Physical Therapy in Sport 34 (2018) 121e128

Contents lists available at ScienceDirect

Physical Therapy in Sport

journal homepage: www.elsevier.com/ptsp

Low load resistance training with blood flow restriction decreases

anterior knee pain more than resistance training alone. A pilot
randomised controlled trial
Vasileios Korakakis a, b, *, Rodney Whiteley a, Giannis Giakas a
a
Aspetar, Orthopaedic and Sports Medicine Hospital, Doha, Qatar
b
Faculty of Physical Education and Sport Sciences, University of Thessaly, Trikala, Greece

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: To evaluate if application of blood flow restriction (BFR) combined with low-load resistance
Received 25 July 2018 training (LLRT) would induce significant anterior knee pain (AKP) reduction compared to LLRT alone.
Received in revised form Design: Randomised Controlled Trial.
17 September 2018
Setting: Institutional physiotherapy clinic.
Accepted 17 September 2018
Participants: Forty males suffering from AKP were randomly allocated in the LLRT-BFR or LLRT group. BFR
was applied at 80% of complete vascular occlusion. Four sets of open kinetic chain knee extensions were
Keywords:
implemented in both groups using a pain monitoring model.
Blood flow restriction
Occlusion
Main outcome measures: Pain (0e10) was assessed immediately after LLRT-BFR or LLRT application and
Resistance training after a physiotherapy session (45 min) during shallow and deep single-leg squat (SLSS, SLSD), and step-
Ischaemia down test (SDT).
Rehabilitation Results: Significant immediate pain reduction was found in LLRT-BFR group in SLSS, SLSD and SDT
(d ¼ 1.32, d ¼ 1.12, d ¼ 0.88 respectively), but no difference was found in LLRT group. Following the
physiotherapy session pain reduction was sustained in LLRT-BFR group in both SLSs and SDT (d ¼ 1.32,
d ¼ 0.78, d ¼ 0.89 respectively). For the control group significant pain reduction was only found in SLSS
(d ¼ 0.56). No significant between-group differences were observed.
Conclusions: The pain reduction induced by LLRT-BFR could indicate this intervention as a pre-
conditioning process prior to the rehabilitation of AKP.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction
Anterior knee pain (AKP) is a non-specific diagnosis often used
to describe patellofemoral disorders and miscellaneous pain in the
knee (Wood, Muller, & Peat, 2011). Patellofemoral pain prevalence
has been estimated to be as high as 40% (Witvrouw et al., 2014),
while it is a well-established complication following anterior cru-
ciate ligament reconstruction and/or meniscal surgery (Culvenor,
Cook, Collins, & Crossley, 2013). Evidence suggests that quadri-
ceps resistance training improves function and decreases patello-
femoral pain (Kooiker, Van De Port, Weir, & Moen, 2014). However,
the effect of exercises may be attenuated in the presence of pain
* Corresponding author. Orthopaedic and Sports Medicine Hospital, Doha, Qatar
Doha, PO Box29222, Qatar.
E-mail addresses: Vasileios.Korakakis@aspetar.com, vkorakakis@hotmail.com
(V. Korakakis).
where resistance training can't be performed with high loads (Chiu,
Wong, Yung, & Ng, 2012; Herrington & Al-Sherhi, 2007).
Recently, significant attention has been drawn to blood flow
restriction (BFR) training which involves low-load resistance
training (LLRT) combined with partial BFR (LLRT-BFR) by using
inflatable cuffs placed at the most proximal part of the exercised
limb. Evidence indicates a moderate effect of LLRT-BFR compared to
low-load resistance training on increasing muscle strength in in-
dividuals suffering musculoskeletal weakness (Hughes, Paton,
Rosenblatt, Gissane, & Patterson, 2017). In terms of musculoskel-
etal pain, preliminary research suggests that the lower load training
associated with BFR had comparable effectiveness with traditional
strengthening despite the reduced absolute load with BFR training
in patients with patellofemoral pain syndrome (Giles, Webster,
McClelland, & Cook, 2017). Interestingly, participants with painful
resisted knee extension in the BFR group had a significantly greater
increase in knee extensor torque than those in the standardised

