The Safety, Feasibility, and Ecacy of Upper Limb Garment-integrated Blood Flow Restriction Training in Healthy Adults

Blood ow restriction training (BFR) has been demonstrated to increase muscle hypertrophy and strength, but has logistical and cost barriers. Garment-integrated BFR has the potential to reduce these barriers, by lowering equipment demands and cost at point of access. The primary aim of the study was to explore the safety and feasibility of garment-integrated BFR in the upper limb of healthy adults, with a secondary aim of exploring ecacy.

3. Garment-integrated BFR can now be safely applied in a randomised controlled trial design to determine its e cacy when compared to either a wait and see control, or another form of BFR.

Background
Blood ow restriction (BFR) training originated in Japan and is known as "kaatsu" training, which translates as "added pressure" (1). BFR involves partial occlusion of limb vasculature using a torniquet or cuff, positioned at the most proximal part of the limb being trained (2). Low-intensity exercise (~30% onerepetition maximum) in isolation facilitates muscle endurance rather than strength (3), but when combined with BFR, promotes greater strength adaptations and hypertrophy (4). Low-intensity exercise combined with BFR can facilitate comparable levels of hypertrophy and human growth hormone plasma concentration to high-intensity exercise (~70% one-repetition maximum; [5][6][7]. It is hypothesised that partial occlusion of the vasculature may cause an acute muscle swelling response via an extracellular uid shift, resulting in the activation of the rapamycin (mTOR) signalling pathway and stimulation of an anabolic response in muscle tissue (8-10).
The safe application of BFR has been reported in many different populations, with a paucity of minor and serious side-effects (11). Many individuals are unable to perform high-intensity exercise for variable logistical and safety reasons. BFR has previously been reported to facilitate hypertrophy and improve functional capacity in both healthy (12) and injured (13) populations. Current application of BFR training involves either the use of a pneumatic cuff or a simple tourniquet (e.g., rubber tubing). Pneumatic cuffs allow for standardisation of occlusion pressure and limb placement but are expensive (£350-£10,000) and require supervision for successful use, limiting their wider application. Simple tourniquets can be unsafe as they do not allow for standardisation of limb placement or the reproducible determination of partial (rather than total) occlusion pressure (14). BFR is generally used by trained professionals with specialist equipment to accurately regulate applied occlusion pressures. Most research to date uses a de ned percentage of total arterial occlusion pressure, which is not re ective of the practical application of BFR training (15). Comparable muscular responses in strength and hypertrophy are reported at varying levels of occlusion pressure (16), which questions the need for de ned measurement of percentage of total arterial occlusion pressure outside of assuring partial (rather than total) occlusion pressure. Garment-integrated BFR has been developed to provide an inexpensive and safe method of BFR application. This allows for consistent placement of an integrated strap of a standardised width and the subjective but reproducible determination of occlusion pressure (17). Garment-integrated BFR can also be used without the supervision of a quali ed professional, increasing the potential application in a variety of settings.
The primary aim of the study was therefore to explore the safety and feasibility of garment-integrated BFR in the upper limb of healthy adults, to inform upon future larger scale studies. A secondary aim was to explore the e cacy of garment-integrated BFR. The null hypothesis was that garment-integrated BFR of the upper limb would be unsafe and infeasible in healthy adults.

Methods
An observational safety and feasibility cohort study was conducted.

Ethical approval
Ethical approval was granted by the Queen Mary Ethics of Research Committee (QMREC2018/48/054).

Participants
Participants were recruited as a sample of convenience and provided written informed consent prior to study commencement using Google Forms (Google LLC, California, USA). Sample size was based on previous guidelines for safety and feasibility studies (18), with a minimum of 12 male and 12 female participants sought. Participants were eligible for inclusion if they were over the age of 18 and currently in good health, injury-free in their upper limbs and willing to perform two BFR sessions per week and cease other forms of upper limb exercise for the study duration. Participants were excluded if they had any history of deep vein thrombosis (DVT) or pulmonary embolism (PE), a previous diagnosis of rhabdomyolysis, haemorrhagic or thrombotic stroke, previous surgery in the past six weeks, were currently or recently pregnant, or had a family history of any blood clotting disorder. Participants were also excluded if they had any prior experience with BFR.

