Received: 16 July 2019 

|
  Accepted: 9 December 2019

DOI: 10.1111/sms.13613

ORIGINAL ARTICLE

Patellofemoral pain: One year results of a randomized trial

comparing hip exercise, knee exercise, or free activity

Alexandra Hott1   | Jens Ivar Brox2,3  | Are Hugo Pripp4  | Niels Gunnar Juel3  |

Sigurd Liavaag5

1
Department of Physical Medicine
and Rehabilitation, Sørlandet Hospital
Objective: Extended follow-up of a randomized trial comparing hip-focused exer-
Kristiansand, Kristiansand, Norway cise, knee-focused exercise, and free physical activity in patellofemoral pain (PFP).
2
Faculty of Medicine, University of Oslo, Methods: A single-blind randomized controlled trial included 112 patients aged
Oslo, Norway
16-40 years (mean 27.6 years) with a clinical diagnosis of PFP ≥3 months (mean
3
Department of Physical Medicine and
39 months) and pain ≥3/10 on a Visual Analog Scale. Patients were randomized to
Rehabilitation, Oslo University Hospital,
Oslo, Norway a 6-week exercise-based intervention consisting of either isolated hip-focused exer-
4
Oslo Centre for Biostatistics and cises (n = 39), traditional knee-focused exercise (n = 37), or free physical activity
Epidemiology, Oslo University Hospital, (n  =  36). All patients received the same patient education. The primary outcome
Oslo, Norway
5
measure was the Anterior Knee Pain Scale (AKPS, 0-100). Secondary outcomes
Department of Orthopedic Surgery,
Sørlandet Hospital Kristiansand, were usual and worst pain, Tampa Scale of Kinesiophobia, Knee Self-Efficacy Score,
Kristiansand, Norway Euro-Qol (EQ-5D-5L), step-down test, and isometric strength. Blinded observers as-
sessed outcomes at baseline, 3, and 12 months. The study was designed to detect a
Correspondence
Alexandra Hott, Department of Physical difference in AKPS >10 at 12 months.
Medicine and Rehabilitation, Sørlandet Results: After 1 year, there were no significant between-group differences in any
Hospital Kristiansand, Norway.
Email: alexandra.hott@sshf.no
primary or secondary outcomes. Between-group differences for AKPS were as fol-
lows: knee versus free physical activity −4.3 (95% CI −12.3 to 3.7); hip versus free
Funding information
This study was funded by The Research
physical activity −1.1 (95% CI −8.9 to 6.7); and hip versus Knee 3.2 (95% CI −4.6
Department of Sørlandet Hospital. to 11.0). The cohort as a whole improved significantly at 3 and 12 months compared
to baseline for all measures except for knee extension strength.
Conclusion: After 1 year, there was no difference in effectiveness of knee exercise,
hip exercise, or free physical activity, when combined with patient education in PFP.

KEYWORDS
anterior knee pain, exercise therapy, hip strengthening, patellofemoral pain syndrome, patient education

1  |   IN T RO D U C T ION Patellofemoral pain is a clinical diagnosis defined by

Patellofemoral pain (PFP) is a common cause of pain in the
lower extremity. Its estimated annual prevalence is 23% in
the general population and 29% in adolescents, with females
approximately twice as commonly affected as males.1
Location: This study was performed at the department of Physical
Medicine and Rehabilitation, Sørlandet Hospital, Kristiansand, Norway.
pain around or behind the patella during loading, in the
absence of other specific pathology. The etiology of PFP
is incompletely understood. Current theories postulate
that anatomical, biomechanical, psychosocial, and behav-
ioral factors contribute to PFP.2 Pathomechanics of the
patellofemoral joint, influenced by factors including mus-
cle strength and control, are thought to be important.2,3