https://doi.org/10.1016/j.ptsp.2018.09.007
1466-853X/© 2018 Elsevier Ltd. All rights reserved.
122 V. Korakakis et al. / Physical Therapy in Sport 34 (2018) 121e128
quadriceps strengthening group (Giles et al., 2017). Additionally, in
a non-controlled study (Korakakis, Whiteley, & Epameinontidis,
2018) a single LLRT-BFR bout reduced AKP immediately post-
exercise and the moderate to large effect was sustained for at
least 45 min. Importantly the pain reduction allowed the patients
to perform exercises with higher absolute training knee loads,
which were typically otherwise associated with symptoms.
Laboratory studies report that acute exercise (isometric, aerobic,
or resistance) reduces sensitivity to painful stimuli in healthy in-
dividuals with effects ranging from small to large depending on
pain induction method and exercise protocol (Naugle, Fillingim, &
Riley, 2012). Regarding the type and intensity of exercise the
hypoalgesic effect has been reported: i) moderate to large for iso-
metric exercise at a low to moderate intensity held for longer
duration (i.e. until task failure, >5 min), ii) moderate to large for
aerobic exercise when performed at moderate to high intensity
pace (i.e. 75% of VO2max) for relatively longer duration (>10 min),
and iii) small to large for dynamic resistance exercise when per-
formed at relatively high intensity (75% of 1 repetition maximum)
(Naugle et al., 2012). While in chronic pain populations the
magnitude and direction of the effect sizes were highly variable and
strongly depended on the condition and the intensity of the exer-
cise (Naugle et al., 2012). To our knowledge, limited reports with
conflicting results exist in the literature regarding lower limb
musculoskeletal conditions; focusing on patellar and Achilles ten-
dinopathy and assessing an acute exercise-induced (heavy 45-s
isometric contractions) hypoalgesic effect (O'Neill et al., 2018; Rio
et al., 2015; Rio et al., 2017).
Given that acute exercise alone can induce a hypoalgesic
response and the preliminary evidence of BFR-induced pain re-
ductions following LLRT, we hypothesised that the hypoalgesic ef-
fect LLRT would be augmented with the addition of BFR. Therefore,
the main objective of this randomised controlled trial was to
evaluate if application of LLRT-BFR would induce significant AKP
reduction compared to LLRT alone.
2. Methods
2.1. Study design
The present study was a pilot randomised controlled trial with
two intervention arms and conducted at the rehabilitation
department of our sports medicine hospital. Ethics approval and
informed consent were sought and obtained.
This study has been designed and reported in line with the
CONSORT recommendations for reporting randomised trials
(Fig. 1).
2.2. Sample size calculation
Recruitment followed a sample size calculation accounting for a
type-I error rate of 0.05, and a power of 95%. The assumption of an
effect size of 0.60 in pain reduction was based on data from a
previous study assessing effectiveness of LLRT-BFR in similar pop-
ulation (Korakakis et al., 2018).
2.3. Participants
We enrolled 40 male adult patients who had consulted a sports
medicine physician in a primary-care setting for AKP complaints.
Inclusion and exclusion criteria are described in detail in Fig. 1.
2.4. Randomisation and blinding
Patients who met the eligibility criteria were randomly allocated
into the LLRT-BFR group (n ¼ 20) or the LLRT group (n ¼ 20). A block
randomisation with randomly mixed block sizes was generated by
using a randomisation web site (http://www.randomization.com).
Random codes were sealed in opaque envelopes and envelopes
were delivered to the physiotherapist providing the intervention by
an administrative assistant not involved in the study. The physio-
therapist who performed the intervention was not involved in the
inclusion process, or any subsequent evaluation of the patient.
Initial assessment for eligibility and physical examination, outcome
assessment (at all follow-up sequence) and physical therapy ses-
sion were conducted by physiotherapists blinded to group alloca-
tion and previous evaluation data of the patient.
2.5. Initial evaluation, outcome measures and intervention group
allocation
A standardised history and physical examination was conducted
by a physiotherapist. Patients that met the inclusion criteria were
evaluated in three functional tests (shallow single leg squat - SLSS,
deep single leg squat - SLSD, and 20 cm step down - SDT) and were
asked to rate their pain on an 11-point numeric rating scale (NRS
0e10) as in a previous study (Korakakis et al., 2018). Upon
completion of the initial assessment and baseline testing the
physiotherapist contacted an administrative assistant who then
delivered the sealed randomisation envelope to the treating phys-
iotherapist shortly before intervention.
2.6. Blood flow restriction pressure calculation
We used partial vascular occlusion of the lower limb set at 80%
of complete arterial occlusion following a previously described
standardised procedure with aim to attain personalised BFR
(Korakakis et al., 2018). A wide BFR cuff (Sports Rehab Tourniquet©)
of 10 cm width and 116 cm length was attached to the proximal
thigh at the inguinal fold region while the patient was lying prone.
By using a vascular Doppler probe on the popliteal artery, the cuff
was inflated to the point at which the pulse was eliminated and this
complete occlusion pressure (mmHg) was recorded and the in-
dividual's 80% partial occlusion calculated for the intervention.
2.7. Interventions
The only difference between groups in LLRT was the application
or not of BFR. Before the initiation of exercise intervention all pa-
tients were familiarised with the exercise protocol. The LLRT con-
sisted of open kinetic chain knee extensions (90 -0 ) performed
slowly (2 s concentric, 2 s eccentric phase, paced by a metronome).
A modified pain monitoring approach (Korakakis et al., 2018;
Thomee, 1997) dictated the loading to a maximum of 5 kg (vari-
able adjustable ankle weights) such that the patient reported a
maximum of 4/10 pain during familiarisation. Patients were
instructed to complete four sets of open kinetic chain knee exten-
sions, where the first set to maximum possible repetitions followed
by three sets of 15 repetitions with 30 s rest between sets. Termi-
nation of the first set was indicated by: i) failure to shadow the pace
of the metronome or ii) failure to fully extend the knee joint
(Korakakis et al., 2018).
The physiotherapist stressed the importance that all repetitions
were performed at each set and in case that the patient was inca-
pable of finishing the repetitions the adjustable weights were
accordingly reduced.
The patients in the LLRT-BFR group had the BFR-cuff inflated
throughout the session.
 V. Korakakis et al. / Physical Therapy in Sport 34 (2018) 121e128
Inclusion criteria
Age >18 years
Enrollment
Diagnosis of anterior knee pain confirmed
by history, physical examination ± imaging
Pain during at least in 1/3 functional tests
BMI<25 or non-overweight individual
Unobstructed knee ROM and no pain in
passive end-range knee extension
Exclusion criteria
Hypertension (SAP≥140 mmHg)
Screening for eligibility
History of deep venous thrombosis (n=1)
History of endothelial dysfunction
Peripheral vascular disease (n=1)
Diabetes (n=6)
Infection
Spinal or referred pain
Bilateral knee symptoms (n=4)
Previous experience of BFR training
Overweight/obese (n=5)
Randomisation
Randomised
LLRT-BFR or LLRT (n=40)
Allocation
LLRT-BFR
Allocated and received allocated
intervention (n=20)
Follow-up
LLRT-BFR
Lost to follow-up (n=0)
LLRT-BFR
Analysis
Lost to follow-up (n=0)
Fig. 1. CONSORT statement flow diagram.; Abbreviations: BMI, body mass index; ROM, range of motion; SAP, systolic arterial pressure; BFR, blood flow restriction, LLRT, low load
resistance training.
2.8. Post-low load resistance training interventions and outcomes
assessment
Immediately following the intervention a physiotherapist blin-
ded to the group allocation of the patient and the previous pain
evaluation rated the participants’ pain at the same functional tests.
Subsequently, all participants independent of group allocation
underwent a standardised physiotherapy session by a blinded
physiotherapist, consisting of lower limb resistance training, core
and balance exercises. Finally, another blinded assessor re-
examined the functional tests and recorded pain levels.
2.9. Statistical analysis
Descriptive statistics (means, standard deviations, and per-
centages) were used to describe the characteristics of the
123
Patients referred for
physiotherapy
Patients in waiting-list referred
for physiotherapy with anterior
knee pain complaints by
physician (n=57)
Excluded (n=17)
History of deep venous
thrombosis (n=1)
Peripheral vascular disease
(n=1)
Diabetes (n=6)
Bilateral knee symptoms
(n=4)
Overweight/obese (n=5)
LLRT
Allocated and received allocated
intervention (n=20)
LLRT
Lost to follow-up (n=0)
LLRT
Lost to follow-up (n=0)
participants. Baseline characteristics of the two groups were
examined and compared with a student's t-test, along with esti-
mation of normality. A 2-way repeated measure ANOVA, with
group as the between-patients factor and time as the within-
patients factor, was used to evaluate significant differences in
pain at the time points (pre-intervention, post-intervention, and
45 min later). Between-group comparison was analysed on an
intention-to-treat basis. The level of significance was set at 0.05,
and the Bonferroni adjustment for multiple comparisons was set at
0.017. All analyses were conducted in SPSS v21 (IBM, Amarok, USA).
As measures of treatment effect we calculated effect sizes for
pain reduction and number needed to treat for an additional
beneficial outcome (NNT) as the reciprocal of the absolute risk
reduction (ARR) (Cook & Sackett, 1995). A standardised effect size
of pain reduction for within and between group differences was
calculated using Cohen's d and interpreted using thresholds of 0.2,
124 V. Korakakis et al. / Physical Therapy in Sport 34 (2018) 121e128

