A total of 159 documents were retrieved from the CNKI, Wanfang, VIP, PubMed, and Web of Science databases up to 4 February 2023. The English search terms used were as follows: (“blood flow restriction training” or “BFR” or “KAATSU training” or “pressure training”) and (“Potentiation after activation” or “PAP” or “muscle recruitment” or “explosive power” or “lower extremity”) and (“RCT”).

The collected literature were imported into EndNote software and underwent independent screening by two researchers. The steps of literature screening and inclusion are illustrated in Figure 1. Ultimately, 18 articles were included in this review.

Data extraction: Two researchers independently extracted information according to a custom-made form, which primarily encompassed the following categories:

1. General information: first author and year of publication.

Statistical analysis was performed using ReviewManager 5.3 software. The outcome indicators included in the literature in this study were continuous variables, and standardized mean differences (SMDs) and 95% confidence intervals were chosen for effect sizes because of the different testing methods used for each indicator. The Cochrane risk bias assessment tool was referenced for literature quality assessment. The homogeneity test (Q test, test level a = 0.1) was used to test for heterogeneity, and I ² values from 0% to 100%, I ² > 50%, and P < a indicated the existence of heterogeneity, and the random effects model was selected. The random effects model was chosen for meta-analysis, while the fixed effects model was chosen for the opposite. The subgroup analysis was used for the treatment of heterogeneity and STATA 16.0 was used for sensitivity analysis to test the stability of the results. The Egger test and funnel map were used to check for the presence of publication bias.

A total of 18 publications were included in this study, all of which were RCTs, which included 267 subjects with mixed genders and an age range of 18–33 years, with the basic characteristics shown in Table 1.

The quality of the collected literature was evaluated with reference to the Cochrane risk of bias assessment tool (Higgins and Altman, 2007). Review Manager 5.3 software was used to assess seven aspects, namely, the random allocation method, allocation scheme concealment, participant blinding, outcome blinding, completeness of outcome data, selective reporting of study results, and other sources of bias (Figures 2A, B). Twelve articles did not clearly describe whether the implementer of the assignment had strictly performed the random assignment, and 18 articles had a high risk of bias due to informed consent being signed before the experiment. Thus, there was a high risk of bias in the blinding.

Out of the 18 included articles, 10 compared 127 subject differences in RMS standard values before and after BFR training (Figure 3). After testing for heterogeneity, I ² = 0.0% < 50% and p = 0.70 > 0.1 for the Q test, suggesting no heterogeneity among the included literature. The fixed effects were selected for the meta-analysis. The SMD value for the pooled 10 studies was 0.98, with a 95% confidence interval from 0.71 to 1.224 and statistically significant, with Z = 7.21 and p ≤ 0.05, indicating that BFR exercises can significantly improve RMS values in lower limb muscles.

For BFR training–induced PAP, 13 articles were included (16 studies with a total of 261 subjects) and were tested for heterogeneity, and I ² = 51% > 50% and Q-test of p = 0.01 < 0.1, indicating a strong heterogeneity among the included literature, and random effects models were selected. A meta-analysis was performed (Figure 4) and the results showed that the combined effect size was SMD = 0.28 and statistically significant (Z = 2.13 and p = 0.03), which indicates that BFR training could significantly induce the production of PAP when compared with the control group.

Based on the data of this study, the authors suspect that the source of heterogeneity may be the inconsistency of PAP effect test indicators. Therefore, according to the test indicators of the PAP effect (MVC: torque, CMJ: longitudinal jump, and P(W): output power), 13 articles (16 studies) were divided into three groups for the subgroup analysis (Table 2). According to the exercise modality, the included studies were divided into three subgroups: ≤30% 1RM resistance exercises, ≥70% 1RM resistance exercises, and self-weight exercises. For compressive strength, the subgroups were categorized based on the arterial occlusion and thigh circumference, according to Loenneke Jeremy et al. (2015), and three subgroups were identified: 40% AOP and below, 40%–60% AOP, and 60% AOP and above.