Scand J Med Sci Sports. 2020;30:741–753. wileyonlinelibrary.com/journal/sms © 2019 John Wiley & Sons A/S.     741 |
Published by John Wiley & Sons Ltd
 |
742       HOTT et al.
Psychological factors and pain sensitization are likely also
important in PFP.4,5
Exercise is considered a mainstay in treating PFP.6,7
Traditionally, exercise for PFP has been knee-focused
with emphasis on strength and coordination in quadriceps
muscles,8 while hip-focused exercise has received increas-
ing attention more recently.8,9 The most recent Cochrane
systematic review concludes with consistent but very
low-quality evidence for exercise as a treatment for PFP,
and insufficient evidence to recommend a specific type of
exercise.8 Although expert recommendations advocate the
combined use of hip and knee exercise, the relative con-
tributions of the components are unclear.6,7 The results of
studies investigating hip-focused exercises compared to
knee-focused exercise are inconsistent, with some studies
reporting superior short- and/or long-term effects of hip
exercise compared to knee exercises,10,11 while others re-
port no difference.12,13
Current recommendations also consider patient educa-
tion an important component of PFP treatment,6,7 although
a few studies exist. Two studies suggest that patient educa-
tion may be equally effective in reducing pain and function
as exercise-based interventions for PFP,14,15 while others find
exercise to be superior to education alone.16,17 Factors such
as kinesiophobia and catastrophizing may be important to
address in PFP.4,18,19 Patient education of this type has been
used successfully in other chronic pain conditions such as
low back pain,20 but has not been studied in PFP.
We have previously reported no significant differences in
short-term effectiveness of isolated hip exercise to traditional
knee exercise or a control group performing free physical
activity in PFP.21 Long-term outcomes within PFP are of
special interest, as newer surveys suggest higher degrees of
chronicity in PFP than has previously been appreciated.1,22,23
The goal of the present study was therefore to assess the long-
term effectiveness of isolated hip exercises, traditional knee
exercise, or free physical activity, through an extended fol-
low-up of a randomized trial for which the 3-month results
are previously reported.21
2  |  M AT E R IA L S A N D ME T HODS
2.1  |  Study design
This is an extended follow-up of a randomized controlled
trial approved by the Ethics Committee Health Region
South-East, Norway (reference number: 2013/1860/
REKsør-øst) and registered with the ClinicalTrials.gov
database (reference number: NCT02114294). Recruitment
took place between September 2014 and September 2017.
The full-trial protocol and 3-month results have been pub-
lished previously.21,24
2.2  |  CONSORT statement and flow chart
This trial has been designed and reported in accordance with
the CONSORT guidelines for reporting of randomized con-
trolled trials.25
2.3  | Patients
The study was performed at the Department of Physical
Medicine and Rehabilitation at Sørlandet hospital. Patients
anticipated to have PFP were referred by primary care phy-
sicians or by other medical specialists (eg, Orthopedic sur-
geons, Rheumatologists). Patients were screened with x-ray,
MRI, and clinical exam by a specialist in physical medicine
and rehabilitation.
2.3.1  |  Inclusion criteria
Eligibility criteria were as follows: age 16-40  years, mini-
mum 3 months history of PFP, pain score for worst pain in-
tensity during previous week of 3 or more on a Visual Analog
Scale (0-10, most pain), reproduced by at least two of the
following activities: Stair ascent or descent, hopping, run-
ning, prolonged sitting, squatting or kneeling and present on
at least one of the following clinical tests: Compression of
the patella or palpation of the patellar facets. In patients with
bilateral pain, the worst knee was included.
2.3.2  |  Exclusion criteria
Clinical, x-ray, or MRI findings indicative of other specific
pathology including osteoarthritis, meniscal, ligament or
cartilage injury, apophysitis, recurrent patellar subluxation
or dislocation, significant knee joint effusion; hip or back
pain interfering with ability to perform prescribed exercises;
Non-steroid anti-inflammatory drug or cortisone use over an
extended period of time; previous surgery to the knee joint;
trauma to the knee joint affecting the presenting clinical
condition; physiotherapy or other similar exercises for patel-
lofemoral pain syndrome within the previous 3 months.
2.4  | Randomization
The randomization sequence was computer-generated,
stratified by sex, and consisted of blocks of a variable size,
unknown to any of the research team. The sequence was con-
cealed in opaque envelopes and stored by an independent
nurse, who delivered them sequentially to the study physi-
otherapist at randomization.
HOTT et al.   
|
   743

2.5  | Blinding
The physiotherapists providing the interventions were blinded
to baseline measures. All outcome measures were collected by
blinded observers. It was not possible to blind the patients or
physiotherapists who provided the interventions. Patients’ ex-
pectations about the effectiveness of each intervention were as-
sessed at baseline (Table 1). Statistical analysis was performed
blinded for the 3 month results.21 As the blinding was opened
in order to complete the 3-month manuscript, a blinded analysis
was not possible for the current report of 12-month outcomes.
2.6  | Interventions
2.6.1  |  Exercise dosage
adherence to the exercise program was recorded at the weekly
session with the physiotherapist.
Patients performed three exercise sessions per week
(one supervised and two home sessions) for 6 weeks. Initial
dosage was three sets of 10 repetitions for each exercise,
progressing to a maximum three sets of 20 repetitions.
Additional resistance thereafter was achieved through
weights or elastic tubing depending on the exercise (see
Appendix).21 Dosage was adjusted individually such that
the last repetitions were difficult while still maintaining
high movement quality throughout the entire program.
To avoid focus on pain, principles of operant condition-
ing were used,29 in which exercise quotas are set below the
patient's limit of tolerance as opposed to training up to the
pain threshold.

Exercise dosages for hip and knee groups were matched and
were based on previous studies.11,26-28 Details are provided in
the Appendix and in a previous publication reporting the out-
comes at 3 months.21 Patients received a written, illustrated
manual in addition to personal instruction. Self-reported
2.6.2  |  Hip-focused exercise
The hip-focused exercises were based on previous studies11,27
and consisted of side-lying hip abduction, hip external rota-
tion (clamshell), and prone hip extension. These exercises

