Blood‐flow‐restricted exercise: Strategies for enhancing muscle adaptation and performance in the endurance‐trained athlete

What is the topic of this review? Blood‐flow‐restricted (BFR) exercise represents a potential approach to augment the adaptive response to training and improve performance in endurance‐trained individuals. What advances does it highlight? When combined with low‐load resistance exercise, low‐ and moderate‐intensity endurance exercise and sprint interval exercise, BFR can provide an augmented acute stimulus for angiogenesis and mitochondrial biogenesis. These augmented acute responses can translate into enhanced capillary supply and mitochondrial function, and subsequent endurance‐type performance, although this might depend on the nature of the exercise stimulus. There is a requirement to clarify whether BFR training interventions can be used by high‐performance endurance athletes within their structured training programme.


Efficiency/ economy VO 2max
. F I G U R E 1 Schematic diagram of the physiological factors that interact as determinants of endurance performance velocity or power output. Figure modified from Joyner & Coyle (2008). Specific focuses of this review (capillary supply and mitochondrial content/function) are highlighted in red. There are close correlations between capillary density and cycling maximal oxygen uptake (V O 2 max ) (Saltin et al., 1977), time to exhaustion during high-intensity endurance exercise (Coyle et al., 1988) and critical power (Mitchell et al., 2018). Endurance training (∼70-80% ofV O 2 max ) can induce capillary growth in skeletal muscle of previously untrained individuals . There is a close correlation between markers of mitochondrial density andV O 2 max (Flück, 2010;Jacobs & Lundby, 2013). Skeletal muscle maximal oxidative phosphorylation capacity is a predictor of time-trial performance in well-trained cyclists (V O 2 max ∼70 ml min -1 kg -1 ; Jacobs et al., 2011). Multiple training modes induce increases in a range of markers of mitochondrial content (Hoppeler et al., 1985;Perry et al., 2010). Abbreviation: Hb, haemoglobin modifications essential for increasing the abundance and/or function of specific proteins that ultimately improve physiological function (Perry & Hawley, 2018). Therefore, it is important to consider the nature of each single exercise stress and the consequent adaptive mechanisms when constructing the weekly training micro-cycle and the broader, periodized macro-cycle of the individual athlete.
An important consideration is that training adaptations are reduced as training status increases (e.g., Laursen & Jenkins, 2002). This is reflected at a molecular level as a blunting of the signalling response to a single session of exercise (Flück, 2010;Granata et al., 2020;Perry et al., 2010). The question is, therefore, what interventions can be used to augment the stress and subsequent signalling response to exercise which, if repeated over time, will lead to enhanced physiological adaptation and endurance performance, particularly in trained populations? Several strategies have been considered in the pursuit of enhanced adaptation, such as manipulation of training intensity distribution (Seiler, 2010) and nutritional interventions (Rothschild & Bishop, 2020). In this review, we examine the evidence for exercise performed with blood flow restriction (BFR) as an effective strategy to augment the exercise-induced stress and subsequent signalling

New Findings
• What is the topic of this review?
Blood-flow-restricted (BFR) exercise represents a potential approach to augment the adaptive response to training and improve performance in endurance-trained individuals.
• What advances does it highlight?
When combined with low-load resistance exercise, low-and moderate-intensity endurance exercise and sprint interval exercise, BFR can provide an augmented acute stimulus for angiogenesis and mitochondrial biogenesis. These augmented acute responses can translate into enhanced capillary supply and mitochondrial function, and subsequent endurance-type performance, although this might depend on the nature of the exercise stimulus. There is a requirement to clarify whether BFR training interventions can be used by high-performance endurance athletes within their structured training programme. responses to enhance the physiological characteristics of the endurance athlete.

