The literature search was conducted across the PubMed, Web of Science, the Cochrane Library, Embase and SPORTDiscus databases using the following keywords: “Blood flow Restriction,” “Kaatsu,” “Ischemic Training,” “BFRT Therapy,” “BFRT Therapies,” “BFRT,” “Blood Flow Restriction Training,” “Blood Flow Restriction Exercise,” “resistance training,” “resistance exercise,” “Resistance,” “Strength Training,” “Strengthening Programs,” “Weight Bearing Exercise Program,” “players,” “sportsman,” “athlete,” “sports person,” and “sportswomen.” The search timeframe covered all publications from the inception of each database until June 29, 2025. The detailed search strategy for each database is provided in Supplementary Appendix 1.

To ensure the accuracy of the literature search, two researchers independently verified the search terms. In cases where there was disagreement between the two researchers (B.D and R. Y) regarding the selection of search terms, a third researcher (G.L) made the final decision.

This meta-analysis was conducted in accordance with the guidelines of the Cochrane Collaboration (Cumpston et al., 2019) and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Moher et al., 2009). The study protocol was registered with the International Platform of Registered Systematic Review and Meta-Analysis Protocols (INPLASY202480005). The Population, Intervention, Comparator, Outcomes, and Study Design (PICOS) criteria were used to define the inclusion and exclusion criteria for this meta-analysis, as outlined in Table 1.

The retrieved records from each database were imported into EndNote X9 software for duplicate removal. Two researchers independently conducted an initial screening of the titles and abstracts to identify potentially relevant studies. Following this, both researchers thoroughly reviewed the full texts and assessed the studies based on the population, intervention, comparator, outcomes, and study design criteria to determine their eligibility according to the inclusion and exclusion standards. The study selection process was conducted independently by two researchers. In cases of disagreement, a third researcher (G.L) was consulted, and any conflicts were resolved through discussion until a consensus was reached. Inter-rater agreement for study eligibility assessment based on full-text review was substantial (Almost Perfect, Cohen’s kappa = 0.85).

Upon completing the search, detailed information was collected from the eligible articles. Two researchers (B.D and R. Y) independently extracted the data into a Microsoft Excel spreadsheet. The extracted data included the article title, publication year, author names, characteristics of the subjects (age, sample size of the experimental and control groups), training protocols (intervention duration, intervention methods, training frequency, training load, BFR-RT duration, cuff location, width, pressure, and intervention methods in both the experimental and control groups), outcome measures, and authors’ conclusions. If discrepancies arose in the data extracted by the two researchers, a third researcher (G.L) would extract the data and confirm the final version. Pre- and post-intervention data were recorded as mean ± standard deviation in the Excel spreadsheet, If the full text or specific data from the study were not accessible, the corresponding author was contacted to obtain the relevant information. Inter-rater reliability for study selection was calculated using Cohen’s kappa statistic (Almost Perfect, Cohen’s kappa = 0.83).

The quality of all included studies was assessed using the Cochrane Risk of Bias Tool. The evaluation criteria included “Low risk,” “Unclear risk,” and “High risk,” and the final results served as the basis for assessing the quality of the included studies. The quality assessment for all studies included in this meta-analysis was independently conducted by two researchers (B.D and R. Y). Inter-rater reliability for the Risk of Bias Assessment was also calculated using Cohen’s kappa statistic (Substantial, Cohen’s Kappa = 0.72).

All statistical analyses were performed using StataMP 17.0. As all variables were continuous, outcomes were reported as means and standard deviations. Standardized mean differences (SMDs) were calculated to estimate effect sizes, and forest plots were generated accordingly. A random-effects model was applied to compute pooled effect sizes across studies, regardless of heterogeneity, to account for potential clinical and methodological differences among included studies (Borenstein et al., 2021; Chandler et al., 2019). Heterogeneity was assessed using the I² statistic, with values interpreted as follows: <25% indicating low heterogeneity, 25%–75% moderate, and >75% substantial heterogeneity (Van Tulder et al., 2003). When I² exceeded 75%, sensitivity analyses were conducted to assess the robustness of the pooled results (Higgins et al., 2003). The magnitude of effect sizes was interpreted according to the scale proposed by Hopkins et al. 2009: <0.2 = trivial; 0.2–0.6 = small; 0.6–1.2 = moderate; 1.2–2.0 = large; 2.0–4.0 = very large; and >4.0 = extremely large. To assess potential publication bias, a funnel plot was constructed to visually inspect asymmetry in the distribution of study effect sizes. In addition, Egger’s regression test was performed to statistically evaluate the presence of publication bias (Sedgwick and Marston, 2015). A p-value of <0.05 was considered statistically significant.