T A B L E 1   Baseline Characteristics
Group Allocation Knee (n = 37) Hip (n = 39) Control (n = 36)
Age in years, mean (SD) 28.5 (6.2) 27.8 (8.6) 26.3 (7.0)
Gender (Female/male) 24/13 25/14 24/12
BMI, mean (SD) 26.9 (4.6) 25.9 (5.2) 26.4 (5.0)
Unilateral/bilateral 10/27 13/26 8/28
Symptom duration
3-6 mo 2 (5) 1 (3) 5 (14)
6-12 mo 7 (19) 5 (13) 11 (31)
12-24 mo 8 (22) 10 (25) 6 (17)
>24 mo 20 (54) 23 (59) 14 (39)
Higher education (>13 y) 11 (30) 10 (25) 12 (33)
Sick-listed 6 (16) 3 (8) 6 (16)
Regular use of analgetics 5 (14) 6 (15) 8 (22)
Use of insoles 4 (11) 3 (8) 9 (25)
Use of knee support or brace 5 (14) 9 (23) 5 (14)
Emotionally distressed 9 (24) 11 (28) 10 (28)
(HSCL ≥ 1.8)
Kinesiophobia
Normal to mild (13-32) 29 (78) 32 (82) 27 (75)
Moderate to high (33-52) 8 (22) 7 (18) 9 (25)
Expectation of effect (VAS 0 to 10), mean (SD)
for knee training 6.6 (2.3) 7.0 (2.3) 6.8 (2.4)
for hip training 6.6 (2.6) 6.9 (2.5) 6.6 (2.3)
for free activity 5.9 (2.3) 6.9 (2.3) 6.4 (2.6)
Note: Data presented as number (percent) unless otherwise specified.
Abbreviations: BMI, body mass index (kg/m2); EQ-5D, Euro-Qol index; HSCL, Hopkins Symptom Checklist;
n, number of participants; VAS, Visual Analog Scale.
 |
744       HOTT et al.
aimed to maximally isolate the posterolateral hip muscles
without stimulating the quadriceps muscles.
2.6.3  |  Knee-focused exercise
The knee-focused exercise regime was based on previous
studies11,28 and consisted of supine straight leg raises, supine
terminal knee extensions (from 10° flexion to full exten-
sion), and a mini-squat (45° flexion) with the back supported
against the wall (to reduce stabilizing requirements from the
hip muscles). The exercises aimed to maximally isolate the
quadriceps muscles without stimulating posterolateral hip
muscles.
2.6.4  |  Control group (free physical activity)
At randomization, the control group met with the study
physiotherapist and were encouraged by the physiothera-
pist to be physically active in accordance with the patient
education component but received no specific exercise
regime.
2.6.5  |  Patient education
All patients received the same standardized oral and writ-
ten information in two sessions: at inclusion and randomi-
zation. The patient education component was intended to
address kinesiophobia and encourage self-mastery of symp-
toms. Previous research suggests potential importance of
these factors.18,19 The main patient education intervention
was delivered at inclusion, when all patients attended an
individual 1-hour consultation with a specialist in physi-
cal medicine and rehabilitation which contained elements
identified as important to patients in a former study inves-
tigating the “good back consultation.”20 These elements
include reassurance and explanations of pain mechanisms,
presented in conjunction with thorough patient-centered
history and physical examination. Information focused on
the benign nature of PFP as a “loading pain” as opposed
to an “injury,” emphasizing that there was no evidence of
injury to the structures of the knee. The presumed impor-
tance of muscle strength and coordination in controlling
the kneecap was explained, stressing that it is thought that
exercise is important but it is unknown what type of exer-
cise is best—including whether structured exercise is bet-
ter than free physical activity. Patients were encouraged
to gradually increase physical activity without excessively
provoking knee pain. Traditional advice to focus on “cor-
rect” biomechanical positions of the lower extremity or
avoid certain activities were not given, as we theorized
that this may increase kinesiophobia.20 A shorter version
of the same information (written and oral) was presented
again to all patients at the first meeting with the physi-
otherapist for randomization (approximately 30  minutes).
All research personnel were instructed to communicate in
the same way with study patients, and meetings were held
at several points before and during the study to standardize
communication.
2.7  |  Outcome measures
For this report, outcomes were assessed at baseline, at 3,
and 12  months after inclusion. Outcomes at 6  weeks and
3  months are previously reported.21 All outcome measures
are explained in detail in the published protocol.24
2.7.1  |  Primary outcome measure
The main outcome was a forward-backward translated
and validated version of the Anterior Knee Pain Scale
(AKPS).30 This self-report questionnaire consists of 13
questions assessing pain, symptoms and function, origi-
nally developed by involvement of patients with anterior
knee pain, patellar subluxation, and patellar dislocation.31
The Norwegian version was validated through hypothesis
testing for construct validity in patients with PFP.30 Range
0 to 100 (least symptoms). The minimal clinically impor-
tant difference (MCID) of the original version is previously
reported to be 10.32
2.7.2  |  Secondary outcome measures
Usual pain and worst pain last week were measured using
a Visual Analog Scale (VAS) from 0 to 10 (most pain). For
these measurements, a MCID of 2 has been reported.32
The Tampa scale for kinesiophobia was used to assess fear
of movement/re-injury. Range 13 to 52 (most kinesiophobia).33
Self-efficacy (how certain the respondent feels about dif-
ferent activities) was assessed with the Knee Self-Efficacy
Score. Range 0 to 10 (highest self-efficacy).34
Euro-Qol-5 Dimensions (EQ-5D-5L) was used to assess
health-related problems and quality of life.35 The Danish
index value calculator was used (range −0.62 to 1.00, best
possible).35 EQ-VAS (0-100, best possible) recorded patients’
self-rated overall health.35
A step-down test was used to measure function.36 The
subject steps forward from a 20 cm step and down toward the
floor. For one repetition, the heel of the down limb touches
the floor before the up limb returns to full extension. The
measure is number of repetitions in 30 seconds.
HOTT et al.
Isometric strength was measured for hip abduction, hip
external rotation, and knee extension. Positioning external
stabilization with straps were used based on validated tech-
niques.37 We used portable fixed dynamometry employing a
force sensor (MuscleLabTM 6000 ML Force Sensor 300 kg,
Ergotest Innovation), which was considered a superior mea-
surement device to handheld dynamometry. Standardized
procedures are based on existing techniques.37 Measures are
reported in Newtons.
We chose a limit of 75% completion of exercise sessions
(home and supervised) to define full adherence to the proto-
col. Patients in the control group who received supervised
exercise therapy were defined as non-adherent.
2.8  |  Sample size
The power and sample size calculations were based on a su-
periority trial design. Standard deviation of AKPS and pain
(VAS) was assumed to be 13.5 and 2.25, respectively, based
on previous studies.17 MCID for AKPS and pain were as-
sumed to be 10 and 2.32 With these assumptions, 27 patients
were required in each of the three treatment groups to obtain
80% statistical power with 5% significance level for AKPS
and 19 in each group for pain. To account for possible miss-
ing data or dropouts, we planned to include 35 patients in
each group.
2.9  |  Statistical analysis
The main analysis used the principle of intention to treat
(ITT) including only cases with available follow-up data. The
differences between the three treatment groups at 12 months
were analyzed with an analysis of covariance model using the
baseline value as a covariate. Assumptions of the model were
checked, and a Bonferroni correction was applied for these
analyses. In addition, repeated measurements were analyzed
using linear mixed models with subject-specific random in-
tercept to assess the time course and the difference between
the groups. The model included the following fixed effects:
Outcome measurement of interest at baseline, follow-up time,
treatment group, and the interaction terms between follow-up
time and treatment group and between follow-up time and
outcome measure of interest at baseline. P-value ≤ .05 was
considered statistically significant.
Post-hoc analyses included a per-protocol analysis com-
prising only patients who had fully adhered to the protocol by
the definition given above, for comparison with ITT analysis.
Improvement from baseline was assessed using paired-sam-
ples t test. Effect size was calculated using Cohen's d for the
paired-samples t test (d = t/√N), where t is the T-statistic and
N is the sample size. Effect sizes were interpreted as follows:
  