DETERMINANTS OF ENDURANCE PERFORMANCE AND SCOPE OF REVIEW
Endurance exercise performance is defined as the ability to sustain dynamic exercise for extended periods of time (typically, >15 min), including high-intensity intermittent exercise. Endurance performance is largely determined by the maximal power of the aerobic energy system and the fraction of this power that can be sustained (Joyner & Coyle, 2008). These determinants are primarily reflected by maximal oxygen uptake (V O 2 max ) and threshold parameters (e.g., lactate threshold and critical power). The power and capacity of anaerobic metabolic processes also contribute, particularly when exercise is performed in the heavy-and severe-intensity domains, where there is a significant anaerobic contribution to total ATP turnover. The efficiency with which total energy turnover is converted to external work also has a major impact. These determinants of endurance performance are limited by several physiological factors (see Figure 1). The focus of the present review is on adaptations associated with capillary growth and mitochondrial biogenesis, because much of the published research on BFR training in relationship to endurance performance has been concentrated on these factors.

PHYSIOLOGICAL AND MOLECULAR SIGNALS FOR ADAPTATION
The mechanisms underlying the adaptive response to exercise training have been proposed to involve transient physiological and metabolic perturbations that activate molecular signalling pathways within skeletal muscle and endothelial cells ( Figure 2).

Capillary growth
The growth of new capillaries from an existing capillary bed, termed angiogenesis, is stimulated by several exercise-related stressors. One key signal is shear stress, which is the tangential, frictional force exerted on the luminal side of endothelial cells as blood flows along their surface, which increases as blood flow velocity increases (Hudlicka & Brown, 2009). Reduced oxygen tension (i.e., hypoxia) is also thought to play a role in capillary growth (Egginton, 2009). These stressors promote the activation of endothelial cell signalling pathways (Chien, 2007), including vascular endothelial growth factor (VEGF), a crucial pro-angiogenic factor in skeletal muscle (Olfert et al., 2010), which stimulates endothelial cell proliferation and migration, hence angiogenesis (Egginton, 2009). Many studies have demonstrated that VEGF is essential for exercise-induced angiogenesis (Delavar et al., 2014;Olfert et al., 2010). The VEGF-mediated angiogenic response to shear stress is regulated mainly by nitric oxide (NO), which is released in proportion to endothelial nitric oxide synthase (eNOS; NOS3) activity that increases with increasing shear stress (Egginton, 2009;Hudlicka & Brown, 2009). The angiogenic response to hypoxia is mediated, in part, by an increase in the activity of hypoxia-inducible factor-1 (HIF-1) and, in particular, its subunit HIF-1α (Rey & Semenza, 2010). This subunit is sensitive to oxygen levels and, upon activation in hypoxic conditions, translocates to the nucleus, where it induces the expression of multiple gene targets, including VEGF (Forsythe et al., 1996).

Mitochondrial biogenesis
Mitochondrial biogenesis can be defined as the making of new components of the mitochondrial reticulum (Granata et al., 2018) and can include changes in mitochondrial content, mitochondrial respiratory function or other aspects of mitochondrial quality, such as the density of the cristae or supercomplex formation.
Exercise-induced mitochondrial biogenesis is initiated by homeostatic perturbations (Coffey & Hawley, 2007) (a) Lack of activation of skeletal muscle 5′-adenosine monophosphate-activated protein kinase (AMPK) during 120 min of cycling exercise at ∼65% of maximal oxygen uptake (V O 2 max ) in untrained (V O 2 max ∼38 ml min -1 kg -1 ) compared with endurance-trained (V O 2 max ∼62 ml min -1 kg -1 ) individuals (figure modified from McConnell et al., 2020). (b) Blunting of protein expression of nuclear fraction of peroxisome proliferator-activated receptor-γ coactivator 1-α (PGC-1α) after 20 days of high-volume (40 sessions), high-intensity interval training (figure modified from Granata et al., 2020). (c) Blunting of the exercise-induced mitochondrial and angiogenic transcript response after the elevation of mitochondrial volume density by endurance training. Participants exercised at the same relative intensity (65% of maximal aerobic power) in the untrained and trained state (6 weeks, five times per week for 30 min at 65% of maximal aerobic power; figure modified from Flück, 2010, based on the data of Schmutz et al., 2006). (d) Temporal responses of PGC-1α mRNA and PGC-1α protein throughout 2 weeks of high-intensity interval training in recreationally active participants (figure modified from Perry et al., 2010) turn, phosphorylate transcription factors and/or transcriptional coactivators involved in the regulation of DNA transcription (Hood, 2009). These events promote a transient increase in the mRNA content of kinases, transcription factors, (co)activators and downstream proteins, increasing the potential for mRNA translation and the subsequent formation of precursor proteins in the mitochondria (Ljubicic et al., 2010). Peroxisome proliferator-activated receptor-γ coactivator 1-α (PGC-1α) is a key transcriptional coactivator coordinating mitochondrial biogenesis (Puigserver & Spiegelman, 2003), which is also an important pro-angiogenic factor (e.g., Arany et al., 2008).