Table 1
Baseline characteristics of the participants.

Characteristic LLRT-BFR group LLRT group P value of difference (95% CI)

Patient (n) 20 20 n/a

Age (years) 29.1 ± 6.6 29.7 ± 7.6 0.776 (5.2 to 3.9)
Height (cm) 179.8 ± 5.1 179.2 ± 4.3 0.701 (2.4 to 3.6)
Weight (kg) 77.0 ± 9.6 73.7 ± 5.5 0.189 (1.7 to 8.3)
Duration (m)(range) 3.2 ± 2.8 (1e12) 4.3 ± 3.3 (1e13) 0.297 (3.0 to 0.9)
Baseline pain (NRS)
Single leg squata 4.6 ± 2.3 3.8 ± 2.3 0.309 (0.7 to 2.2)
Single leg squatb 5.6 ± 2.6 5.1 ± 1.8 0.480 (0.9 to 1.9)
Ste-down test 4.2 ± 2.4 4.1 ± 2.6 0.851 (1.4 to 1.7)
Anterior knee pain diagnostic category (n/%)
Patellofemoral pain 10/50% 9/45% n/a
Post-operative ACL AKP 7/35% 8/40% n/a
Meniscal repair/removal 3/15 3/15% n/a

Values are presented as mean ± standard deviation unless otherwise indicated.