|
   745
<0.50 weak, from 0.50 to 0.79 moderate, and ≥0.80 strong.38
Between-group differences in number of patients seeking ad-
ditional treatment and analgetic use were tested using Pearson
chi-square analysis.
Missing values in the questionnaires AKPS, TSK, KSES
and HSCL were treated as follows: If <25% of items were
missing, the values were substituted with the arithmetic mean
of values from the available items.39 If ≥25% of items were
missing, the outcome measure was regarded as missing for
the patient.
3  |  RESULTS
3.1  | Patients
317 patients with knee pain were evaluated for inclusion
between September 2014 and September 2017. One hun-
dred and twelve patients were included and randomized,
while 205 were excluded. Eighty-eight percent of patients
completed 12-month follow-up. For details, see flowchart
(Figure 1). Demographic and baseline characteristics were
similar between the groups (Table 1).
3.2  |  Missing data
There were generally few missing data (<5% for all patient-
reported outcomes), with the exception of muscle strength
testing at 12  months, which 14 patients (Hip, n  =  6; knee,
n = 3; control, n = 5) were unable to attend and thus only
submitted patient-reported outcomes. Missing data analysis
found no systematic differences between those attending
muscle strength testing compared to those who did not. Thus,
no imputation was performed.40
3.3  |  Primary outcome
We found no between-group differences in AKPS at
12  months by ITT analysis (Table 2 and Table 3). Linear
mixed models analysis detected no significant group or
time*group effect for repeated measures. The cohort as a
whole continued to improve slightly, and statistically sig-
nificantly, between 3 and 12  months: from 74.8 to 77.5
(P  =  .01). The effect size for AKPS improvement from
baseline was 0.81 (large).
3.4  |  Secondary outcomes
We found no significant between-group differences at
12 months for any secondary outcome measures (Table 3). We
 |
746       HOTT et al.

F I G U R E 1   CONSORT (Consolidated Standards of Reporting Trials) flowchart. n, number

found no significant group effect for any secondary outcomes
in linear mixed models analysis. In general, 3-month improve-
ments were maintained at 12 months (see Table 2 and Figure 2).
The effect size for improvement from baseline to 12 months was
0.66 for usual pain (moderate), 0.70 for worst pain (moderate).
Significant between-group differences in hip abduc-
tion strength and knee extension strength at 3 months were
no longer significant at 12 months (Table 3). The hip and
knee groups improved significantly for nearly all muscle
strength measures at 3 and 12  months, while the control
group did not improve significantly on any muscle strength
measures (Table 2 and Figure 3). There was one extreme
outlier in the knee group for percent change in knee ex-
tension at 12 months (>8 SD from the mean). Sensitivity
analysis using non-parametric tests showed no major dif-
ference in result with or without the outlier. Thus, Figure 3
is presented with the outlier.
3.5  | Compliance
Compliance was high, with 88% of patients adhering to the
protocol (knee group 84%, hip group 92%, control group 92%).
There were no significant differences in compliance between
HOTT et al.   
   747
|
T A B L E 2   Primary and secondary outcomes at baseline, 3, and 12 mo

Outcome Knee (n = 37) Hip (n = 39) Control (n = 36) All (n = 112)