THE ADAPTIVE SCOPE OF TRAINED MUSCLE
Endurance-trained individuals, defined as those with aV O 2 max of >65 ml min -1 kg -1 , are typically accustomed to varied and voluminous exercise stimuli, such that an increase in training volume alone is insufficient to improve performance (Costill et al., 1988;Hoppeler et al., 1985). The reduced adaptability of trained individuals is also reflected at the tissue level. Cross-sectional and longitudinal studies have demonstrated that the large increases in skeletal muscle capillarity and oxidative enzyme activity (citrate synthase, succinate dehydrogenase and cytochrome oxidase) during the first few months of endurance training are progressively attenuated as training is continued (e.g., Saltin et al., 1977).  (Flück, 2010;Granata et al., 2020;Høier et al., 2012;Jensen et al., 2004;Perry et al., 2010;Richardson et al., 2000;Schmutz et al., 2006).
The reduced plasticity of trained skeletal muscle suggests that adaptations in highly trained individuals primarily serve to maintain their physical performance and/or that the training stimulus is insufficient to promote further adaptation. These observations demonstrate the necessity for introducing variations in training to prevent the attenuation of the molecular signalling responses that drive long-term adaptations essential for the enhancement of performance.
Moreover, as an athlete's training volume and intensity increase, this consideration becomes more important, particularly to avoid over-training and/or injury.

CAN BLOOD-FLOW-RESTRICTED EXERCISE BE USED TO ENHANCE THE ENDURANCE EXERCISE STIMULUS?
Blood flow restriction or complete circulatory occlusion has been used for many years to permit a greater understanding of physiology (e.g., Bull et al., 1989), metabolic function (e.g., Greenhaff et al., 1993;Larsson & Bergström, 1978) and the adaptive responses to exercise interventions (e.g., Esbjörnsson et al., 1993;Sundberg et al., 1993). In an early series of investigations, blood flow was restricted to the legs by placing the participants' legs inside a hyperbaric chamber (Eiken & Bjurstedt, 1987), in which the pressure was increased to 50 mmHg above atmospheric pressure. This caused a decrease of ∼20% in leg blood flow during knee-extensor exercise at the prescribed submaximal work rates (Sundberg & Kaijser, 1992). Blood flow restriction is more commonly applied using an inflatable cuff (Abe et al., 2006), a tourniquet (Shinohara et al., 1998) or an elastic bandage (Loenneke et al., 2010) around the proximal portion of the limb to restrict arterial inflow and prevent venous outflow from the exercising muscles. Recently, a gravity-induced BFR exercise model was developed, whereby cycling was performed with the participants' legs above the level of the heart, which promoted deoxygenation in the exercising muscles in comparison to cycling with the legs below the heart (Preobrazenski et al., 2020).
Multiple variables of cuff application can be manipulated, which has a significant impact on the level of BFR and associated physiological/metabolic responses. Wide cuffs are more effective than narrow cuffs in restricting arterial blood flow at lower inflation pressures (Crenshaw et al., 1988). Compared with wide cuffs (13.5 cm), complete arterial occlusion did not occur using narrow cuffs (5 cm) in some individuals, even at pressures of ≤300 mmHg (Loenneke et al., 2012). Cuff pressure also has a significant impact on the level of BFR. In many studies, it has been common to inflate the cuff to the same absolute pressure (e.g., Credeur et al., 2010;Evans et al., 2010;Kacin & Strazar, 2011;Takarada et al., 2000). However, owing to inter-individual anthropometric and physiological differences (e.g., limb circumference, subcutaneous fat and arterial blood pressure), a standard occlusion pressure might not reduce blood flow to the same degree in different individuals. Moreover, a sex difference in the cuff pressure required to achieve a given level of BFR has been reported (Hunt et al., 2016). Therefore, it has been proposed that the level of BFR should be set relative to the individual (Patterson et al., 2019).
Another important factor to consider is the timing of cuff application in relationship to exercise. In many studies, the cuff is inflated immediately before exercise and remains inflated throughout exercise and during the recovery periods between exercise bouts (Scott et al., 2015). In other studies, the cuff is released during the recovery periods (Christiansen et al., , 2019a(Christiansen et al., , b, 2020, which provides additional episodes with reperfusion. If the exercise intensity is high, it is more feasible to use postexercise BFR, whereby the cuff is inflated rapidly upon completion of each exercise interval and deflated before the subsequent exercise bout (Mitchell et al., 2019;Taylor et al., 2016).
Based on the evidence above, many variables inherent to the BFR exercise protocol can be adjusted, depending on the purpose of the exercise session, to achieve a desired degree of blood circulation and physiological stress in the exercising muscles.