Abbreviations: LLRT: low load resistance training; BFR: blood flow restriction; 95%CI: 95% confidence intervals; n: numbers; cm: centimetres; kg: kilograms; m: months; NRS:
numeric rating scale; AKP: anterior knee pain; n/a: not applicable.
a
Shallow - To the mid-point between greater trochanter and popliteal crease.
b
Deep - To the level of the popliteal crease.

0.5, 0.8 for small, moderate and large effect, respectively (Cohen,
1988). We aimed to estimate clinically meaningful effects of the
intervention in patient rated terms, (Farrar, Young, LaMoreaux,
Werth, & Poole, 2001) therefore the substantial benefit in ARR
calculation was set at 30% reduction from baseline pain.
Individual participant pain data for all three functional tests
were also presented for clinical interpretation of the results.
3. Results
No significant differences were observed between LLRT-BFR and
LLRT groups for any of the variables assessed at baseline (p > 0.05,
Table 1).
3.1. Within-group differences
Significant immediate pain reduction with large to very large
effect sizes were found in the LLR-BFR group in SLSS, SLSD and SDT
(all P < 0.001; d ¼ 1.32, d ¼ 1.12, d ¼ 0.88 respectively), but no sig-
nificant differences (all P > 0.017) were found in LLRT group where
the effect sizes were small to moderate (Table 2 and Fig. 2).
Following the physiotherapy session (45 min) pain reduction
was sustained in LLRT-BFR group in both SLS, and SDT (P < 0.001;
d ¼ 1.32, d ¼ 0.78, d ¼ 0.89 respectively). For the LLRT group sig-
nificant pain reduction was only found for the SLSS (d ¼ 0.56), but
not for SLSD and SDT (Table 2 and Fig. 2).
3.2. Between-group differences
No significant between-group differences were found immedi-
ately post intervention or following the physiotherapy session
(Table 2). The effect sizes calculated were very small to moderate
(range d ¼ 0.06e0.35), except the SLSD immediately post-
intervention that presented a non-significant moderate to large
effect size (d ¼ 0.61). There was a 117%, 237% and 82% greater
reduction in pain immediately post-intervention in the LLRT-BFR
group relative to the LLRT group in SLSs, SLSD and SDT, respec-
tively. Also, there was a 92%, 73% and 82% greater reduction in pain
post physiotherapy session in the LLRT-BFR group relative to the
LLRT group in SLSs, SLSD and SDT, respectively.
The NNT calculated are presented in Table 3 and the individual
participant data for pain in functional tests used as outcome
measures are presented in Fig. 3. No participant's symptoms in
LLRT-BFR worsened compared to baseline at all follow-up points in
both SLSs and SDT, except one participant in SDT immediately post-
intervention (0.5/10 increase). In the LLRT group for each individual
test 20% of the participants reported increased symptoms post-
intervention and 15% following the physiotherapy session.
4. Discussion
LLRT-BFR significantly reduced pain in functional activities
immediately post-intervention, and this was sustained for at least
45 min, with large effect sizes, an effect not seen in the LLRT group.
There were no non-responders in the LLRT-BFR group in contrast to
the LLRT group where 20% of the participants reported symptoms
worsening post-intervention. We suggest that LLRT-BFR may be
used to reduce knee pain before rehabilitation, providing clinicians
a window of opportunity for knee loading exercises, which are
typically otherwise associated with symptoms. The variability of
clinical presentations in terms of broad pain severity and different
pathologic categorisations suggests this may be a generalizable
finding. We note that no patients in LLRT-BFR group reported any
adverse effect during this trial.
4.1. LLRT-BFR and LLRT independently reduce pain post-
intervention
Low load exercise with or without BFR reduced pain, however
only the effect of LLRT-BFR was significant and large. Evidence
suggests that acute exercise reduces sensitivity to painful stimuli in
healthy individuals and specific chronic pain populations (Naugle
et al., 2012), however there is limited and/or conflicting evidence
for painful lower limb musculoskeletal conditions (Korakakis et al.,
2018; O'Neill et al., 2018; Rio et al., 2015; Rio et al., 2017). Laboratory
studies measuring pain intensity in healthy individuals using
resistance exercise reported large effect sizes immediately post-
exercise (similar to the current study) and small effects 15 min
post-exercise (Focht & Koltyn, 2009; Koltyn & Arbogast, 1998). By
observing the individual data (Fig. 3) the additive effect of BFR
seemed to be sustained 45 min later compared to LLRT group that
presented variable responses following the physiotherapy session.
Both studies used load equal to 75% of 1RM, a load not applicable to
clinical populations with AKP as usually aggravation of symptoms
forces clinicians to reduce resistance load (Chiu et al., 2012;
Herrington & Al-Sherhi, 2007). To our knowledge, only one study
(Korakakis et al., 2018) in AKP patients has examined the acute pain
response to LLRT-BFR reporting moderate effect sizes. The broader
 V. Korakakis et al. / Physical Therapy in Sport 34 (2018) 121e128 125

Table 2
Mean NRS pain score at baseline, immediately post intervention, and following the physical therapy session and within- and between-group differences.