Primary outcome, mean (95% CI)
AKPS (0 to 100)c
Baseline 67.2 (62.7 to 71.7) 64.5 (61.1 to 97.9) 65.3 (61.2 to 69.5) 65.3 (63.0 to 67.6)
a a a
3 mo 74.4 (69.8 to 79.0) 73.1 (69.5 to 76.7) 73.1 (68.2 to 78.0) 73.5 (71.1 to 75.9)a
a a,b a
12 mo 76.6 (70.4 to 83.0) 77.6 (73.4 to 81.8) 78.7 (73.0 to 84.4) 77.7 (74.7 to 80.7)a,b
Secondary outcome, mean (95% CI)
Usual pain (VAS 0 to 10)
Baseline 4.3 (3.6 to 5.0) 4.4 (3.8 to 4.9) 3.7 (3.0 to 4.4) 4.1 (3.8 to 4.5)
3 mo 2.6 (1.8 to 3.4)a 2.9 (2.4 to 3.5)a 3.2 (2.5 to 3.9) 2.9 (2.5 to 3.3)a
a a a
12 mo 2.8 (1.7 to 3.9) 2.8 (2.1 to 3.4) 2.2 (1.4 to 3.0) 2.6 (2.1 to 3.1)a
Worst pain (VAS 0 to 10)
Baseline 6.0 (5.2 to 6.8) 6.5 (5.8 to 7.1) 5.8 (5.1 to 6.5) 6.1 (5.7 to 6.5)
a a
3 mo 4.0 (3.0 to 5.1) 4.9 (4.1 to 5.7) 5.0 (4.1 to 5.9) 4.7 (4.1 to 5.1)a
12 mo 3.7 (2.5 to 4.9)a 4.4 (3.6 to 5.3)a 3.5 (2.4 to 4.6)a,b 3.9 (3.3 to 4.5)a,b
Kinesiophobia (TSK 13 to 52)
Baseline 27.2 (24.9 to 29.5) 26.8 (24.9 to 28.6) 27.2 (25.0 to 29.4) 27.1 (25.9 to 28.2)
3 mo 24.5 (21.8 to 27.1)a 24.5 (22.5 to 26.4)a 25.9 (23.6 to 28.3) 24.9 (23.6 to 26.2)a
12 mo 24.5 (21.7 to 27.3)a 25.1 (22.8 to 27.3) 23.7 (21.3 to 26.1)a,b 24.5 (23.1 to 25.8)a
c
Knee self-efficacy (KSES 0 to 10)
Baseline 6.2 (5.6 to 6.8) 6.2 (5.6 to 6.8) 5.8 (5.1 to 6.3) 6.0 (5.7 to 6.3)
a a a
3 mo 7.1 (6.3 to 7.8) 7.0 (6.5 to 7.6) 6.4 (5.7 to 7.1) 6.8 (6.5 to 7.2)a
a a a,b
12 mo 6.9 (6.1 to 7.1) 6.8 (6.1 to 7.5) 7.1 (6.3 to 8.0) 6.9 (6.5 to 7.4)a
Euro-Qol-5D-5L index (−0.62 to 1.00)c
Baseline 0.75 (0.70 to 0.80) 0.76 (0.73 to 0.80) 0.73 (0.69 to 0.78) 0.75 (0.72 to 0.77)
a
3 mo 0.82 (0.78 to 0.85) 0.80 (0.76 to 0.83) 0.77 (0.73 to 0.82) 0.80 (0.77 to 0.82)a
12 mo 0.80 (0.75 to 0.86) 0.79 (0.75 to 0.83) 0.83 (0.79 to 0.88)a,b 0.81 (0.78 to 0.83)a
c
Euro-Qol VAS (0 to 100)
Baseline 64.3 (58.5 to 70.2) 66.6 (59.4 to 73.7) 62.9 (57.4 to 68.4) 64.6 (61.1 to 68.1)
3 mo 67.0 (60.4 to 73.5) 71.1 (64.3 to 77.9)a 68.7 (62.8 to 74.6)a 69.0 (65.4 to 72.6)a
12 mo 67.0 (60.5 to 75.0) 74.5 (70.1 to 80.6)a 68.7 (61.9 to 75.4)a 70.8 (67.1 to 74.4)a
Step-downd
Baseline 15.6 (13.6 to 17.6) 16.7 (15.2 to 18.2) 16.4 (14.4 to 18.4) 16.2 (15.2 to 17.3)
a a a
3 mo 19.2 (16.8 to 21.6) 19.5 (18.1 to 20.9) 19.2 (16.7 to 21.6) 19.3 (18.1 to 20.5)a
a a a
12 mo 18.9 (16.1 to 21.8) 20.0 (18.4 to 21.6) 19.3 (16.4 to 22.2) 19.4 (18.1 to 20.8)a
Hip abduction strengthe
Baseline 126 (110 to 141) 138 (124 to 152) 138 (122 to 154) 134 (126 to 142)
3 mo 145 (131 to 159)a 149 (134 to 163)a 135 (119 to 151) 143 (135 to 151)a
12 mo 144 (126 to 161)a 153 (137 to 168)a 140 (119 to 161) 146 (136 to 156)a
e
Hip external rotation strength
Baseline 103 (88 to 118) 111 (96 to 125) 114 (97 to 131) 109 (101 to 118)
3 mo 119 (104 to 133)a 123 (110 to 136)a 117 (101 to 133) 120 (112 to 128)a
12 mo 125 (108 to 142)a 127 (112 to 142)a 123 (102 to 144) 125 (115 to 135)a,b
Knee extension strengthe
Baseline 317 (272 to 362) 321 (283 to 358) 337 (289 to 386) 325 (300 to 349)