A total of 181 articles were retrieved from the databases, and an additional four articles were identified through manual search of the references in related studies. After removing duplicates, an initial screening was conducted based on the titles and abstracts. Full texts were then reviewed and re-screened, excluding studies that did not meet the inclusion criteria. Ultimately, 15 studies were included in the meta-analysis. The detailed process is illustrated in Figure 1.

A total of 15 studies were included in this meta-analysis, encompassing 372 participants, all of whom were athletes. The experimental group, which underwent BFR-RT, included 185 participants, while the control group, which performed traditional RT, included 205 participants. The majority of the interventions were conducted at a frequency of 2–3 sessions per week, with intervention durations ranging from 3 to 10 weeks. The types of athletes included in the studies were volleyball players (Wang et al., 2022), long jumpers (Hopkins et al. 2009), weightlifters (Van Tulder et al., 2003), rugby players (Jessee et al., 2018; Geng et al., 2024; Vergara et al., 2024; Adhitya et al., 2025), football players (Wang et al., 2022; Garber et al., 2011; Sarfabadi et al., 2023; Luebbers et al., 2019; Kami̇Ş et al., 2024), netball athletes (Manimmanakorn et al., 2013), trampoline athletes (Yang et al., 2022), Basketball players (Adhitya et al., 2025; Smith H. et al., 2025), and canoeists (Ugur et al., 2023). Detailed characteristics of the included studies are presented in Table 2.

The quality assessment of the 15 included studies was conducted based on the Cochrane Risk of Bias Tool criteria, and the results are presented in Figure 2. All the included studies were randomized controlled trials. Regarding allocation concealment, 13 studies were assessed as having unclear risk, while only 2 studies were considered to have a low risk. Due to the nature of the interventions, all studies were rated as high risk for blinding of participants and personnel, as blinding was not feasible. For blinding of outcome assessment, 10 studies were assessed as having unclear risk, and 5 studies were rated as low risk. In terms of incomplete outcome data, 2 studies were assessed as having unclear risk, and 13 studies were rated as low risk. Regarding selective reporting, 2 studies were assessed as having unclear risk, while the remaining 13 studies were considered to have low risk. Regarding other biases, 1 study was assessed as having unclear risk, while the remaining studies were considered to have low risk.

Funnel plots for all outcomes showed approximate symmetry, and Egger’s regression tests revealed no significant publication bias, suggesting a low risk of publication bias (Figure 7).

The meta-analysis results indicate that BFR-RT is not an effective method for enhancing lower limb muscle hypertrophy in athletes when compared to RT alone. This finding is consistent with prior systematic reviews by de Ruiter et al. (2007) and Sadamoto et al. (1983), as well as many of the studies included in the present analysis (Scott et al., 2017; Yang et al., 2022; Ugur et al., 2023; Luebbers et al., 2014; Yamanaka et al., 2012), all of which reported similar results.

Individual studies agree. Laurentino et al. (2008) showed that applying an occlusion cuff during 8 weeks of heavy leg extensions did not augment quadriceps CSA gains relative to training without cuffs. Likewise, Dankel et al. (2016) found that blood flow restriction added to high-load biceps curls produced no greater muscle activation or hypertrophy than high-load exercise alone–concluding that BFR “would seem unlikely to induce greater muscle hypertrophy” under those conditions pubmed.ncbi.nlm.nih.gov. Teixeira et al. (2021) reported similar results: after 8 weeks of unilateral knee extensions, all protocols (high-load with BFR during rest, with BFR during lifting, or with no BFR) showed ∼6–7% quadriceps CSA increases with no between-group differencesjournals.lww.com. In sum, adding BFR to an already intense program did not confer any extra muscle size gain.

A likely reason is redundancy of the hypertrophic stimulus. Maximal contractions themselves generate very high intramuscular pressure, effectively occluding blood flow and causing the same hypoxic, metabolite‐rich environment that BFR aims to induce. In other words, heavy lifting “per se” creates near-occlusion in the muscle (Teixeira et al., 2021), so supplemental cuff pressure adds little new stress. Moreover, the once‐popular hormonal hypothesis has been largely disproven: acute surges of GH, IGF-1, testosterone, etc. from exercise have no clear link to long-term muscle growth (Lixandrão et al., 2018). Recent evidence indicates that systemic hormonal spikes do not predict how much muscle is gained–mechanical and metabolic stress appear to be the true drivers (Lixandrão et al., 2018). Thus, expecting BFR to boost growth via extra hormone release or “metabolic shock” beyond what heavy loads already provide is not supported by current knowledge.