Demographics
Eligible participants self-reported their age (in years), height (to the nearest cm), mass (to the nearest kg) and activity level using the Tegner scale. Combining both work and sports activities, the Tegner scale is a reliable and valid measure to re ect the average activity levels of recruited participants (19).

Experimental protocol
The study was conducted online using Microsoft Teams (1.400.11161, Microsoft, Washington, USA) due to the SARS-CoV-2 pandemic. All included participants followed a ve-week upper limb BFR programme consisting of two sessions per week and a total of ten sessions and were advised to avoid other forms of upper limb exercise for the duration of the study. Participants were provided with a garment with integrated BFR (Hytro Limited, London, UK). This uses a standardised elastane strap (width 4cm) located at the most proximal part of the upper limb and secured with a Velcro mechanism (YKK Fastening Corp, Tokyo, Japan; see gure 1) to allow for standardisation of compression stimulus. Participants were also provided with a standardised resistance band (male=red [tension 7-16kgs] and female=yellow [tension 2-9kgs]), a exible tape measure, and a pulse oximeter (NHS Pulse Oximeter, London, UK). Participants had an initial familiarisation meeting with a researcher (BD or EM), where they were introduced to the BFR programme and protocol. Participants were then instructed to pull the BFR strap on their dominant arm to its maximal position (re ecting 100% compression stimulus), before releasing to a perceived 50% compression stimulus, noting the corresponding number on the Velcro mechanism. Participants then underwent a three-minute passive BFR session at 50% compression stimulus to familiarise them to the sensation of BFR, before releasing. Participants nally completed four sets of a single exercise (banded bicep curls) at 50% compression stimulus, resting for 30 seconds in between sets and taking a pulse oximeter reading before and after the exercise was completed with compression applied for familiarisation.

BFR programme
Each BFR session involved four exercises (push-ups, banded bent over rows, banded triceps extensions, and banded bicep curls). Male participants were instructed to complete full push-ups, whilst female participants were instructed to complete kneeling push-ups. Participants were required to complete four sets of each exercise (30/15/15/15 repetitions), taking a 30 second rest interval between sets and a two minute rest interval between exercises (1). If participants reached volitional failure prior to the prescribed number of repetitions in a set, they were asked to record their number of successful repetitions.
Participants were instructed to tighten the BFR strap to a perceived compression stimulus of 50% for their rst session and increase to 60% for their second session as a familiarisation week. Participants were then instructed to increase to a perceived compression stimulus of 70% for the remaining eight sessions (weeks two-ve). The BFR strap was to be applied prior to commencing an exercise and remain secured for all four sets (i.e., 75 repetitions), releasing at the start of the two-minute rest interval between exercises.

Safety outcomes
Safety was determined by recording the occurrence of adverse events during and after BFR and by monitoring for potential arterial occlusive limb pressure using pulse oximetry.

Questionnaire
Potential adverse events that could re ect thrombosis or ischaemia (20) were monitored (see table 1).
Participants were instructed to report the occurrence of these events by completing a safety questionnaire after each BFR session and attend a weekly virtual meeting with a researcher (BD/EM) to discuss any adverse events that may have occurred during the preceding week. Participants were required to con rm the presence of an upper limb pulse by taking a pulse oximeter reading before commencing each exercise (once their BFR strap had been applied at the required perceived compression) and once each exercise had been completed (prior to releasing their BFR strap and commencing their rest period). Pulse oximetry has been reported to be a valid method of ensuring sub-occlusive arterial pressure (i.e., the presence of a pulse) in the upper limb when compared to the gold standard of ultrasound doppler (21).

Feasibility outcomes
Successful recruitment was determined by the time within which a minimum of 24 participants could be recruited, with a maximum of three months de ned a priori.
Successful adherence was determined by monitoring the number of sessions completed by each participant (x/10), with a minimum of 80% required a priori.
Successful data collection was determined by outcome measure capture, with a minimum of 80% required a priori.