Shallow Single Leg Squat (SLSS)

Baseline Immediately post intervention Post physical therapy session

Mean ± SD Mean ± SD DW (95%CI) P value/d Mean ± SD DW (95%CI) P value/d

LLRT-BFR 4.6 ± 2.3 2.0 ± 1.6 2.6 (1.5e3.7) <0.001/1.32 2.0 ± 1.5 2.5 (1.5e3.6) <0.001/1.32
LLRT 3.8 ± 2.3 2.6 ± 2.7 1.2 (0.1e2.3) 0.028/0.48 2.5 ± 2.3 1.3 (0.2e2.4) 0.013/0.56
DВ (95%CI) 0.75 (0.7 to 2.2) 0.6 (2.0 to 0.8) 0.5 (1.8 to 0.8)
P value/d 0.309/0.32 0.374/0.29 0.428/0.26

Deep Single Leg Squat (SLSD)

Baseline Immediately post intervention Post physical therapy session

Mean ± SD Mean ± SD DW (95%CI) P value/d Mean ± SD DW (95%CI) P value/d

LLRT-BFR 5.6 ± 2.6 2.9 ± 2.3 2.7 (1.6e3.8) <0.001/1.12 3.7 ± 2.3 1.9 (0.8e3.0) <0.001/0.78
LLRT 5.1 ± 1.8 4.2 ± 2.2 0.8 (0.2 to 1.9) 0.167/0.43 4.0 ± 2.2 1.1 (0.1e2.2) 0.055/0.55
DВ (95%CI) 0.5 (0.9 to 1.9) 1.3 (2.8 to 0.1) 0.3 (1.7 to 1.1)
P value/d 0.480/0.23 0.052/0.61 0.674/0.13

Step-down test (SDT)

Baseline Immediately post intervention Post physical therapy session

Mean ± SD Mean ± SD DW (95%CI) P value/d Mean ± SD DW (95%CI) P value/d

LLRT-BFR 4.2 ± 2.4 2.2 ± 2.2 2.0 (0.8e3.2) 0.001/0.88 2.2 ± 2.1 2.0 (0.9e3.1) <0.001/0.89
LLRT 4.1 ± 2.6 3.0 ± 2.5 1.0 (0.2 to 2.3) 0.112/0.41 2.9 ± 2.2 1.1 (0.1e2.2) 0.033/0.47
DВ (95%CI) 0.1 (1.4 to 1.7) 0.8 (2.3 to 0.7) 0.7 (2.1 to 0.7)
P value/d 0.851/0.06 0.281/0.35 0.316/0.32

Note: Statistically significant differences are indicated in bold.

Abbreviations: NRS: numeric rating scale; SD: standard deviation; DW: within-group difference; 95%CI: 95% confidence intervals; d: Cohen's d; LLRT: low load resistance
training; BFR: blood flow restriction; DB: between-group difference.

Shallow Single Leg Squat Deep Single Leg Squat Step-down Test
10 10 10
9 9 d = 0.78 9
8 d = 1.32
8 d = 1.12 8
d = 0.89
7 d = 1.32
7 7
d = 0.88
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
Pre Post 45m Pre Post 45m Pre Post 45m Pre Post 45m Pre Post 45m Pre Post 45m
LLRT-BFR LLRT LLRT-BFR LLRT LLRT-BFR LLRT

Fig. 2. Pain reduction and effect sizes following interventions in LLRT-BRF and LLRT groups in the functional tests used as outcome measures at the three time points (pre-
intervention, immediately post-intervention, and 45 min). In all tests numerical rating pain scale was improved immediately after intervention and sustained following a physical
therapy session only in the LLRT-BFR group, but no significant between-group effects were found. Brackets denote significant difference to pre-intervention (p < 0.017).; Abbre-
viations: LLRT: low load resistance training; BFR: blood flow restriction; Pre: baseline pain; Post: post-intervention pain; 45 m: pain following a physical therapy session lasted
45 min.

