M: 3	H: 12M: 12	6/10
Soh et al. (2020) (42); South Korea	Patients after minor stroke	H: 18 (72.2)M: 18 (66.7)	H: 56.3 ± 5.3 M: 57.4 ± 7.2	H: 13 (72.2)M: 12 (66.7)	H: Lateral push-off skater exerciseM: Conventional treadmill aerobic exercise	H: NI for intensity regarding HI and LI intervals. RPE ≤ 14 for a HIIT session was reported.M: RPE ≤ 14 and 40–80% of HRR	H: 30M: 30	H: 3M: 3	H: 12M: 12	5/10

Table 2. Methodological quality assessment of the selected randomized controlled trials (RoB 2.0).

Study	Bias from the randomization process	Bias caused by deviations from intended interventions	Bias caused by missing outcome data	Bias in measurement of the outcome	Bias in selection of the reported results	Overall risk of bias
Boyne et al. (35)	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Boyne et al. (34)	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Hsu et al. (36)	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Lapointe et al. (37)	Low risk	Low risk	Low risk	Moderate riska	Low risk	Moderate risk
Mahrukh et al. (38)	Moderate riskb	Moderate riskc	Low risk	Moderate riskd	Low risk	High risk
Marzolini, et al. (39)	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Moncion et al. (25) and Rodrigues et al. (41)	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Soh et al. (42)	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk

Full text

Introduction

1

Stroke is the leading cause of disability and the third most common cause of death globally, affecting approximately 93.8 million people worldwide, and causing about 160.5 million disability-adjusted life-years lost in 2021 (1). Patients after chronic stroke often face motor impairments, leading to mobility challenges such as walking and balance difficulties (1). Reduced physical activity can further impact cardiorespiratory fitness (CRF), which is defined as the capacity of the circulatory and respiratory systems to supply oxygen to skeletal muscle mitochondria for energy production required during physical activity (2). Patients after stroke typically demonstrate low values of peak oxygen uptake (V̇O_2-peak_) (3), a key indicator of CRF. Therefore, improving CRF by implementing aerobic training in this population is critical in stroke rehabilitation (4).

Currently, moderate-intensity continuous training (MICT), typically targeting a heart rate of 40–80% of heart rate reserve (HRR), is the most commonly used aerobic training program and is recommended in stroke rehabilitation guidelines (5–7). The benefits of MICT in improving V̇O_2-peak_, motor function, and cardiovascular risk factors (e.g., blood pressure and glucose levels) are well established (8, 9). However, evidence suggests that the therapeutic effects of aerobic training may be associated with training intensity (10). The intensity of MICT may not sufficiently challenge the cardiovascular system to elicit maximal adaptations, nor meet the threshold required to complete many activities of daily living (10, 11).

Previous studies have proposed the use of high-intensity training (>60% of HRR) could augment outcomes, although such high intensity exercise can be challenging for patients after stroke (12, 13). Therefore, high-intensity interval training (HIIT), a modality that maximizes exercise intensity by alternating bursts of high intensity effort with recovery periods to enable higher sustained intensities at lower perceived exertion than high-intensity continuous exercise, has gained popularity in recent years (14–16).

Several meta-analyses have shown that HIIT leads to significantly better outcomes than MICT in healthy individuals and in patients with chronic diseases such as cancer, obesity, coronary artery disease, and heart failure (17–23). However, few clinical trials have compared the effects of HIIT and MICT in patients after stroke, leaving its therapeutic efficacy unclear. To date, only two meta-analyses have explored the effects of HIIT among patients after stroke. Both of them primarily compared HIIT with control group (usual care) or low-intensity continuous training, which lead to obviously better results (24, 25). With emerging clinical trials that explores the difference between HIIT and MICT in stroke population, we conducted this review study to provide further insight for clinical practice.

In the present study, we aimed to explore the effects of HIIT versus MICT on improving CRF, functional performance in mobility, and the differences in training fidelities between HIIT and MICT.

Methods

2

This systematic review and meta-analysis was registered in the International Prospective Register of Systematic Reviews of the UK National Institute for Health Research (PROSPERO; ID: CRD42025645342), and was performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (50).

