Feasibility study.

Patients were referred to the study by orthopedic specialists and physiotherapists working in orthopedic outpatient clinics at three public hospitals (Horsens Regional Hospital, Aarhus University Hospital, Silkeborg Regional Hospital).

The intervention and testing took place at two hospitals only (Horsens Regional Hospital and Aarhus University Hospital). The intervention was conducted by two physiotherapists, while one of these (MH) conducted all tests.

Inclusion criteria were (1) subjective complaints of lateral hip pain, (2) palpable tenderness or pain at the insertion point of the gluteus medius/minimus tendon at the greater trochanter, (3) a positive Single Leg Stance test (SLS), by reproduction of know lateral hip pain within 30 s of one leg stance, (4) lateral hip pain provoked by the FADER (Flexion Adduction External Rotation) and/or the FADER-R (Flexion Adduction External Rotation with resisted isometric internal rotation) test, (5) age 18–75 years and (6) ability to read and understand Danish. The diagnosis was based on the clinical tests described by Grimaldi et al. (2017). Exclusion criteria were (1) corticosteroid injection in the affected hip within the last 6 weeks prior to the intervention (2) any prior surgery in the affected hip (3) unregulated hypertension (≥ 180 mmHg/ ≥ 110 mmHg) (4) complaints or clinical signs of bilateral GT (5) MRI or X-ray verified osteoarthritis or (6) pregnancy. See Supplementary Material no. 1 for the clinical tests.

The study was conducted in accordance with the Declaration of Helsinki. The study was presented to the ethics committee of Central Region Denmark, who decided that no formal ethical approval was required (record number: 1-10-72-181-20). The Danish Data Protection Agency (record number: 1-16-02-548-20) approved the study and all patients gave written informed consent prior to inclusion.

The intervention had an overall duration of 8 weeks and consisted of exercises and patient education. Once a week, exercise sessions were performed at the hospital under supervision of the primary investigator (MH) or a trained physiotherapist (LCUR). Remaining exercises sessions were performed at home without supervision, yielding a total of eight supervised sessions and 48 sessions at home.

At baseline and at eight-week follow-up, patient-reported outcomes, two tests of physical function, isometric hip and knee muscle strength tests were completed.

The exercises chosen in this study were inspired by a previous study by Mellor et al. consisting of four to six exercises per session (Mellor et al., 2018). However, due to the duration of rest between sets and exercises using LL-BFR, only four exercises per session were chosen for this study. Static abduction, sidestepping, glute bridging, and bodyweight squats were chosen, as these exercises included functional retraining, strengthening of the hip and thigh muscles and control of adduction during function as proposed by Mellor et al. (2018).

Week 1 and 2, each training session consisted of the following: 5 × 5 s of standing static abduction, side-stepping; 10 steps to each side, 10 repetitions of glute bridging and 10 repetitions of squats. The exercises were to be performed once a day with bodyweight only. From week 3 to 8, static abduction and side-stepping were continuously performed daily. Glute bridging and squats were only performed every second day, and these two exercises were exclusively performed with the application of an 11.7 cm wide pneumatic BFR nylon cuff (Occlude Aps, Denmark) around the affected leg (Figure 1).Cuff pressure was 60% of the pressure required to fully restrict blood flow to the exercising limb. Given the relatively low volume in regards of repetitions of the BFR exercises from week 3 to 7, the restriction time was much shorter than suggested by Patterson et al. (5–10 min per exercise) (Patterson et al., 2019). That is, the squat and the bridging exercise combined, would until week 7 only last ~4 min. Thus, in order to reach the proposed restriction time, both the rest between exercises and a 30-s extra period after completion of the exercises were carried out. Given the increased number of repetitions, the participants released the pressure in the pause between the exercises in week 7 and 8. A LL-BFR session progressed from ~15 min in week 3 to ~25 min in the final 2 weeks. Repetitions in double leg bridging, and double leg squats were alternately increased by 10 repetitions every week until patients performed three sets of 20-10-10 repetitions. The full exercise programme with progressions is presented in Figure 2. The whole intervention is described according to the Template for Intervention Description and Replication (TIDieR) checklist (Hoffmann et al., 2014) in Supplementary Material No. 2. In case of illness during the intervention-period, which would result in absence from a supervised session, the progression for the following week and the patient education was managed by a telephone call. Throughout the intervention-period, patients were required to complete a daily training diary. In addition to reporting the number of repetitions, patients reported their perceived hip-related pain on a Numerical Rating Scale (NRS: 0-10) for every exercise and scored their perceived rating of exertion (RPE: 0-10) for the entire session. Diaries for the previous week were collected at the supervised sessions.

At baseline testing, all patients received written information on the anatomy of the gluteus muscles, common GT-symptoms, pain management, load management, appropriate movement patterns, and resting positions. Patients were encouraged to avoid hip adduction across midline, prolonged single leg stance, lying on the affected side, and to place a pillow between the knees during sleep, and rest on the non-affected side to reduce tension and compression of the gluteus medius muscle and tendon. During supervised sessions, patients were continuously educated to manage their GT-symptoms according to the written information received at baseline, and time was taken to verbally instruct and ensure the understanding of the information regarding tendon care and load and pain management.