inclusion criteria and the sample size configuration used in the
previous study may explain the difference in effect sizes reported
here. Potentially the pain reduction induced by LLRT-BFR and
documented by individual patient data could indicate this inter-
vention as a preconditioning process prior to the rehabilitation
session.
Is it the BFR or the low load and the pain monitoring system that
augmented the effect of exercise?
We aimed to challenge the knee with minimal load by using a
pain monitoring approach dictated the loading to a maximum of
5 kg, such that the patient reported a maximum of 4/10 pain during
training. The message transferred to the participants by this
approach e that has several clinical implications - was that exercise
activity that does not cause pain above 4 of 10 is safe. In previous
studies (Silbernagel, Thomee, Eriksson, & Karlsson, 2007; Thomee,
1997) the use of a pain monitoring approach has been advocated as
resulting in avoidance of excessively heavy tissue loading. In the
present study this assumption was partially supported by the re-
sults. Both groups were comparable at baseline, followed the same
pain monitoring system and loading, and presented no between-
group differences; nevertheless, only some of the LLRT group the
exercise had worsening of their pain (Fig. 3). It should be noted that
between-group differences in individual patients’ characteristics
and confounders other than those examined here might have
126 V. Korakakis et al. / Physical Therapy in Sport 34 (2018) 121e128

Table 3 LLRT group had the same conditions with patients that were
Treatment effect expressed as number needed to treat. benefited in the LLRT-BFR group. A plausible explanation can be
Outcome Immediately post Post physical therapy session attributed to the addition of BFR that, through unclear and
intervention currently unexplored mechanisms, was found to induce a beneficial
Substantial benefit 30% Substantial benefit 30% effect.
LLRT-BFR LLRT NNT LLRT-BFR LLRT NNT

SLSS 17/20 9/20 2.5 16/20 9/20 2.9 4.1.1. BFR-induced pain reduction may be evident but the
SLSD 15/20 9/20 3.3 9/20 8/20 20.0 mechanisms of action are unclear
SDT 15/20 9/20 3.3 12/20 8/20 5.0 BFR training is a relatively novel approach in rehabilitation of
Substantial pain reduction was based on at least 30% pain reduction from baseline musculoskeletal clinical conditions (Hughes et al., 2017). Some
pain. Patient benefitted from interventions are presented as absolute numbers. mechanisms of action in terms of muscle strength improvement
Abbreviations: LLRT: low load resistance training; BFR: blood flow restriction; SLSS:
and size gain have been explored (Hwang & Willoughby, 2017); but
shallow single leg squat; SLSD: deep single leg squat; SDT: 20 cm step down test;
NNT: number needed to treat.
no study has examined the mechanism of BFR training inducing
acute pain reduction. Recently, it has been suggested that the
mechanisms may include: i) conditioned pain modulation through
driven this outcome. From a clinical perspective, observation of the
the diffuse noxious conditioning controls (DNIC)-like effect, ii) ex-
pain reduction pattern at the individual level does not indicate the
ercise related release of endogenous substances which inhibit
presence of sample/condition confounders, nor “washed out ef-
nociceptive pathways, and iii) induced hypoxia following BFR
fects” due to the heterogeneity of AKP. Both groups had equally
training (Korakakis et al., 2018). We cannot discount the possibility
distributed clinical conditions and the patients worsened in the
that the effects seen are the result of the placebo effect, given that

LLRT-BFR group LLRT group

5
10 10
9 9
8 8
7 7
6 6
5 5
4 4
3 3
2 2
1 1
0 0
Pre Post 45min Pre Post 45min
10 10
9 9
8 8
7 7
6 6
5 5
4 4
3 3
2 2
1 1
0 0
Pre Post 45min Pre Post 45min