E cacy outcomes
Push-ups to volitional failure All participants performed a push-ups to volitional failure test in their familiarisation meeting, but before their BFR familiarisation. Total push-ups were observed and recorded by the researcher during the video call and the test was ceased once participants failed to meet the minimum movement standard of 90e lbow exion (i.e., volitional failure). This was then repeated in the nal meeting after the ve-week BFR programme. Push-ups to volitional failure was chosen as a proxy measure of strength as it has been reported to correlate well with a one repetition maximum bench press using an equivalent load (22). It could also be performed virtually and without requiring participants to attend a human performance laboratory during the SARS-CoV-2 pandemic.

Arm girth
Arm girth was measured by the participant using a exible tape measure according to the International Society for the Advancement of Kinanthropometry (ISAK) guidelines (23). Participants were instructed to measure from their acromion process to their cubital fossa on their right arm, marking the midpoint.
Participants were then instructed to take a circumferential measurement of their arm at this point to the nearest 0.5cm, with their arm relaxed in the anatomical position (23).

Number of prescribed repetitions completed
The total number of repetitions completed were compared from week two to week ve, excluding week one as a familiarisation week, as a measure of muscular endurance (24).

Statistical analysis
Data were collected and collated using a customised spreadsheet (Microsoft Excel 16.0.13426320270, Microsoft, Washington, USA). Safety and feasibility data were analysed using Microsoft Excel. Safety outcomes were calculated by dividing the incidence of reported adverse events by the total number of BFR sessions (x/280) and expressed as a percentage. Adherence outcomes were calculated by dividing the number of completed sessions by the total number of prescribed sessions (x/280) and expressed as a percentage.
E cacy data were analysed using JAMOVI (v.1.6.23, the JAMOVI project, Sydney, Australia). Mean change and associated standard deviation (SD) were calculated for push-ups to volitional failure and arm girth (follow up -baseline), and total number of repetitions completed (week ve -week two). A Shapirowilk normality test was conducted to determine if data were normally distributed. As a feasibility study not powered apriori to detect statistical signi cance, dependent sample t-tests were not performed, and pvalues not reported, because of the potential for type II error and to avoid giving the impression of there being robust ndings from a feasibility design. Instead, mean change with 95% con dence intervals (CI) and effect sizes (25)   Safety outcomes

Adverse events
One participant reported one incidence of excessive pain during exercise (0.36%). Two participants reported one incidence of excessive pain post-exercise (0.72%). One participant reported one incidence of bruising in the locality of the BFR strap post-exercise (0.36%). No other adverse events were reported.

Pulse oximetry
An absent pulse oximeter reading was reported by four participants either before or after an exercise a total of 82 (out of 2240) times (3.7%), with one participant reporting an absent pulse oximeter reading 56 times and another participant reporting an absent pulse oximeter reading on 24 occasions. The remaining two incidences of absent pulse oximeter readings came from two separate participants.

Number of prescribed repetitions completed
The total number of repetitions completed each week increased from week two (576.7 ±30.5) to week ve (581.9 ±34.0). Shapiro-Wilk normality test indicated non-normally distributed data (p≤0.001) and so a mean change of 11.5 (95% CI 0.50, 28.00, RBC 0.50; see gure 2c and table 3).

Discussion
This study aimed to explore the safety, feasibility, and e cacy of garment-integrated BFR in the upper limb of healthy adults. Consistent with our hypotheses, garment-integrated BFR was identi ed to be safe and feasible, with increases observed in push-ups to volitional failure and total repetitions completed, but not participant-measured arm girth.

Safety
Garment-integrated BFR was identi ed to be safe for use in the upper limb of healthy adults, re ected by the minimal presence of adverse events and con rmation of sub-occlusive pressure using perceived compression stimulus. Muscle soreness is a common side effect of low load resistance exercise combined with BFR (1), which may persist for up to 72 hours (20). Two participants (8%) in this study reported excessive muscle pain post-exercise. Upon further exploration, one of these participants failed to cease the protocol at the point of volitional failure, instead completing the maximum possible repetitions with small rest intervals. The other participant went against the advised study protocol and combined his BFR protocol with regular golf (36 holes in a ve-day period). After a week of prescribed rest and reinforcement of instructions these participants returned to the protocol without any further excessive muscle soreness, indicating that their initial episodes are unlikely to be the direct result of garmentintegrated BFR.
A maximum of 80% arterial occlusive pressure is advocated when combining BFR with resistance exercise (1). Use of BFR with total occlusive pressure is not advised to minimise the potential for more serious adverse events (1). In the absence of a gold standard of ultrasound doppler, pulse oximetry was used to ensure sub-occlusive arterial pressure (21). A successful pulse oximeter reading was achieved 2158/2240 times, with a pulse oximeter reading absent 82 times (3.7%). Upon further exploration, one participant accounted for 56/82 absent readings, which was recti ed by the provision of a new pulseoximeter, suggesting equipment failure. A further participant accounted for a further 24 absent pulse oximeter readings, likely to be explained by the use of acrylic false nails (28). Participant determined compression stimulus is therefore a safe method to use in future trials to ensure sub-occlusive pressure when using garment-integrated BFR, with the secure Velcro mechanism allowing for consistent replication of the required compression stimulus.