It is worth noting that measurable muscle hypertrophy typically requires several weeks of training. For example, in Teixeira’s study, all groups achieved about 7% muscle growth after 8 weeks (Teixeira et al., 2021). Most RT studies run ≥8–12 weeks to capture muscle growth. Our meta included few very long trials, so we cannot exclude subtle effects over longer periods. Future work should employ longer interventions (e.g., ≥10–12 weeks) to determine if chronic adaptations diverge when BFR is combined with heavy lifting. Until then, the evidence suggests that for trained athletes, BFR + RT offers no hypertrophy advantage over RT alone (Teixeira et al., 2021; Lixandrão et al., 2018). Taken together, these findings suggest that in trained athletes, combining blood flow restriction with resistance training does not produce additional hypertrophic benefits compared to resistance training alone.

The studies included in this analysis assessed strength by measuring the one-repetition maximum in both full and half squats, with results indicating that BFR-RT has a positive impact on lower limb strength. Additionally, the study by Yasuda et al. (2006) demonstrated that 6 weeks of blood flow restriction resistance training led to an average strength increase of 0.3 kg, further validating the effectiveness of BFR-RT in enhancing lower limb strength. Other studies have also reported significant improvements in muscular strength following BFR-RT (36), reinforcing its applicability in athletic training.

However, not all investigations concur. Gavanda et al. (2020) found no additional improvement in leg-press 1RM after 8 weeks of low-load BFR-RT (20% 1RM, 40% limb-occlusion pressure) compared with volume-matched RT in healthy males. Likewise, Lixandrão et al. (2015) observed only trivial changes in squat strength when cuff pressure was set below 40% LOP, suggesting that sub-threshold pressures may fail to elicit sufficient mechanical and metabolic stress.

Most meta-analyses and studies confirm that BFR-RT significantly enhances lower limb strength (Ugur et al., 2023; Yamanaka et al., 2012; Cook et al., 2014). The primary mechanisms underlying this effect include increased mechanical tension and metabolic stress, both of which are considered key factors in muscle adaptation and strength development (Luebbers et al., 2014; Cook et al., 2014). BFR-RT creates a hypoxic and acidic environment (Yasuda et al., 2006), promoting lactate accumulation, which in turn affects muscle contraction and sustains force output by increasing motor unit recruitment (Yamanaka et al., 2012). Additionally, the accumulation of metabolites and cellular swelling (Jessee et al., 2018; Loenneke et al., 2012) may further enhance training adaptations, while the additional recruitment of type II muscle fibers (Loenneke et al., 2015) is regarded as one of the crucial mechanisms by which BFR-RT improves strength performance. Furthermore, research suggests that BFR-RT can stimulate the secretion of anabolic hormones and accelerate the fatigue process, thereby promoting greater motor unit recruitment and overall strength gains (Yasuda et al., 2014).

From a physiological perspective, the high training frequency and appropriate loading levels in BFR-RT enhance muscle activation and increase motor unit recruitment rates during slow, high-tension contractions (Wang et al., 2022). The application of cuff pressure not only facilitates metabolite accumulation but also influences hormone secretion levels (Takano et al., 2005), thereby amplifying training adaptations. Specifically, BFR-RT elicits strength-type neural adaptations—greater motor-unit recruitment and firing rates under sustained tension—while simultaneously promoting protein synthesis, lactate accumulation, and growth-hormone release (Takano et al., 2005; Takarada et al., 2000). Notably, although BFR-RT is effective in improving strength, its adaptive mechanisms may rely more on neural adaptations rather than muscle hypertrophy (Gabriel et al., 2006).

However, several trials indicate that BFR-RT may function mainly as a complement to high-load RT rather than provide additional strength adaptations—especially when strength is evaluated via 1RM, where it often fails to surpass traditional ≥70% 1RM training (Held et al., 2020). Discrepancies across studies appear to hinge on programme and cuff-application variables. For example, Chen (Chen et al., 2022) employed twice-weekly sessions at 30% 1RM with 40% limb-occlusion pressure (LOP) and reported modest gains, whereas Hosseini Kakhak (Hosseini et al., 2022) used four sessions per week at 40% 1RM with 80% LOP and a 13 cm pneumatic cuff, yielding significantly larger improvements. Such contrasts suggest that load (<30% vs. ≥ 40% 1RM), weekly frequency (≤2 vs. ≥ 3 sessions), cuff width (≤5 cm vs. ≥ 13 cm), and—most critically—relative pressure (<50% vs. 60%–80% LOP) jointly modulate mechanical tension, metabolic stress, and thus strength adaptation. Therefore, future research should systematically manipulate these parameters to define an optimal BFR-RT prescription and to clarify its suitability across diverse athletic populations.

The results of the meta-analysis indicate that, compared to RT, BFR-RT is not an effective method for improving lower limb explosiveness in athletes. This finding is not entirely consistent with Xiaolin’s meta-analysis (Wang et al., 2023), which may be due to the inclusion of a “healthy population,” while our study specifically targeted athletes. This difference may have led to contrasting trends in the effects of BFR-RT on lower limb explosiveness.