10 10
9 9
5
8 8
7 7
6 6
5 5
4 4
3 3
2 2
1 1
0 0
Pre Post 45min Pre Post 45min

Fig. 3. Individual participant data for the pain in functional tests used as outcome measures at the three time points (pre-intervention, immediately post-intervention, and 45 min).
Note due to the same numerical rating pain scores reported for different participants some lines in the graphs overlap. Cases that did not improve or worsened immediately
following intervention are shown in red, and those that improved are shown in grey.; Abbreviations: LLRT: low load resistance training; BFR: blood flow restriction; Pre: baseline
pain; Post: post-intervention pain; 45 m: pain after a 45 min physical therapy session.
 V. Korakakis et al. / Physical Therapy in Sport 34 (2018) 121e128
the “control” group in the study underwent LLRT. Evidence sug-
gests that in RCTs the placebos are “near inert” as their analgesic
effect is very small; 3.2 points on a 100-point pain scale (Kamper,
Machado, Herbert, Maher, & McAuley, 2008). The current study
showed no subjects to have increased pain with exercise in the
presence of BFR, whereas the exercise group had a more variable
response e although on average the group's pain was reduced.
Clinically the implication here is that BFR may be a more reliable
method of reducing pain such that therapeutic loading can be
applied than simple low load exercise.
Recent research (Giles et al., 2017) in patellofemoral pain
patients assessed the effectiveness of low-load BFR strength-
ening (30% of 1RM, 1 set of 30 repetitions and 3 sets of 15 rep-
etitions) compared to usual strengthening (approximately 70% of
1RM, 3 sets of 7e10 repetitions) in pain reduction with activities
of daily living (ADL) among other outcome measures. They re-
ported a 93% greater relative reduction in ADL pain at BFR
compared to usual strengthening group over the 8 weeks of
intervention. While this group did not evaluate the acute effect
on pain, but these results were suggestive that BFR may have a
potential acute effect on pain that was not evident at the follow-
up at 6 months where the BFR intervention had stopped. In the
present study pain reduction may partially explain the fact that
almost none of the participants in the LLRT-BFR group reported
worse pain from baseline, despite the knee loading during the
physiotherapy session - an observation that was not evident in
the LLRT group. The pain reductions in the LLRT-BFR group
relative to the LLRT group in all clinical tests used in the present
study, both immediately post-intervention and at 45 min,
ranging from 73% to 237%.
4.1.2. Limitations and future research
It cannot be excluded in our study that the lack of between-
group differences in pain reduction represents a type II error
due to sample size. The present study employed a heterogeneous
group of patients despite the robust inclusion criteria, all of whom
were recreationally active males. Different outcomes may be
found for sub-groups of pathologies, different activity levels, or
females. Despite the randomised controlled design of the study
the control group also received active intervention, and we have
not assessed follow-up beyond the immediate period of the
intervention. In addition, the occlusion pressure was calculated
with the subject prone - recent research (Hughes et al., 2018;
Sieljacks, Knudsen, Wernbom, & Vissing, 2018) indicates that
body position will alter this pressure, and it has been suggested
that the training pressure should be estimated by testing in the
same position. The current research has not strictly controlled for
the total volume of exercise. Future research should consider time
under tension and/or total load as independent factors which may
influence any hypoalgesic effect independent of BFR. Also, future
studies should include larger sample sizes and a control sham BFR
intervention with the aim to evaluate the magnitude of any pla-
cebo effect. We suggest that an important area of future research
is determining the mechanisms of action, particularly if they are
central or peripheral. If the mechanisms are central, for example,
BFR might be successfully applied away from the injured limb. If
the mechanisms are completely local (tissue-related), then the
BFR parameters (pressure, time, load volume) need to be explored
to allow optimisation of the intervention. Additionally, there may
be sub-groups of tissue pathology for whom BFR has a differential
response.
Conflicts of interest and source of funding
None declared.
127
Ethics approval
Granted from Anti-Doping Lab Qatar (ADLQ e ethics board).
Contributors
VK and RW contributed to the conception and design of the
study. VK performed BFR assessment and protocol, extracted and
analysed the data, and wrote the manuscript. RW and YG critically
revised and edited the manuscript for intellectual content.
Funding
None declared.
Provenance and peer review
Not commissioned; externally peer reviewed.
Acknowledgements
The authors would like to thank the following physiotherapist
for their contribution in the study: Andrew Cole, Pedro Nunnes,
Matthew Azzopardi, and Abdallah Itani.
References
Chiu, J. K., Wong, Y. M., Yung, P. S., & Ng, G. Y. (2012). The effects of quadriceps
strengthening on pain, function, and patellofemoral joint contact area in per-
sons with patellofemoral pain. American Journal of Physical Medicine & Reha-
bilitation, 91(2), 98e106. https://doi.org/10.1097/PHM.0b013e318228c505.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New
Jersey: Lawrence Erlbaum Associates.
Cook, R. J., & Sackett, D. L. (1995). The number needed to treat: A clinically useful
measure of treatment effect. BMJ, 310(6977), 452e454.