Feasibility
All three of the apriori de ned feasibility outcomes were satis ed. The required number of participants were comfortably recruited within three months, and complete data were obtained from 96% of participants. Adherence to the garment-integrated BFR protocol used was high (99.3%), which is comparable to adherence rates reported by other BFR feasibility studies in clinical populations (29). The one male participant who failed to complete the nal two BFR sessions wished to return to his typical upper limb training routine, which led to his withdrawal. Overall, this gives a high degree of con dence that garment-integrated BFR could be scaled up and investigated using a randomised controlled trial design.

E cacy
Whilst not designed for hypothesis testing, a mean increase of eight push up repetitions to volitional failure was observed. This may re ect an improvement in muscle strength, as push-ups are a valid predictor of upper body strength (22). An increase in strength would also expected after combining low load resistance training with BFR (4). The absence of a control group in this study means that it is impossible to separate a dependent training effect from an independent effect of BFR. With feasibility established, future research should look to investigate the e cacy of garment-integrated BFR in an adequately powered trial with an appropriate control.
No change in participant-measured arm girth was observed amongst the participants in this study.
Because of the limitations placed on clinical research during the SARS-CoV-2 pandemic, participants were required to measure their own arm girth using a exible tape measure, whilst observed by a researcher, leading to questionable reliability. Limb girth is also affected by several variables beyond muscle hypertrophy, including water retention and adipose tissue. Future studies are advised to apply valid and reliable methods of body composition testing, such as muscle cross-sectional area magnetic resonance imaging (30,31) once the pandemic-related restrictions on clinical research are lifted.
A modest mean increase of 11.5 repetitions was observed between weeks two to ve. Total number of repetitions completed is a valid predictor of muscular endurance (24) and an improvement in muscular endurance would be expected when combining low load resistance training with BFR (4). The ceiling of the prescribed repetitions was 600 repetitions per week and eight participants were completing this maximum volume from week two, indicating that the prescribed training protocol was not an appropriate challenge for almost one third of the cohort. Participants were unable to access gym facilities and perform a more speci c protocol re ective of their baseline condition due the SARS-CoV-2 pandemic. Future studies should look to explore low load resistance training that is speci c to the individual, combined with BFR, and include a muscular endurance test using a xed percentage of one repetition maximum to volitional failure (32), to appropriately evaluate the effect of garment-integrated BFR on muscular endurance.

Interpretation
Garment-integrated BFR can be used safely in the upper limb of healthy adults when a subjective measure of compression stimulus is used. It is plausible that this safety outcome is generalisable to wider populations beyond healthy adults, but future trials should look to con rm this ahead of commencement if investigating a population beyond healthy adults. Recruitment, adherence, and data collection were all successful, meaning that a future larger scale trial is feasible. The feasibility of randomisation was not investigated in this study, and a future trial should therefore have appropriate stop/go criteria should initial randomisation prove infeasible.

Conclusion
Garment-integrated BFR is safe and feasible in the upper limb of healthy adults and can be applied in a resistance training setting using existing BFR protocols. Further work is required to investigate the e cacy of garment-integrated BFR and determine its comparability or superiority to existing BFR methods with respect to facilitating muscle hypertrophy, strength and endurance.

Declarations
Ethics approval and consent to participate Ethical approval was granted by the Queen Mary Ethics of Research Committee (QMREC2018/48/054).

Consent for publication
We consent for this manuscript to be published upon completion of peer-review.

Availability of data and material
Raw data can be provided upon request.  Figure 1 garment integrated BFR with Velcro mechanism