Searching strategy

2.1

We searched the PubMed, Embase, and Cochrane Library for studies published from inception to February 2025 using the following search terms: ([high-intensity interval training OR HIIT] OR [moderate-intensity continuous training OR MICT] OR aerobic training) AND (stroke OR cerebrovascular accident OR cerebrovascular disorder OR cerebral infarction OR brain infarction OR intracranial arteriosclerosis OR intracranial thrombosis OR intracranial embolism OR CVA). The detailed searching strategy is presented in the Supplementary Data 1. Moreover, all retrieved abstracts, studies, and citations were reviewed. No language restrictions were applied.

Study inclusion and exclusion criteria

2.2

We included randomized trials on the basis of their titles and abstracts in accordance with the following selection criteria: (1) compared the outcomes of HIIT with MICT in patients after stroke, and (2) reported the inclusion and exclusion criteria for patient selection. We excluded trials that (1) were observational or nonrandomized trials, (2) primarily used intervention other than HIIT (ie, high intensity continuous training), (3) compared HIIT with non-MICT intervention (ie, low intensity exercise or usual care), and (4) had not been published as full-text articles in a peer-reviewed journal.

Data extraction

2.3

Two reviewers independently extracted baseline and outcome data, including the number, age, and sex of the participants; inclusion and exclusion criteria; number of patients after ischemic stroke; intervention regimens, frequency, and duration; and outcome parameters. Individually recorded data from two reviewers were compared, and any disagreements were resolved by a third reviewer. In case of missing data or interesting data that were not reported in the article, the corresponding authors of the original study for information were contacted through e-mail.

Outcomes

2.4

The primary outcome was the CRF indicators, including V̇O_2-peak_ (mL/kg/min) or VO_2_ at ventilation threshold (VO_2-VT_, mL/kg/min) measured by cardiopulmonary exercise test. Previous studies have determined the minimum detectable change for V̇O_2-peak_ in patients after stroke to be 1 mL/kg/min (26, 27). The secondary outcomes were the functional performance in mobility, involving walking speed (10-meter walk speed), walking endurance (6-min walk test), and postural balance (Berg Balance Score). Resting blood pressure, a cardiovascular risk indicator, was surveyed. In addition, parameters of training fidelity representing the intensity the patient sustained during a training session were included.

Study quality assessment

2.5

The methodological quality of the randomized trials was assessed independently by two reviewers, in accordance with the revised Cochrane Risk of Bias (RoB 2.0) tool, PEDro scale, and the Tool for the Assessment of Study Quality and Reporting in Exercise (TESTEX).

RoB 2.0 includes the following domain: bias deriving from the randomization process, bias caused by deviations from intended interventions, bias caused by missing outcome data, bias arising from outcome measurement, bias deriving from selection of the reported results, and overall risk of bias (28). Each domain was rated as having low risk of bias, some concerns, or high risk of bias. The PEDro scale assesses the quality of randomized controlled trials in physiotherapy and rehabilitation, focusing on validity, statistics, and design. It includes 11 criteria, with scores below 4 rated as poor, 4–5 as fair, 6–8 as good, and 9–10 as excellent (29). The TESTEX scale is a 15-point scale that is designed specifically for use in exercise training studies, with scores below 6 rated as ‘low quality’, 7–11 as ‘good quality’, and 12–15 as ‘high quality’ (30). After individual assessments, the two reviewers discussed any potential discrepancies, which were subsequently resolved by a third reviewer.

Data synthesis and presentation

2.6

Data were analyzed by Review Manager, Version 5.3 (Cochrane Collaboration, Oxford, England). Standard deviations were estimated from the provided confidence interval (CI) limits or standard errors. For the primary and secondary outcomes, the means and standard deviations of pretreatment–posttreatment changes were used in the meta-analysis. When necessary, these values were estimated in accordance with the reported pretreatment and posttreatment data (31). Continuous outcomes were analyzed using the mean difference (MD) or the standard mean difference (SMD). The precision of the effect sizes was reported as 95% CIs. We used the DerSimonian and Laird random-effects model to compute a pooled estimate of the MD or the SMD (32). The inverse variance method was used to analyze continuous variables. The I^2^ test was used to quantify the heterogeneity of the outcomes. Heterogeneity was classified as small, moderate, and large for I^2^ values of 25, 50, and 75%, respectively. Results were reported only when 2 or more studies were available for meta-analysis on the same outcome.