PHYSIOLOGICAL SIGNALS ALTERED BY BLOOD-FLOW-RESTRICTED EXERCISE
The changes in blood flow caused by BFR during exercise and/or in the recovery from exercise expose the peripheral vasculature and exercising skeletal muscle to a distorted level of blood perfusion and oxygenation that gives rise to shear, hypoxic, metabolic and oxidative stress signals. These stressors play putative roles in the adaptive response of the microvasculature and skeletal muscle.

Skeletal muscle blood flow and shear stress
When the BFR cuff is inflated around the limb, different levels of the vascular tree are subjected to altered blood flow profiles. Venous blood pressure is typically <5 mmHg at rest and increases to 8-12 mmHg during dynamic exercise (Rådegran & Saltin, 1998). Therefore, even the relatively low cuff pressures used in most BFR exercise studies (∼80-120 mmHg) result in venous occlusion, increased venous pressure distal to the cuff, reduced distal perfusion pressure and decreased capillary and resistance vessel flow (Hudlicka & Brown, 2009 (Credeur et al., 2010). More moderate levels of BFR (∼60 mmHg) can completely attenuate exercise-induced anterograde shear stress (Tinken et al., 2010), although the effect of BFR exercise on remote sites of the peripheral vasculature, where retrograde shear stress could increase (Green et al., 2002), are not fully known. Upon cuff deflation after BFR exercise, all vessels are exposed to a substantial increase in the rate of blood flow (reactive hyperaemia), which can increase more than 3-fold relative to that observed after intensity-matched exercise without BFR (Christiansen et al., 2019b;Takarada et al., 2000). The increase in blood flow can be maintained up to 1 h after deflation (Gundermann et al., 2012), which is likely to sustain the anterograde shear stress stimulus.

Skeletal muscle oxygenation and hypoxia
Given the reduction in skeletal muscle blood flow, muscle oxygen saturation is reduced in a manner proportionate to the level of BFR (Karabulut et al., 2011). In several studies, reduced muscle oxygenation, assessed using near-infrared spectroscopy, has been observed during low-load resistance exercise or low-intensity cycling with BFR to levels equal to, or even lower, than that achieved during higher-load/intensity exercise in normal blood flow conditions (Corvino et al., 2017;Downs et al., 2014). Christiansen et al. (2019b) demonstrated that the levels of muscle hypoxia during interval treadmill running with BFR (cuff pressure ∼175 mmHg) were commensurate with those achieved during intensity-matched running in systemic hypoxia with an inspired oxygen fraction of 14% (equivalent to an altitude of ∼3250 m above sea level). When BFR is maintained during rest intervals, the reduction in muscle oxygen saturation is sustained (Downs et al., 2014). In all experimental situations, when the BFR cuff was released either between sets or at the end of the exercise bout, the subsequent augmented reactive hyperaemia resulted in a more rapid reoxygenation compared with a non-BFR control.