First, Enhancing lower-limb explosiveness relies on rapid motor-unit firing synchrony, a high rate of force development, and efficient stretch–shortening-cycle behaviour (Enoka, 1997; Moritani, 1993). The BFR-RT interventions included in our meta-analysis mainly used slow-tempo, single-plane resistance exercises and did not incorporate plyometric or ballistic elements—stimuli that are essential for eliciting the explosive-type neural adaptations underlying vertical-jump improvements (Markovic and Mikulic, 2010; Iacono et al., 2016). These findings suggest that the specificity of the movement stimulus—rather than the blood-flow-restriction modality itself—determines whether strength-type gains translate into explosive ability.

Secondly, the impact of BFR-RT on the tendon and tendon-aponeurosis complex is also an important consideration. Existing research indicates that tendon stiffness and elasticity play a critical role in lower limb explosiveness (Maciejewska-Skrendo et al., 2020). Kubo et al.’s study found that, after 12 weeks, LL-BFR-RT did not significantly alter the tensile properties and stiffness of the tendon-aponeurosis complex. Tendon stiffness is believed to be positively correlated with explosive performance in athletes (Kubo et al., 2006; Madarame et al., 2011), especially in actions such as jumping and sprinting, where tendons contribute to force output by storing and releasing elastic energy. Therefore, the lack of significant improvement in lower limb explosiveness in this study may be related to insufficient effects of BFR-RT on tendon characteristics.

Furthermore, in Li et al. (2023) ’s research, BFR-RT was shown to indirectly promote vertical jump height by enhancing muscle metabolic adaptability and motor unit recruitment. However, the results of this study differ from LI et al.’s meta-analysis, which could be due to differences in experimental design, sample characteristics, and research conditions. Li et al. (2023)’s study may have included various types of training methods or sample groups, while our study focused specifically on resistance training, which might have had a different impact on enhancing lower limb explosiveness.

In conclusion, although BFR-RT shows potential for enhancing strength and explosiveness, its effects appear to be no greater than traditional RT. Future research could further explore the effects of BFR-RT under different conditions, such as varying athletic populations, training loads, and cycles, through subgroup analyses, to better understand its role in explosive power training.

The meta-analysis results indicate that BFR-RT does not significantly improve sprint performance in athletes compared to RT alone. This finding is consistent with the majority of studies included in this analysis (Scott et al., 2017; Manimmanakorn et al., 2013; Hosseini et al., 2022; Cook et al., 2014). Research by Fostiak et al. (2022) also found that 30-m sprint performance did not improve following BFR training. The included studies evaluated sprint performance using measures such as 5-m, 10-m, 20-m, 40-m, and 40-yard sprints to comprehensively assess the impact of BFR-RT on sprint performance.

Critical factors influencing sprint performance include reaction time, technique, electromyography (EMG) activity, and neural factors. Reaction time and technique are largely influenced by genetics and training (Mero et al., 1992). In terms of other factors, EMG is often used to assess changes in motor unit recruitment, firing frequency, and synchronization (Wang et al., 2022). However, studies have shown no significant differences in EMG amplitude between BFR-RT and RT (Wang et al., 2022; Moore et al. 2004). Additionally, neural adaptations, such as increased motor unit activation and synchronization, are considered crucial for improving maximal power output (Enoka, 1997; Moritani, 1993; Fujita et al., 2008). RT research indicates that maximal voluntary contraction can increase motor unit activation (Folland and Williams, 2007). However, studies on BFR-RT have not shown significant changes in motor unit activation (Wang et al., 2022; Gabriel et al., 2006; Kubo et al., 2006; Scott et al., 2016), which may explain why BFR-RT does not significantly outperform RT in sprint performance.

It is noteworthy that one study included in our analysis found that while there was no significant difference between LL-BFR-RT and LL-RT in the 10-m sprint, LL-RT outperformed LL-BFR-RT in the 5-m sprint (Manimmanakorn et al., 2013). This suggests that BFR-RT might be more effective than RT for certain short-distance sprints. Systematic reviews by Wortman (Wortman et al., 2021) and research by Mckee et al. (2023) also pointed to the positive effects of BFR-RT on sprint performance. This might be because BFR increases metabolic stimuli during exercise, which more effectively triggers sympathetic nervous system activity, enhancing reaction speed and, consequently, displacement speed (Scott et al., 2016; Kiyohara et al., 2006).

Given these conflicting results, it is possible that BFR-RT’s effects vary with different sprint distances, and differences in BFR settings, training loads, and athlete adaptation levels may lead to varying outcomes. Therefore, more studies are needed to verify the effects of BFR-RT on specific sprint distances in different sports.