Assessment of quality of evidence

2.7

We used the Grading of Recommendation Assessment, Development, and Evaluation (GRADE) approach to access the quality of evidence (33). Rating aspects include (1) study limitations, (2) inconsistency, (3) indirectness, (4) imprecision, and (5) publication bias. We consequently graded the quality of evidence for each outcome as high, moderate, low, or very low quality.

Results

3

Search results and study characteristics

3.1

A flowchart describing the screening and selection process is presented in Figure 1. Nine articles involving eight different trials were included in the systematic review and meta-analysis (34–42).

Flowchart of study identification and selection.

The characteristics of the included trials are summarized in Table 1. Two articles reported different aspects of the same trial; one concerns CRF benefits (40) and the other concerns psychosocial responses (41). Three trials were conducted in Canada (37, 39–41), two in The United States of America (34, 35), one in Pakistan (38), one in South Korea (42), and one in Taiwan (36). Two articles reported different aspects of the same trial; one concerns CRF benefits and the other concerns psychosocial responses after training (40, 41). A total of 371 patients aged from 50 to 80 were included in the meta-analysis.

In the training modality, three trials used treadmill exercise as their HIIT program (34, 35, 39), two trials used bicycle ergometer (36, 37), one trial used adaptive recumbent steppers (40, 41), one trial used lateral push-off skater exercise (42), and the other trial did not provide information regarding training modality (38). The intensity of high-intensity intervals was measured by heart rate (60–100% of HRR) in three trials (34, 35, 40, 41), VO_2_ (80–100% of V̇O_2-peak_) in two trials (36, 39), and power output in one trial (37). Two trials did not provide detailed information on defining the intensity of interventions (38, 42), though we have contacted the corresponding authors for more detailed information. Training duration and frequency varied from two to five sessions per week (34–37, 39–42); and the duration of intervention lasted for 4 weeks in one trial (35), 12 weeks in five trials (34, 36, 38, 40–42), and 24 weeks in the other two trials (37, 39).

Methodological quality

3.2

The assessment of ROB 2.0 of the included trials is summarized in Table 2 and Supplementary Figure 1. Overall, one trial was rated as having high risk of bias (38), one trial was rated as having moderate risk of bias (37), and the remaining six trials were rated as having low risk of bias (34–36, 39–42).

The assessment of PEDro scale of the included trials is summarized in Table 1 and Supplementary Table 1. The PEDro scores ranged from four to eight. Most of the trials were downgraded due to lack of subject blinding and therapist blinding. Overall, three trials were rated as having fair methodological qualities (37, 38, 42), and the remaining six trials were rated as having good methodological qualities (34–36, 39–41).

The assessment of the TESTEX scale of the included trials is summarized in Supplementary Table 2. The TESTEX scores ranged from 10 to 15. Overall, two trials were rated as having good methodological qualities (37, 38), and the remaining six trials were rated as having high methodological qualities (34–36, 39–42).

The assessment results of the three assessment tools are generally consistent with each other.

Effects on CRF outcomes

3.3

Effects on V̇O_2-peak_ and VO_2-VT_ were shown in Figure 2. The pooled results from all included studies (34–42) showed that HIIT exhibited a significantly greater improvement in V̇O_2-peak_ than MICT (MD = 1.88 mL/kg/min, 95% CI = 1.20 to 2.55, p < 0.00001; Figure 2A). A sensitivity analysis excluding 2 studies (37, 38) with moderate and high risk showed a similar result (MD = 2.02 mL/kg/min, 95% CI = 1.27 to 2.77, p < 0.00001). In addition, the pooled results from three included studies (34, 35, 39) showed that HIIT exhibited a significantly greater improvement in VO_2-VT_ than MICT (MD = 2.20 mL/kg/min, 95% CI = 0.46 to 3.95, p = 0.01; Figure 2B).

Forest plot presenting a comparison of changes in (A) VO2-peak and (B) VO2-VT.