Metabolic stress
Studies using complete circulatory occlusion have shown substantial alterations in skeletal muscle metabolite content after exercise (e.g., Greenhaff et al., 1993;Krustrup et al., 2009 These exacerbated metabolic perturbations have been demonstrated to be of a similar magnitude to those induced by higher-load exercise (Suga et al., 2009(Suga et al., , 2010.

Reactive oxygen species and oxidative stress
The hypoxic intramuscular environment accompanying BFR exercise provides favourable conditions to generate reactive oxygen species (ROS) (Christiansen, 2019), both in the exercising muscles and in the circulation. Moreover, a second, more potent window for increasing muscle ROS production occurs when the BFR cuff is released in the subsequent recovery period, owing to the marked rise in oxygen availability (Christiansen, 2019). In line with this concept, increases in skeletal muscle markers of ROS accumulation, such as heat shock protein-27, have been observed after a single session of BFR interval running in physically active men . In contrast, direct assessment of mitochondrial bioenergetics in permeabilized muscle fibres has revealed attenuated mitochondrial ROS emission rates 2 h after low-load resistance exercise with BFR (Petrick et al., 2019). However, an earlier elevation (e.g., during and immediately after exercise) in ROS emission rates cannot be excluded. Furthermore, BFR interval training provides more potent protection against excessive cytosolic versus mitochondrial ROS production (Christiansen et al., 2019a), indicating the need to differentiate between ROS sources when assessing the effects of BFR exercise on ROS accumulation.

POTENTIAL FOR AUGMENTING MOLECULAR SIGNALS AND PHYSIOLOGICAL ADAPTATION TO EXERCISE WITH BLOOD FLOW RESTRICTION
The augmented physiological and metabolic stress associated with BFR exercise has the potential to promote the molecular signalling networks and the resultant physiological adaptation beyond what is observed with exercise alone (Figure 4). In this section, we explore this potential of BFR when applied to various types of exercise in relationship to endurance-type adaptations (i.e., capillary growth and mitochondrial biogenesis). For clarity, we make a distinction between low-load resistance exercise (LLRE), low-and moderate-intensity endurance exercise (LI/MI) and sprint interval exercise (SIE). Within each of the following sections, we focus on: (i) primary signals/protein activation; (ii) changes in mRNA content; and (iii) subsequent long-term adaptation.

Low-load resistance exercise
Although not typically associated with endurance adaptations, evidence is accumulating that LLRE combined with BFR (Table 1)  F I G U R E 4 Hypothetical schematic diagram of the potential for blood flow restriction (BFR) to augment the molecular signals and associated physiological adaptations with repeated bouts of exercise training. 'mRNA response' represents the acute transcriptional response to single exercise sessions. 'Protein response' represents changes in protein content and/or function (or post-translational modification). 'Adaptation' represents changes in physiological properties associated with capillary growth and mitochondrial biogenesis (mitochondrial content and respiratory function). Black and red lines represent the magnitude and timing of training-induced adaptation in non-BFR conditions (based on Granata et al., 2018) and BFR conditions, respectively observed after LLRE (30% 1RM) with BFR compared with high-load resistance exercise (HLRE; 70% 1RM; Groennebaek et al., 2018).
Although AMPK and CaMKII activity did not change, phosphorylation of acetyl-CoA carboxylase (ACC), a downstream target of AMPK, also increased to a similar extent after LLRE with BFR and HLRE . The long-term effects of LLRE training with BFR on capillary growth and mitochondrial adaptations remain to be studied comprehensively.
Indirect assessment of microvascular filtration capacity indicates increased skeletal muscle capillarity after LLRE with BFR in healthy men in comparison to work-matched, non-BFR exercise (Evans et al., 2010;Hunt et al., 2013). Recent studies have confirmed that LLRE with BFR increases muscle capillarity. In elite powerlifters, the number of capillaries around type I fibres increased in response to 6.5 weeks of high-frequency (daily) LLRE (∼30% 1RM) with BFR, but not after LLRE alone (Bjørnsen et al., 2019). An increase in capillary-to-fibre ratio (C:F; ∼23%) and capillary area (∼30%) in response to 3 weeks of high-frequency (one or two daily sessions) LLRE (20% 1RM) was observed with BFR performed to volitional failure, in comparison to work-matched, non-BFR exercise (Nielsen et al., 2020