(Continues)
 |
748       HOTT et al.
T A B L E 2   (Continued)
Outcome Knee (n = 37)
3 mo 313 (279 to 357)
a
12 mo 327 (285 to 369)
Note: Unadjusted values are given.
Abbreviations: AKPS, Anterior Knee Pain Score; VAS, Visual Analog Scale.
a
Significant improvement from baseline at the P ≤ .05 level using paired-samples t test.
b
Significant improvement from 3 mo at the P ≤ .05 level using paired-samples t test.
c
Higher value indicates better function.
d
Step-down, number in 30 s.
e
Muscle strength, Newtons.
the groups. Post-hoc per-protocol analyses did not result in
differing conclusions from the intention-to-treat analyses.
3.6  |  Additional treatment
Additional treatment was sought by 4 (12%) of the knee group,
1 (3%) of the hip group, and 6 (19%) of the free training group
(P  =  .1). Analgetic use did not differ significantly between
groups at any point during the study (P > .4 at all time-points).
4  |   D IS C U S S ION
We found no difference in the long-term effectiveness of either
knee exercise, hip exercise, or free physical activity in patients
with patellofemoral pain. The cohort as a whole improved
in clinical outcomes. The clinical improvements observed at
3 months were maintained or further improved at 12 months.
The two guided exercise groups experienced superior gains in
muscle strength compared to the control group, but this did not
translate to better clinical outcomes in the short or long term.
4.1  |  Comparison to previous studies
4.1.1  |  Hip versus knee exercises
Previous studies show conflicting results comparing hip exer-
cises with knee exercises. Two small studies (16-18 patients
per group) have found isolated hip exercises to be more effec-
tive than knee exercises in the short and/or long term10,11 while
one multi-center RCT including 199 patients found no differ-
ences in pain and function at 6  weeks12 or 6  months.13 The
latter is consistent with our lack of between-group differences
in clinical outcomes between hip- and knee-focused exercise
in the short term or long term. We note that many hip exercise
protocols also stimulate the knee and vice versa, especially ex-
ercise performed in weight-bearing positions,11 which is likely
to influence comparisons between exercise components.
Hip (n = 39) Control (n = 36) All (n = 112)
a
342 (307 to 376) 319 (270 to 368) 327 (304 to 350)
343 (303 to 384) 335 (280 to 389) 335 (310 to 360)
4.1.2  |  Exercise dosage
The ideal exercise dosage in PFP is not known.41 To our
knowledge, only one previous study has specifically addressed
the question of dosage in PFP.42 Their results suggested a dose
effect, finding superior results with higher number of repeti-
tions than usual (30 repetitions per set vs 10 repetitions per
set). Based on this and other studies having shown success
with a similar approach,10,26,27 we chose a higher number of
repetitions for the present study. With respect to exercise type,
schedules and dosages in the guided exercise programs in the
current study are comparable to previous studies.10-12,15,26,28,41
4.1.3  |  Muscle strength
In the current study, we found that improvements in muscle
strength for the two guided exercise groups did not trans-
late to superior clinical results. Esculier et al15 observed the
same phenomenon in their study comparing gait retraining,
therapist-guided exercise, and patient education for runners.
They found no between-group differences in pain and func-
tion despite superior gains in muscle strength for the exercise
group, and superior gains in step rate for the gait retraining
group. That clinical improvements are not necessarily tied to
improvements in muscle strength is observed in two previous
systematic reviews.9,43 This may suggest that clinical improve-
ments may be moderated by factors other than muscle strength.
4.1.4  |  Patient education
Previous research regarding patient education compared to exer-
cise-based intervention in PFP is conflicting. Four studies have
compared patient education to complex exercise-based interven-
tions. Rathleff et al16 found that the addition of exercise resulted
in superior results to education alone in adolescents and that
better adherence to exercise improved odds of recovery. Van
Linschoten et al17 found exercise superior to education alone for
patients recruited from general practice and not those recruited
HOTT et al.   
|
   749

T A B L E 3   Adjusted mean differences (95% CI) between-group at 3, and 12 mo

Outcome Hip versus Control Knee versus Control Hip versus Knee
Primary outcome, mean (95% CI)
AKPS (0 to 100)a
3 mo 1.0 (−4.6 to 6.6) 0.2 (−5.5 to 6.0) 0.8 (−4.8 to 6.4)
12 mo −1.1 (−8.9 to 6.7) −4.3 (−12.3 to 3.7) 3.2 (−4.6 to 11.0)
Secondary outcome, mean (95% CI)
Usual pain (0 to 10)
3 mo −0.6 (−1.6 to 0.4) −0.9 (−1.9 to 0.2) 0.2 (−0.8 to 1.3)
12 mo 0.2 (−1.1 to 1.5) 0.3 (−1.0 to 1.7) −0.1 (−1.4 to 1.2)
Worst pain (0 to 10)
3 mo −0.4 (−1.8 to 1.1) −1.0 (−2.4 to 0.5) 0.6 (−0.9 to 2.0)
12 mo 0.6 (−1.0 to 2.3) 0.1 (−1.6 to 1.9) 0.5 (−1.2 to 2.2)
Kinesiophobia (13 to 52)b
3 mo −1.0 (−3.8 to 1.7) −1.5 (−4.3 to 1.3) 0.5 (−2.3 to 3.2)
12 mo 1.7 (−1.6 to 5.0) 1.1 (−2.3 to 4.4) 0.6 (−2.6 to 3.8)
a
Knee self-efficacy (0 to 10)
3 mo 0.3 (−0.5 to 1.2) 0.3 (−0.5 to 1.2) 0.0 (−0.8 to 0.8)
12 mo −0.7 (1.8 to 0.4) −0.7 (−1.9 to 0.4) 0.0 (−1.1 to 1.1)
Euro-Qol 5D-5L (−0.62 to 1.00)a
3 mo 0.02 (−0.04 to 0.07) 0.03 (−0.02 to 0.09) −0.02 (−0.07 to 0.03)
12 mo −0.07 (−0.14 to 0.01) −0.05 (−0.13 to 0.03) −0.02 (−0.09 to 0.06)
Euro-Qol VAS (0 to 100)a
3 mo −0.8 (−8.4 to 6.8) −2.2 (−10.0 to 5.6) 1.5 (−6.1 to 9.0)
12 mo 3.7 (−5.4 to 12.9) −2.1 (−11.4 to 7.2) 5.9 (−3.2 to 14.9)
c
Step-down
3 mo −0.2 (−2.3 to 2.0) 0.4 (−1.8 to 2.5) −0.5 (−1.6 to 2.6)
12 mo −0.1 (−3.2 to 3.1) 0.2 (−3.1 to 3.4) −0.2 (−3.3 to 2.9)
Hip abductiond
3 mo 10.9 (−3.7 to 25.5) 17.9 (2.7 to 33.2)* −7.2 (−21.7 to 7.6)
12 mo 8.5 (−9.5 to 26.5) 9.8 (−8.2 to 27.8) −1.3 (−18.9 to 16.2)
d
Hip external rotation
3 mo 8.1 (−6.6 to 22.9) 10.3 (−4.9 to 25.6) −2.2 (−16.8 to 12.4)
12 mo 10.2 (−7.3 to 27.7) 15.2 (−2.6 to 33.0) −5.0 (−22.1 to 12.0)
Knee extensiond
3 mo 34.9 (3.9 to 65.9)* 19.4 (−12.8 to 51.5) 15.6 (−15.1 to 46.2)
12 mo 20.8 (−14.3 to 56.0) 25.2 (−9.9 to 60.3) −4.4 (−38.7 to 29.9)
Abbreviations: AKPS, Anterior Knee Pain Score; VAS, Visual Analog Scale.
a
Higher value indicates better function.
b
Lower value indicates less kinesiophobia.
c
Step-down, number in 30 s.
d
Muscle strength, Newtons.
*P = <0.05.