Culvenor, A. G., Cook, J. L., Collins, N. J., & Crossley, K. M. (2013). Is patellofemoral
joint osteoarthritis an under-recognised outcome of anterior cruciate ligament
reconstruction? A narrative literature review. British Journal of Sports Medicine,
47(2), 66.
Farrar, J. T., Young, J. P., Jr., LaMoreaux, L., Werth, J. L., & Poole, R. M. (2001). Clinical
importance of changes in chronic pain intensity measured on an 11-point nu-
merical pain rating scale. Pain, 94(2), 149e158. https://doi.org/10.1016/s0304-
3959(01)00349-9.
Focht, B. C., & Koltyn, K. F. (2009). Alterations in pain perception after resistance
exercise performed in the morning and evening. The Journal of Strength &
Conditioning Research, 23(3), 891e897. https://doi.org/10.1519/
JSC.0b013e3181a05564.
Giles, L., Webster, K. E., McClelland, J., & Cook, J. L. (2017). Quadriceps strengthening
with and without blood flow restriction in the treatment of patellofemoral
pain: A double-blind randomised trial. British Journal of Sports Medicine, 51(23),
1688e1694. https://doi.org/10.1136/bjsports-2016-096329.
Herrington, L., & Al-Sherhi, A. (2007). A controlled trial of weight-bearing versus
non-weight-bearing exercises for patellofemoral pain. Journal of Orthopaedic &
Sports Physical Therapy, 37(4), 155e160. https://doi.org/10.2519/
jospt.2007.2433.
Hughes, L., Jeffries, O., Waldron, M., Rosenblatt, B., Gissane, C., Paton, B., et al. (2018).
Influence and reliability of lower-limb arterial occlusion pressure at different
body positions. PeerJ, 6, e4697. https://doi.org/10.7717/peerj.4697.
Hughes, L., Paton, B., Rosenblatt, B., Gissane, C., & Patterson, S. D. (2017). Blood flow
restriction training in clinical musculoskeletal rehabilitation: A systematic re-
view and meta-analysis. British Journal of Sports Medicine, 51(13), 1003e1011.
https://doi.org/10.1136/bjsports-2016-097071.
Hwang, P., & Willoughby, D. S. (2017). Mechanisms behind blood flow restricted
training and its effect towards muscle growth. The Journal of Strength & Con-
ditioning Research. https://doi.org/10.1519/jsc.0000000000002384.
Kamper, S. J., Machado, L. A., Herbert, R. D., Maher, C. G., & McAuley, J. H. (2008).
Trial methodology and patient characteristics did not influence the size of
placebo effects on pain. Journal of Clinical Epidemiology, 61(3), 256e260. https://
doi.org/10.1016/j.jclinepi.2007.03.017.
Koltyn, K. F., & Arbogast, R. W. (1998). Perception of pain after resistance exercise.
British Journal of Sports Medicine, 32(1), 20e24.
Kooiker, L., Van De Port, I. G., Weir, A., & Moen, M. H. (2014). Effects of physical
therapist-guided quadriceps-strengthening exercises for the treatment of
patellofemoral pain syndrome: A systematic review. Journal of Orthopaedic &
Sports Physical Therapy, 44(6). https://doi.org/10.2519/jospt.2014.4127,
391eb391.
128 V. Korakakis et al. / Physical Therapy in Sport 34 (2018) 121e128
Korakakis, V., Whiteley, R., & Epameinontidis, K. (2018). Blood Flow Restriction
induces hypoalgesia in recreationally active adult male anterior knee pain pa-
tients allowing therapeutic exercise loading. Physical Therapy in Sport, 32,
235e243. https://doi.org/10.1016/j.ptsp.2018.05.021.
Naugle, K. M., Fillingim, R. B., & Riley, J. L., 3rd (2012). A meta-analytic review of the
hypoalgesic effects of exercise. The Journal of Pain, 13(12), 1139e1150. https://
doi.org/10.1016/j.jpain.2012.09.006.
O'Neill, S., Radia, J., Bird, K., Rathleff, M. S., Bandholm, T., Jorgensen, M., et al. (2018).
Acute sensory and motor response to 45-s heavy isometric holds for the plantar
flexors in patients with Achilles tendinopathy. Knee Surgery, Sports Traumatol-
ogy, Arthroscopy. https://doi.org/10.1007/s00167-018-5050-z.
Rio, E., Kidgell, D., Purdam, C., Gaida, J., Moseley, G. L., Pearce, A. J., et al. (2015).
Isometric exercise induces analgesia and reduces inhibition in patellar ten-
dinopathy. British Journal of Sports Medicine. https://doi.org/10.1136/bjsports-
2014-094386.
Rio, E., van Ark, M., Docking, S., Moseley, G. L., Kidgell, D., Gaida, J. E., … Cook, J.
(2017). Isometric contractions are more analgesic than isotonic contractions for
patellar tendon pain: An in-season randomized clinical trial. Clinical Journal of
Sport Medicine, 27(3), 253e259. https://doi.org/10.1097/
jsm.0000000000000364.
Sieljacks, P., Knudsen, L., Wernbom, M., & Vissing, K. (2018). Body position in-
fluences arterial occlusion pressure: Implications for the standardization of
pressure during blood flow restricted exercise. European Journal of Applied
Physiology, 118(2), 303e312. https://doi.org/10.1007/s00421-017-3770-2.
Silbernagel, K. G., Thomee, R., Eriksson, B. I., & Karlsson, J. (2007). Continued sports
activity, using a pain-monitoring model, during rehabilitation in patients with
Achilles tendinopathy: A randomized controlled study. The American Journal of
Sports Medicine, 35(6), 897e906. https://doi.org/10.1177/0363546506298279.
Thomee, R. (1997). A comprehensive treatment approach for patellofemoral pain
syndrome in young women. Physical Therapy, 77(12), 1690e1703.
Witvrouw, E., Callaghan, M. J., Stefanik, J. J., Noehren, B., Bazett-Jones, D. M.,
Willson, J. D., … Crossley, K. M. (2014). Patellofemoral pain: Consensus state-
ment from the 3rd international patellofemoral pain research retreat held in
vancouver, september 2013. British Journal of Sports Medicine, 48(6), 411e414.
https://doi.org/10.1136/bjsports-2014-093450.
Wood, L., Muller, S., & Peat, G. (2011). The epidemiology of patellofemoral disorders
in adulthood: A review of routine general practice morbidity recording. Primary
Health Care Research & Development, 12(2), 157e164. https://doi.org/10.1017/
s1463423610000460.