To individualize the cuff pressure during the exercise session, measurement of the pressure required to fully restrict blood flow to the affected lower limb [limp occlusion pressure (LOP)] was determined prior to week 3. A rigid nylon cuff (Occlude Aps, Denmark) with a removable manometer was fitted around the proximal thigh on the affected leg. The posterior tibial artery was located to detect the auscultatory pulse using a vascular doppler probe (EDAN SD3, China). The cuff was gradually inflated until the auscultatory pulse was undetectable as described previously elsewhere (Jørgensen et al., 2020). During LL-BFR exercises a cuff pressure corresponding to 60% LOP was applied and remained constant throughout the entire intervention period. As LOP is affected by body position (Sieljacks et al., 2018), LOP was measured in both seated and supine reflecting the body position of the torso during mini-squats and glute bridging. Subsequently, patients were carefully instructed to apply and inflate the cuff correctly, and further to regulate the pressure between the exercises according to either the mini-squats or glute bridging, without totally deflating the cuff.

Baseline characteristics were described using numbers (n), proportions (%) for categorical data, and mean and standard deviations (SD) for continuous data if the values were normally distributed. Otherwise, median and interquartile range (25th−75th percentiles) were presented.

Adherence was calculated as the proportion of completed exercise sessions for both the overall number of sessions and for the LL-BFR-sessions alone. Formula used to calculate the adherence:

Changes in secondary outcome measures from baseline to follow-up were evaluated by using paired t-tests given the data were normally distributed, otherwise a Wilcoxon signed-rank test was used. Level of significance was set at 5%. Stata 16.0 (Statacorp, Texas, TX USA) was used for the statistical analysis. Proportion (%) of successful global improvements was presented using descriptive statistics.

The study was conducted from October 2020 to April 2021. Sixteen females were included, and 14 completed the intervention. Baseline characteristics are presented in Table 1. Two patients had their intervention period extended from 8–9 weeks (one due to COVID-19, and one due to excessive hip-related pain at the time of initiation of LL-BFR).

The mean (SD) adherence to all exercise sessions was 96% (5.76), with the patient with the lowest adherence displaying an adherence of 79%. The mean (SD) adherence to the LL-BFR-sessions only was 94% (8.52), with the patient with the lowest adherence achieving an adherence of 72%. One patient dropped out due to illness prior to the initiation of the LL-BFR. Another withdrew due to excessive pain during LL-BFR at home.

One adverse event, leading to a drop out, was registered during the intervention. The patient experienced a sudden and ongoing pain sensation during a LL-BFR session. The intensity of the pain was reported as 8-9 NRS. The pain was located directly beneath the LL-BFR-cuff and radiating throughout the leg alongside the affected leg. The patient counselled her general practitioner who advised her to stop her participation in the study. Two weeks later, the patient reported to have resumed her daily exercise sessions without the LL-BFR-cuff and did not experience pain with exercise.

Patient-reported hip pain, function, and quality of life from baseline to follow-up are presented in Table 2. Overall mean lateral hip pain (NRS) decreased significantly. Both VISA-G, EQ-5D-VAS, OHS and the majority of the subscales in HAGOS (5 out of 6) improved significantly. Nine out of 14 patients (64%) reported successful improvements on the GRoC. Mean (SD) hip pain during exercise throughout the intervention-period was 2.20 (1.43).

The mean number of repetitions in the 30s-CST as well as the mean power output in the SCT increased significantly from baseline to follow-up (Table 3).

Isometric muscle strength outcomes are presented in Table 3. Significant improvements of isometric strength were seen in the affected leg in both MVIC HA and MVIC KE, and in both legs in MVIC HE. The difference in strength change between the affected and unaffected leg was only significant in MVIC HA.

In the present study we observed an exercise adherence corresponding to 96%, which is higher than the adherence previously reported with daily exercise sessions planned for patients suffering from GT (Ganderton et al., 2018; Mellor et al., 2018). Mellor et al. reported a mean (SD) adherence at 88.8% (13.7) after 8 weeks of daily home-based/supervised low-load exercise intervention (Mellor et al., 2018). Ganderton et al. reported an adherence corresponding to 75.80% (23.49) for the intervention group after 12 weeks of a low-load exercise and 75.99% (25.35) for the control group, engaging in 12 weeks of a sham exercise involving primarily seated exercises with no external load (Ganderton et al., 2018). Recently, other homebased LL-BFR exercise protocols have emerged and demonstrated excellent adherence when applied in clinical populations (Kilgas et al., 2019; Petersson et al., 2020; Jørgensen and Mechlenburg, 2021). Both Petersson et al. (2020) and Jørgensen and Mechlenburg (2021) have reported adherence of 100% to a 5-week combined homebased and supervised BFR walking exercise protocol (three sessions/week, one of these supervised) in a patient suffering from knee OA and to a 12-week homebased LL-BFR intervention (exercise session every second day, supervision only during the first week) in a patient suffering from reactive arthritis, respectively (Petersson et al., 2020). In line with this, Kilgas et al. reported an adherence of 100% to a 4-week entirely homebased LL-BFR intervention (40 sessions in total) in a patient who had received a total knee arthroplasty (Kilgas et al., 2019). Thus, the adherence in this present study was consistent with previous studies, even though our study had more patients included.