Effects on functional performance in mobility

3.4

The pooled results revealed that there was no significant difference between HIIT and MICT in change in 6-min walk test (MD = 17.63 m, 95% CI = −1.44 to 36.70, p = 0.07; Figure 3A) (34, 35, 38–40), the change in 10-meter gait speed (MD = 0.08 m/s, 95% CI = −0.01 to 0.17, p = 0.08; Figure 3B) (34, 35, 39, 40), or the change in Berg Balance Score (MD = 2.57, 95% CI = −4.53 to 9.68, p = 0.48; Figure 3C) (39, 42).

Forest plot presenting a comparison of changes in motor function: (A) 6-min walk test; (B) 10-meter gait speed; (C) Berg Balance Score.

Training fidelity regarding the intensity

3.5

The pooled results from three studies (34, 35, 39) showed that patients undergoing HIIT sustained greater mean heart rate than MICT during the training session (MD = 19.36% of HRR, 95% CI = 13.83 to 24.90, p < 0.00001; Figure 4A). In addition, patients undergoing HIIT also sustained greater peak heart rate than MICT during the training session (SMD = 1.00, 95% CI = 0.40 to 1.59, p = 0.0010; Figure 4B) (34, 39, 42).

Forest plot presenting a comparison of changes in training fidelity: (A) Mean heart rate and (B) Peak heart rate.

Effects on resting blood pressure

3.6

During the intervention period, no significant differences were observed between HIIT and MICT in systolic blood pressure changes (MD = −1.72 mmHg, 95% CI = −5.92 to 2.48, p = 0.42; Supplementary Figure 2A) (37, 39, 40, 42) and diastolic blood pressure (MD = −1.04 mmHg, 95% CI = −3.78 to 1.69, p = 0.48; Supplementary Figure 2B) (37, 39, 40, 42).

Adverse effects

3.7

No adverse events of HIIT were observed in two of the included studies (37, 42). In another four studies, no study-related serious adverse events occurred, and the number of adverse effects was similar between the HIIT group and the MICT group. Common adverse events included pain, soreness, fatigue, lightheadedness, and fall (34, 35, 39, 40). No information on adverse effects was provided in the other two studies (36, 38).

Quality of evidence

3.8

The summary of findings and the grading of the evidence certainty are presented in Supplementary Table 3. The GRADE certainty of the evidence of the above outcomes were moderate or low because of inconsistency [unexplained heterogeneity was detected in the pooled result (I^2^ > 50%)] and imprecision (the number of studies and patients are small).

Discussion

4

To the best of our knowledge, this is the first systematic review to directly compare the effects of HIIT and MICT in patients after stroke. Overall, HIIT generated a 1.88 mL/kg/min improvement in V̇O_2-peak_ compared to MICT, which exceeded the reported minimum detectable change (26, 27) (1 mL/kg/min) in patients after stroke. Given that every 3.5 mL/kg/min (equal to 1 metabolic equivalent) increase in V̇O2-peak was associated with a 15 to 20% decrease in overall mortality among various populations (51), a difference of 1.88 mL/kg/min of V̇O_2-peak_ was clinically meaningful. Considering that MICT remained the most commonly used training mode in stroke rehabilitation, our results support the suggestion that HIIT may be also used as an alternative for aerobic training in stroke population.

Previous studies have shown that acute bout of HIIT activates signaling pathways involving mitochondrial biogenesis, and repeated activation of these pathways may lead to specific muscular adaptations, including increased mitochondrial density, capacity and respiration in skeletal muscle (43–45), which is related to improvements in V̇O2-peak. In addition, previous studies have shown that HIIT can lead to better improvements in cardiac contractility, systolic function, and endothelial function in patients with cardiovascular diseases when compared to MICT (21, 46). Moreover, one included study in our meta-analysis also compared the physiological effects of HIIT to those of MICT on patients after stroke, and the results showed that HIIT significantly improved cardiac output (mean improvement: 1.45 L/min, p = 0.038) and serum brain-derived neurotrophic factor level (mean improvement: 1.85 ng/mL, p = 0.012). Improvement in aerobic capacity by increasing systemic tissue oxygen extraction, and increased cerebral oxygen utilization in the involved hemisphere was also observed after HIIT when compared with MICT (36). All these mechanisms discussed above may lead to better improvements in CRF in the HIIT group when compared with the MICT group, which is the primary result in this meta-analysis.