Low-and moderate-intensity endurance-type exercise
Blood flow restriction has been combined with LI or MI dynamic exercise (  (Norrbom et al., 2011;Preobrazenski et al., 2020) or intermittent   phosphorylation have been reported (Norrbom et al., 2011;Ozaki et al., 2014;Preobrazenski et al., 2020), although similar changes were evident in the non-BFR conditions (Norrbom et al., 2011;Preobrazenski et al., 2020). In contrast, p38MAPK phosphorylation remained unchanged after BFR exercise of a shorter duration (15 vs. 30-45 min) and with a lower intensity (40% ofV O 2 max ; Smiles et al., 2017). These contrasting findings might be explained by differences in exercise duration and intensity, BFR parameters or the timing of muscle biopsies. In addition, CaMKII phosphorylation was lowered to the same extent after a session of interval exercise with or without BFR, but not when the session was performed in normobaric systemic hypoxia , suggesting that Ca 2+ -related signalling through CaMKII, when assessed at exercise termination, is not augmented by BFR exercise.
Consistent with the stimulating effect of BFR exercise on AMPK and its role in inducing mRNAs associated with mitochondrial biogenesis, PGC-1α mRNA content increased more after various types of BFR exercise compared with work-matched exercise without BFR, including single-leg knee-extension exercise (Norrbom et al., 2004(Norrbom et al., , 2011, intermittent running  and cycling (Preobrazenski et al., 2020). Indeed, when the same intermittent running protocol Despite the potential for LI/MI exercise with BFR to promote indicators of mitochondrial biogenesis, few studies have explored the long-term effects of BFR training on mitochondrial content and respiratory function. Early work by Kaijser et al. (1990) and Esbjörnsson et al. (1993) showed that citrate synthase activity was higher in a BFR-trained leg (using the hyperbaric method) compared with a non-BFR control leg. In agreement, Christiansen et al. (2020) recently showed that BFR interval training, in contrast to a work- Several studies have investigated how BFR exercise affects angiogenic mRNA content. Initially, no differences were observed between BFR and non-BFR exercise for the increase in VEGF mRNA content in skeletal muscle (Gustafsson et al., 1999). In a subsequent study, greater increases in VEGF mRNA splice variants (VEGF-A 121 , VEGF-A 165 and VEGF-A 189 ) and VEGFR-1 mRNA were evident 2 h after BFR exercise compared with the non-BFR control (Gustafsson et al., 2005). The discrepancy between these two studies might be attributable to the timing of the postexercise biopsies, which were obtained only 30 min postexercise in the initial study (Gustafsson et al., 1999), typically before any measurable changes in VEGF mRNA content occur (Kuang et al., 2020). More recently, similar increases in VEGF-A mRNA content were observed after BFR exercise and the non-BFR control (Preobrazenski et al., 2020). Thus, the effect of a single session of BFR exercise on VEGF mRNA content remains unclear.
Longitudinal studies have explored the effect of LI/MI training with BFR on protein markers of angiogenesis. A greater increase in the resting muscle abundance of VEGF-A protein was reported after 5 weeks of continuous BFR training compared with a training control (Gustafsson et al., 2007). Furthermore, the basal level of Ki67 mRNA increased more in the BFR-trained leg (Gustafsson et al., 2007). In contrast, muscle VEGF protein abundance remained unaltered after 6 weeks of interval cycling with BFR (Christiansen et al., 2019a).
The contrasting findings for VEGF protein might be attributed to differences in the exercise training protocols. However, the selective increases in leg blood flow (Christiansen et al., 2019b) and oxygen delivery and uptake (Christiansen et al., 2020) during single-leg exercise for the BFR-trained leg indicates that a microvascular adaptive response had taken place early during the intervention. Indeed, although capillary supply was not assessed (Christiansen et al., 2020), the classic studies of Sundberg and colleagues had already established the potential of LI/MI exercise training with BFR to increase muscle capillary density (Esbjörnsson et al., 1993).
Overall, the present data indicate that LI/MI BFR exercise can invoke a greater stimulation of endurance-type adaptations than absolute intensity-matched exercise without BFR. To elaborate on the potency of LI/MI BFR exercise for promoting endurance-type adaptations, future studies should incorporate a high-intensity exercise control. Furthermore, the findings above also underscore that although mRNA and protein measurements can provide interesting insights about the potential of a training strategy for increasing endurance-type adaptations, they do not necessarily reflect adaptations in muscle functional properties (e.g., blood perfusion, O 2 delivery and mitochondrial function) at a given point in time. Further research is warranted to confirm and expand these initial observations, particularly in cohorts of endurance athletes.