from sport physicians and that exercise did not improve the rate
of self-reported recovery compared to education alone. Clark et
al14 found no significant difference in pain or function between
exercise and patient education, although patients performing ex-
ercise were more likely to be discharged from care, indicating
higher satisfaction. As detailed above, Esculier et al15 compared
exercise to patient education based on load modification for
runners and found no between-group differences in clinical
outcomes. Thus, previous studies have not consistently shown
additional effect of complex exercise-based interventions over
patient education alone. The content and form of patient edu-
cation, exercise interventions, and patient characteristics varies
 |
750       HOTT et al.
F I G U R E 2   Time trends for mean values at baseline, 3, and 12 mo. Error bars represent 95% confidence interval. AKPS, Anterior Knee Pain
Scale (0-100, highest function); VASU, usual pain (0-10); VASW, worst pain (0-10)
between all of these studies. As such, existing literature does not
support conclusions as to type of educational intervention which
can be recommended for patients with PFP.
4.2  |  Interpretation of the results
The lack of between-group differences in clinical outcomes in
this study raises the question whether free training might be as
effective as therapist-guided exercises. In interpreting these re-
sults, it should be considered that the exercise programs aimed
to isolate the muscle groups in question, as opposed to employ-
ing a complex or multimodal exercise-based intervention as is
most commonly used in PFP.6,7,9 Further, isolated exercises
might be more effective if targeted to specific deficits, as theo-
rized by Selfe et al44 These factors might affect relative effec-
tiveness of guided exercise compared to the control group.
To our knowledge, this is the first study to use pa-
tient education attempting to influence kinesiophobia and
self-mastery in PFP, despite the potential importance of
F I G U R E 3   Percent change in
isometric strength for hip abduction (HAB),
hip external rotation (HER), and knee
extension (KEX) at 3, and 12 mo. Error
bars represent 95% confidence interval.
*Indicates significant change from baseline,
P < .05. Outliers are included
these factors.4,18,19 Our improvements in kinesiophobia and
self-efficacy are relatively small—though this may partially
be explained by baseline measures being collected after the
main patient educational component (which was delivered at
inclusion) for logistical reasons. Regardless, we suggest that
a more comprehensive cognitive-based intervention might
result in further improvements in kinesiophobia.29
4.3  |  Implications for future research and
clinical practice
Although our results suggest the importance of patient edu-
cation in PFP, the design of the present study did not allow
us to discriminate between the natural course, placebo and
nocebo and the control intervention (patient education com-
bined with free training). To our knowledge, no study has
compared the effectiveness of patient education alone to no
treatment or placebo in PFP. Thus, this topic could be of in-
terest for exploration in future studies. Further, a cost-benefit
HOTT et al.
analysis comparing structured therapist-guided exercise to
free training would be of relevance.
4.4  |  Strengths of the current study
The main strengths of the present study were adequate power,
good compliance, low rates of dropout (12%), and few missing
data. All observers collecting outcome measures were blinded
to group allocation. The physiotherapists who provided the in-
terventions were blinded to baseline measures. Blinding of the
patients and physiotherapists to group allocation was not pos-
sible. Analysis of the 3-month results was performed blinded,
while was not possible for the 12-month results. Standardized
oral and written information presenting all three interventions
as potentially equal aimed to reduce expectation bias. Patient
expectations to intervention effectiveness were similar between
groups, suggesting success in this respect. The use of MRI at
inclusion ensured that other knee pathologies were excluded.
4.5  |  Limitations and considerations
The relatively low recruitment rate (35%) may limit external
validity, although we have no reason to suspect that excluded
patients differed significantly from those who were included.
The study population has a long pain duration and/or low
baseline AKPS compared to some other studies,12,17,28,45 per-
haps reflecting the setting of a specialist Physical medicine
and Rehabilitation department. This may limit its generaliz-
ability to primary care. The long pain duration may affect the
prognosis, as pain duration is a negative prognostic factor.46
A further limitation is the inability to quantify the activity
level of the control group during the intervention period.
A higher rate of missing data for muscle strength testing at
12 months limits the robustness of these measures.
4.6  | Conclusion
We found no difference in the long-term effectiveness of ei-
ther knee exercise, hip exercise, or free physical activity, when
combined with patient education in PFP. Muscle strength im-
provements in the guided exercise groups did not translate into
improved outcomes compared to the control group.
5  |   P ER S P E C T IV E
The current study found no difference in clinical outcomes
between traditional quadriceps based exercise, isolated hip-
based exercise or free physical activity, when combined
with the same patient education for patients with PFP. The
  