However, our results revealed that HIIT did not result in superior functional mobility improvements, indicating that HIIT is not superior to MICT in those variables. This may be attributed to the relatively small number of included studies assessing these outcomes and the high heterogeneity observed (e.g., I^2^ = 65% for the 6-min walk test and 94% for the Berg Balance Score), suggesting variability in patient characteristics and intervention protocols. While CRF plays a crucial role in mobility, stroke-related impairments such as neuromuscular dysfunction, reduced motor coordination, muscle weakness, proprioceptive deficits, and impaired balance may limit the direct translation of improved aerobic capacity into functional mobility gains (47). Further research is needed to better understand the effects of HIIT on functional mobility in patients after stroke and to identify factors that may influence its efficacy.

Two recent review studies have addressed the effects of HIIT in the stroke population (24, 25). Although Anjos et al. supported the superiority of HIIT over continuous aerobic training, their conclusion was based on only 4 trials with 91 patients. Moreover, their continuous aerobic training group was mixed with low-intensity walking, which caused bias in interpreting the true difference between HIIT and MICT (24). On the other hand, Moncion et al. conducted a systematic review on the effects of various aerobic exercise interventions (i.e., HIIT, high-intensity continuous training, MICT, low-intensity continuous care, usual care) (25). Their network meta-analysis showed no significant mean difference between HIIT and MICT (i.e., MD = 0.82 mL/kg/min, 95%CI: −0.92 to 2.56). It should be noted that their result was based on only 4 trials involving direct comparison between HIIT and MICT. Additionally, the methodological bias arising from indirect comparison in network meta-analysis was also concerned (48). In contrast, we updated the knowledge with 4 new trials and provided more solid evidence regarding the comparison between HIIT and MICT.

Although our results supported the superiority of HIIT over MICT in improving V̇O2-peak, the underlying mechanism remained unclear. Our meta-analysis showed that both peak and mean heart rate was higher during HIIT sessions than during MICT sessions (Figure 4). This finding may imply that the HIIT session provided higher averaged training intensity than the MICT did. Since it has been well established that the effect of aerobic training on CRF is associated with training intensity (52), a higher averaged training intensity may contribute to the superiority of HIIT. This notion should be confirmed with more evidence involving the comparison regarding the training fidelity between HIIT and MICT.

Training intensity (i.e., high or moderate) and duration (i.e., interval or continuous or time spent on a session) are two controlling factors of training volume. It is rational to compare two interventions at different intensity but with same duration, or with different duration but at same intensity. However, comparing two training modes that differ in both intensity and duration leads to methodological challenges. At least one included study controlled for training volume and still found evidence favoring HIIT. In the study by Hsu et al., the HIIT protocol consisted of alternating intervals at 40 and 80% of V̇O_2-peak_, while the MICT protocol was performed at a steady intensity of 60% V̇O_2-peak_, with a matched total training duration (36). This isoenergetic expenditure design minimized the confounding effect of training volume. We advocate for future trials to control training volume (i.e., the integral of training intensity and duration) when comparing HIIT and MICT. Doing so will provide clearer insights into the underlying factors contributing to HIIT’s potential superiority.

Limitations

5

Our study has several limitations. First, the number of included randomized trials and the sample sizes were relatively small, potentially affecting result precision. Larger randomized trials should be conducted to compare the effects of HIIT and MICT in patients after stroke. Moreover, variations in cardiopulmonary exercise test protocols, exercise intervention regimens, frequency, and duration among trials may cause considerable heterogeneity. In addition, previous article has stated that V̇O_2-peak_ may not precisely represent CRF, especially in clinical population (49). However, only two CRF indicators (V̇O_2-peak_ and VO_2-VT_) were available for data pooling. Future studies may investigate the effects of HICT on different CRF indicators. Finally, only the changes between pretreatment and posttreatment data were analyzed; therefore, leaving the long-term sustainability of improvement unclear.

Conclusion

6

In conclusion, our results indicate that no difference was detected between HITT and MICT for functional mobility. However, HIIT significantly improves CRF indicators compared to MICT in patients after stroke (moderate certainty evidence on GRADE), supporting HIIT as a preferred option over MICT for enhancing cardiorespiratory fitness in this population.