Sprint interval exercise
Although the combination of BFR with LI/MI exercise augments some of the acute signals and long-term physiological adaptations associated with endurance performance, the maintenance of a high training intensity is considered important to optimize adaptations and performance in trained athletes (Laursen et al., 2005). To date, only two studies have investigated the effect of a single session of SIE with BFR (Table 3). In one of these studies, Taylor et al. (2016) used an SIE protocol on trained participants (V O 2 max ∼60 ml min -1 kg -1 ), which consisted of repeated, maximal 30 s sprints separated by 4.5 min of recovery, during which BFR (∼130 mmHg) was rapidly applied and maintained for 2 min. The addition of BFR did not augment the increase in AMPK phosphorylation immediately after the exercise session (Taylor et al., 2016). Consistent with this finding, there was no further increase in PGC-1α mRNA content 3 h after the session, increase with SIT plus BFR in trained subjects (Mitchell et al., 2019).
This adaptation often precedes any measurable changes in skeletal muscle capillary supply (Høier et al., 2010(Høier et al., , 2020. However, there was no increase in any index of skeletal muscle capillarity (capillary density, capillary-to-fibre ratio and capillary-fibre contacts). Moreover, citrate synthase, COXII and COXIV protein abundance remained unchanged (Mitchell et al., 2019).
Although shear stress was likely to be high during SIE with postexercise BFR, the accumulated shear stress stimulus might have been lower than that invoked by the continuous, moderate-intensity exercise protocols discussed in the previous section, owing to the lower overall training volume (Høier et al., 2013). Thus, the insufficient training volume and an attenuated BFR-induced shear stress stimulus could explain the absence of changes in capillary markers in the study by Mitchell et al. (2019). Another explanation might be the high training status of the participants, in whom the resting muscle C:F ratio (∼2.9) was almost double that of the untrained participants (∼1.6) in a 6 week SIT study that induced a ∼30% increase in C:F ratio (Cocks et al., 2013). ↑ indicates greater increase in BFR compared with the control conditions (e.g., no BFR or high-load resistance exercise conditions). ↔ indicates no difference in changes between BFR and CON. → indicates no change in BFR or CON. Abbreviations: Ang-2, angiotensin II; BFR, blood flow restriction; CON, control; COXII, cytochrome C oxidase subunit II; COXIV, cytochrome C oxidase subunit IV; eNOS, endothelial nitric oxide synthase; HIF-1α, hypoxia-inducible factor-1α; MMP-9, matrix metalloproteinase 9; p38MAPK, p38 mitogen-activated protein kinase; PGC-1α, peroxisome proliferator-activated receptor-γ coactivator 1-α; rep or reps, repetition;