The subgroup analysis of the outcome indicators showed that the heterogeneity of the three groups was 66%, 28%, and 0%. When compared with the overall combined effect (I ² = 51%),the within-group heterogeneity of the MVC moments had increased to 66%, which suggests strong heterogeneity among the studies with respect to the outcome indicator of the MVC moment. The heterogeneity within the literature on the outcome indicator of the MVC moment was strong. The CMJ longitudinal jump group had the highest effect size and was statistically significant (SMD = 0.45 and p = 0.0004 < 0.01). The available results suggest that BFR training can significantly improve CMJ performance. The available results also suggest that BFR training can significantly increase the vertical jump height of CMJ.

The subgroup analysis of exercise patterns showed that the heterogeneity of the three groups was 40%, 0%, and 33%, which is lower than the overall combined effect (I ² = 51%). It is speculated that different exercise methods may be the influencing factors that lead to the heterogeneity between studies. When compared to resistance exercises at 30% 1RM and below, as well as resistance exercises at 70% and above, self-weight exercise had the largest effect, with an SMD of 0.57, and it was statistically significant (p < 0.01). This indicates that BFR training combined with self-weight exercise can significantly induce the generation of PAP.

The subgroup analysis of pressure intensity revealed varying degrees of heterogeneity among the three groups: 73%, 19%, and 0%, respectively, in comparison to the overall combined effect (I ² = 51%). Notably, within the subgroup of compressive strength of 40% AOP and below, the intra-group heterogeneity increased to 73% (I ² = 73%), which indicates significant heterogeneity among the studies in this category. Among these, the effect size of the 40%–60% AOP group was the highest and statistically significant (SMD = 0.57 and p < 0.01), which suggests that the pressurization intensity of BFR training at 40%–60% AOP significantly induces PAP production.

Sensitivity analysis was conducted on the included literature, both by including and excluding individual groups of studies, to assess heterogeneity.

In Table 3, the combined effect of BFR training on lower limb muscle activation, as included in all studies, resulted in an SMD of 0.98 (p < 0.01 and I ² = 0%). The SMD range from individual studies was from 0.94 to 1.06, with I² = 0%, and all p-values were <0.01. There was no single literature that posed a threat to the meta-analysis results, which indicates the study’s stability.

Table 4 presents the combined effect of BFR training on the PAP effect across all included studies, with an SMD of 0.28 (p = 0.03 and I ² = 51%). After excluding the studies by Hongwen and Xiang (2022) and Lauver et al. (2017), the combined effects were 0.09 and 0.002, respectively, with I ² values of 37% and 31%. This resulted in a significant reduction in heterogeneity. The combined effect, after excluding other single studies, had an SMD of 0.02–0.07, with p-values < 0.1.

We conducted bias tests separately by drawing funnel plots within the subgroups to assess the publication bias of the included studies. The funnel plots exhibited symmetry, which indicates no publication bias, as illustrated in Figures 5A–D.

Furthermore, we continued to examine potential bias using the Egger test, and all p-values were found to be >0.05, as presented in Table 5. Therefore, it can be concluded that there is no publication bias present in the literature analyzed in this study.

The induction method employed in this study is BFR training. By combining low-intensity resistance exercises (Lauver et al., 2017; Cleary Christopher and Cook Summer, 2020; Hu, 2020; Li, 2020; Che et al., 2021; Wang, 2021), high-intensity resistance exercises (Pedro et al., 2016; Pan et al., 2019), and aerobic activities such as running (Amani et al., 2019; Yun-Tsung et al., 2019), we assessed their impact on electromyography (EMG) readings of lower limb muscles. As for the criteria for evaluating lower limb muscles, specific sites below the iliac crest point (Wu et al., 2020) were selected, which primarily included the vastus medialis (VM), rectus femoris (RF), gluteus maximus (GM), biceps femoris (BF), and other muscle groups.