To our knowledge, only a few smaller studies have investigated the feasibility of LL-BFR in clinical settings for patients with tendinopathies (Sata, 2005; Skovlund et al., 2020). A case series by Skovlund et al. investigated the feasibility of LL-BFR on patients with patellar tendinopathy (21). This study reported a 50% pain-reduction on NRS during a single-leg decline squat test, while the tendon vascularity was decreased by 31% after 3 weeks of low-load BFR resistance exercise (Skovlund et al., 2020). In line with these findings, a case report from 2005 observed reduced pain levels (without having reported the magnitude of this) and no thigh muscle atrophy after 3 weeks LL-BFR exercise in a patient showing patellar tendinitis (Sata, 2005). In both studies the adherence was comparable to the findings in the present study. Further, none of the studies reported any adverse events. However, both studies had few participants, and the exercise protocols consisted exclusively of supervised training sessions.

A concern regarding supervised exercise session in clinical populations is the limited flexibility in time, since both the patient and the supervisor has to have compatible timetables (Collado-Mateo et al., 2021). In this study, the supervised exercise sessions were performed at two sites to allow patients access to the supervised exercise session closer to either their home or worksite. Further, a session consisted of maximum four exercises (every second day from week 3) with no required exercise equipment beside the cuff, which was handed out at the hospitals at baseline. Additionally, in both the present study and the study by Mellor et al. patient education took place during the weekly, supervised sessions. This study design could enhance the patients understanding of their condition and expected benefit of the exercises (Collado-Mateo et al., 2021), resulting in a higher compliance to the protocol.

In the present study, one adverse event, leading to a drop out, was registered. There has been some discussion on the safety of LL-BFR, underreporting adverse events (Minniti et al., 2020) and the degree to which LL-BFR can cause serious muscle damage, such as exertional rhabdomyolysis (Wernbom et al., 2020). However, the patient reporting an adverse event in this study was not displaying symptoms related to rhabdomyolysis or other muscle damages, as the patient reported pain ease and continuation of daily exercise 2 weeks after the incident.

In general, LL-BFR seems to be a well-tolerated and a safe exercise modality in both healthy and clinical populations, when safety precautions concerning cuff application, cuff pressure, and time with blood flow restriction are taken (Hughes et al., 2017; Patterson et al., 2019).

Improvements in patient reported outcomes shown in this study were comparable to Mellor et al. and Ganderton et al., although our patient population was slightly younger and had a longer period of pain duration before inclusion (Ganderton et al., 2018; Mellor et al., 2018). Changes in HAGOS score after an intervention in GT patients has to our knowledge, not been reported before. Five of the six subscales on HAGOS showed significant improvements. Additionally, a mean change within the minimal important change (MIC) of 10-15 points were seen in these five subscales (Thorborg et al., 2011). Even though the patients experienced less pain and better functioning, the subscale “participation in physical activities” (PA) only improved by 7 points. However, 14 patients had a duration of symptoms for at least 7 months suggesting that fear-avoidance or habits may have changed their approach to physical activities. Further, the PA subscale only consist of two items which make it a challenge to achieve a minimal important change.

In the present study, hip abduction strength measured at follow-up was comparable to the abduction strength in the exercising group in the study by Mellor et al. at 8 weeks follow-up (Mellor et al., 2018). Further, we found that mean hip extension peak torque increased for both legs and knee extension mean peak torque increased on the affected leg. Interestingly, the strength deficits observed at baseline between the affected and non-affected leg was minimized at follow-up. These gains in strength are reflected in the functional capacity tests, where the number of repetitions in the 30s-CST and the power output used in the SCT improved, indicating better function. Previous research by Fearon et al., suggested pain to be the main driver to activity limitation following GT (Fearon et al., 2017). Hence, the clinically relevant reduction in pain observed in the present study may have contributed to improvements in functional performance and muscle strength.

Intuitively, the muscular adaptations to LL-BFR are expected to occur distal to the cuff. Nevertheless, Hedt et al. focused a review on proximal muscle responses to BFR training and suggested a potential benefit of the tissue directly proximal to the occlusion site (Hedt et al., 2022). The mechanisms for proximal benefits of BFR requires more research. However, the authors suggested that increased muscle activation of the muscle proximal to the cuff due to downstream fatigue, mechanotransduction signaling (due to muscle cell swelling), metabolite signaling during release of cuff pressure, and systemic anabolic signaling (i.e. insulin growth factor-1, growth hormone) as potential mediators of the adaptations proximal to the occlusion site. Even though studies on proximal effects after BFR training primarily have focused on the upper extremity (Hedt et al., 2022), a RCT study by Bowman et al. has, in line with the present study, observed increased isometric hip abduction muscle strength following 6 weeks of unilateral LL-BFR training (Bowman et al., 2019).