TA B L E 3 Studies investigating skeletal muscle adaptive responses to sprint interval exercise combined with blood flow restriction
SIT, sprint interval training; VEGF, vascular endothelial growth factor; VEGFR-2, vascular endothelial growth factor receptor-2;V O

ENHANCING ENDURANCE PERFORMANCE WITH BLOOD-FLOW-RESTRICTED EXERCISE TRAINING
In the previous sections of this review, we have examined how different types of BFR exercise affect the physiological stressors and some of the molecular signalling networks related to mitochondrial and vascular adaptations that support endurance performance. In this section, we explore how the different types of BFR training (LLRE, LI/MI and SIE) impact endurance-related exercise performance (Table 4). Kaijser et al. (1990) and Esbjörnsson et al. (1993) made the first observations of an improved endurance exercise performance with BFR training. In these studies, physically active participants completed 4 weeks of single-leg, continuous training with BFR (using the hyperbaric method). Greater improvements in single-leg exercise time to exhaustion in ischaemic conditions were observed after BFR training compared with the control group who performed the same training without BFR. In parallel work by Sundberg et al. (1993), greater improvements in single-leg peak oxygen uptake were achieved with continuous BFR training than with work-matched training without BFR.
Studies of LLRE with BFR have consistently demonstrated enhanced skeletal muscle endurance exercise capacity. In most cases, greater improvements in the number of repetitions performed at a fixed percentage repetition maximal or maximal voluntary contraction have been observed after BFR compared with non-BFR training (e.g., Groennebaek et al., 2018;Kacin & Strazar, 2011;Manimmanakorn et al., 2013). Pignanelli et al. (2020) demonstrated a greater increase in the average power through the mid-portion of a 30-repetition maximal effort muscle endurance task after 6 weeks of single-leg squat training to failure at 30% 1RM with BFR compared with a non-BFR control.
Several studies have also shown that LI/MI training (using various types of exercise) with BFR improves endurance-related performance outcomes. For example, greater improvements inV O 2 max have been found after LI/MI exercise with BFR than after intensity-matched training alone (Abe et al., 2010;de Oliveira et al., 2016), although this is not a universal finding (Kim et al., 2016;Paton et al., 2017). Other indices of endurance performance, such as time to exhaustion (Abe et al., 2010;Christiansen et al., 2019a) and power output during an incremental test (Christiansen et al., 2020;de Oliveira et al., 2016), are also improved in response to this type of BFR training in trained individuals.

CONCLUSIONS AND FUTURE RESEARCH
de Oliveira et al.

Participants and study design
Method Performance Paton et al. (2017) Healthy, recreationally active males and females (n = 16) V O 2 max ∼46 ml min −1 kg −1 Independent groups design: BFR group n = 8, CON group n = 8 4 weeks of treadmill run training: 2 sessions week −1 of two or three sets of 5-8 reps of 30 s running at 80% of peak running velocity BFR: elastic wraps, applied during the running bouts, width 7.5 cm, at a perceived pressure of 7 out of 10, released between each set CON: no wrap applied; work-rate matched ↔ increase inV O 2 max in BFR (6%) and CON (4%) ↑ peak running velocity achieved during maximal incremental test in BFR (4%) versus CON (1%) ↑ TTE at peak running velocity in BFR (27%) versus CON (17%) Taylor et al.
& White, 1999). Indeed, given the substantial role played by muscle afferent activity for endurance capacity (Amann et al., 2020), any such attenuated feedback after BFR training could contribute to improved performance.
Notwithstanding the safety concerns, the present review supports that BFR exercise may be used to maintain adaptations already gained as a component of tapering phases where the training volume is reduced and the intensity maintained (or increased). Models of BFR training might have applications during specific periods of the athlete's season, for example in rehabilitation after injury, when high mechanical loads are contraindicated. Either way, the potential for BFR exercise to provide 'more bang for buck' for the endurance-trained individual is an exciting prospect that demands further attention.

COMPETING INTERESTS
None declared.

R.
A.F. wrote the initial manuscript. All authors critically revised and contributed to the revised manuscript. All authors approved the final manuscript and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.