Muscle mass and strength development are governed by neural modulation and stimuli (Marcotte et al., 2015). The results of the meta-analysis indicate that the effect sizes from the overall effect tests in all 10 studies were consistently positive (p < 0.01). Therefore, BFR training significantly enhances the root mean square (RMS) values of lower limb muscles. The heterogeneity test results indicate complete homogeneity among the study outcomes (I ² = 0%), and the sensitivity analysis showed no significant changes in heterogeneity or combined effects when any single study was removed. This suggests that there are no differences related to participant characteristics, exercise methods, pressure intensity, etc., with all included studies pointing to the same outcome.

It is well known that high-intensity exercise is achieved through the recruitment of both fast and slow muscle fibers by the nervous system. Studies have demonstrated a positive correlation between the degree of muscle activation and number of nerve fibers (Westing et al., 1991). When compared to centrifugal exercise alone, combining centrifugal exercise with BFR leads to greater neuromuscular activation (Schoenfeld Brad, 2012). This increased neuromuscular activation observed during BFR exercises is attributed to reduced muscle oxygen availability due to compression, which inhibits α motor neurons. To meet the altered energy demands, more type II (fast-twitch) muscle fibers are recruited for participation in the exercise (Yasuda et al., 2015). In the literature within this study, the control group’s intervention measures involved non-pressurized resistance exercises. Therefore, in comparison to traditional resistance exercises, BFR exercises may facilitate the recruitment of more fast-twitch muscle fibers. This recruitment potentially has a positive impact on improving explosive strength and muscle strength in athletes.

As a physiological phenomenon characterized by an acute increase in explosive force, methods aimed at inducing PAP are primarily based on heavy-load activities such as squats, bench presses, and high-intensity exercises (Zhuang, 2021). In this study, we conducted a meta-analysis of various explosive force measurement indices, such as MVC moment, CMJ vertical jump height, and P maximum output power, before and after BFR training. The results revealed an overall effect size of 0.28 (p < 0.05) across 16 studies, which indicates that BFR training can induce PAP.

Traditional PAP induction methods typically require heavy training equipment for short-term maximum or sub-maximum resistance training, which can pose a risk of sports injuries and subsequent muscle fatigue due to their high-intensity and heavy-load nature. By contrast, BFR training is characterized by low intensity and rapid recovery. Furthermore, creatine kinase (an index representing the degree of muscle damage) did not significantly increase during BFR training, suggesting that muscle fiber damage is lower than it is in high-intensity training (Xu et al., 2021).

In recent years, new warm-up programs based on the PAP effect have gained attention in the academic community. These programs combine resistance training with isometric training to enhance sports performance and have been adopted as pre-match warm-up routines to induce PAP (Yang and Huan, 2022). In summary, BFR training offers a relatively simple and convenient method for effectively inducing the PAP phenomenon when compared to traditional high-intensity resistance exercises.

The production of PAP is primarily attributed to factors such as muscle acidification, strengthening of the H-reflex, and changes in muscle fiber recruitment angles under load stimulation, which result in an increased number of recruited motor units due to motor nerve excitation (Chong et al., 2020). This phenomenon aligns with the neuromuscular activation observed in BFR exercises as mentioned earlier. Cleary Christopher and Cook Summer (2020) also found that EMG amplitudes in activated hamstrings were the highest under BFR conditions during blood restriction training, suggesting a potential link between lower limb muscle activation and BFR training–induced PAP.

The 18 articles included were not blinded and the methodology is limited in this respect. It is difficult to use blinding in this type of study because of the corresponding ethical requirement for human subjects to sign an informed consent form. In the few articles retrieved for this study, the authors could not be contacted to obtain the required data, which is theoretically biased, but the authors do not believe that this would have a subversive effect on the results of this study. Limitations in the research field and theme, such as the small number of studies evaluating lower extremity muscle function, the generally low sample sizes, and variations in intervals after BFR exercises among studies, may have introduced potential bias. Furthermore, despite extensive research, there remains a shortage of large-scale, high-quality studies on BFR training. Future studies should improve the credibility of the study by expanding the sample size as much as possible by improving the experimental design. For PAP induction, the effects of the training status, exercise interval time, and other factors on the PAP effect should be further clarified.