|
   751
guided exercise groups showed superior improvement in
muscle strength compared to the control group, but this did
not translate to superior clinical outcomes, which may sug-
gest that clinical improvement is mediated by factors other
than muscle strength. Our results raise the question whether
free training might be as effective as guided exercise for
PFP. We suggest future randomized studies comparing free
training to guided exercise, for example, in a combined ex-
ercise strategy as is commonly used in PFP. Further, com-
paring patient education to no treatment or placebo should
be of interest, as no previous studies have investigated this
question.
ACKNOWLEDGEMENTS
This study is funded by the Research Department of
Sørlandet Hospital. All authors declare no competing inter-
ests. We wish to thank the physiotherapists at the department
for Physical Medicine and Rehabilitation, Sørlandet Hospital
Kristiansand for performing the interventions and outcome
measures in this trial. We thank Anders Östling for coordi-
nating the intervention and Eirin Fidje Wessman for perform-
ing most of the data collection.
ORCID
Alexandra Hott  https://orcid.org/0000-0003-4438-3042
R E F E R E NC E S
1. Smith BE, Selfe J, Thacker D, et al. Incidence and prevalence of
patellofemoral pain: a systematic review and meta-analysis. PLoS
ONE. 2018;13(1):e0190892.
2. Powers CM, Witvrouw E, Davis IS, Crossley KM. Evidence-
based framework for a pathomechanical model of patellofemoral
pain: 2017 patellofemoral pain consensus statement from the 4th
International Patellofemoral Pain Research Retreat, Manchester,
UK: part 3. Br J Sports Med. 2017;51(24):1713-1723.
3. Dutton RA, Khadavi MJ, Fredericson M. Patellofemoral Pain. Phys
Med Rehabil Clin N Am. 2016;27(1):31-52.
4. Maclachlan LR, Collins NJ, Matthews MLG, Hodges PW,
Vicenzino B. The psychological features of patellofemoral pain: a
systematic review. Br J Sports Med. 2017;51(9):732-742.
5. De Oliveira SD, Rathleff MS, Petersen K, Azevedo FM, Barton CJ.
Manifestations of pain sensitization across different painful knee
disorders: a systematic review including meta-analysis and metare-
gression. Pain Med. 2019;20(2):335-358.
6. Collins NJ, Barton CJ, van Middelkoop M, et al. 2018 Consensus
statement on exercise therapy and physical interventions (ortho-
ses, taping and manual therapy) to treat patellofemoral pain: rec-
ommendations from the 5th International Patellofemoral Pain
Research Retreat, Gold Coast, Australia, 2017. Br J Sports Med.
2018;52:1170-1178.
7. Barton CJ, Lack S, Hemmings S, Tufail S, Morrissey D. The 'Best
Practice Guide to Conservative Management of Patellofemoral
Pain': incorporating level 1 evidence with expert clinical reasoning.
Br J Sports Med. 2015;49(14):923-934.
8. van der Heijden RA, Lankhorst NE, van Linschoten RT,
Bierma-Zeinstra SMA, van Middelkoop M. Exercise for
 |
752       HOTT et al.
treating patellofemoral pain syndrome. Cochrane Database Syst
Rev. 2015;1:Cd010387.
9. Nascimento LR, Teixeira-Salmela LF, Souza RB, Resende RA.
Hip and knee strengthening is more effective than knee strength-
ening alone for reducing pain and improving activity in individuals
with patellofemoral pain: a systematic review with meta-analysis. J
Orthop Sports Phys Ther. 2018;48(1):19-31.
10. Khayambashi K, Fallah A, Movahedi A, Bagwell J, Powers
C. Posterolateral hip muscle strengthening versus quadriceps
strengthening for patellofemoral pain: a comparative control trial.
Arch Phys Med Rehabil. 2014;95(5):900-907.
11. Dolak KL, Silkman C, Medina McKeon J, Hosey RG, Lattermann
C, Uhl TL. Hip strengthening prior to functional exercises reduces
pain sooner than quadriceps strengthening in females with pa-
tellofemoral pain syndrome: a randomized clinical trial. J Orthop
Sports Phys Ther. 2011;41(8):560-570.
12. Ferber R, Bolgla L, Earl-Boehm JE, Emery C, Hamstra-Wright K.
Strengthening of the hip and core versus knee muscles for the treat-
ment of patellofemoral pain: a multicenter, randomized controlled
trial. J Athl Train. 2012;50(4):366-377.
13. Hamstra-Wright KL, Aydemir B, Earl-Boehm J, Bolgla L, Emery
C, Ferber R. Lasting improvement of patient-reported outcomes 6
months after patellofemoral pain rehabilitation. J Sport Rehabil.
2017;26(4):223-233.
14. Clark DI, Downing N, Mitchell J, Coulson L, Syzpryt EP, Doherty
M. Physiotherapy for anterior knee pain: a randomised controlled
trial. Ann Rheum Dis. 2000;59(9):700-704.
15. Esculier JF, Bouyer LJ, Dubois B, et al. Is combining gait retrain-
ing or an exercise programme with education better than education
alone in treating runners with patellofemoral pain? A randomised
clinical trial. Br J Sports Med. 2018;52(10):659-666.
16. Rathleff MS, Roos EM, Olesen JL, Rasmussen S. Exercise during
school hours when added to patient education improves outcome
for 2 years in adolescent patellofemoral pain: a cluster randomised
trial. Br J Sports Med. 2015;49(6):406-412.
17. van Linschoten R, van Middelkoop M, Berger MY, et al.
Supervised exercise therapy versus usual care for patellofemoral
pain syndrome: an open label randomised controlled trial. BMJ.
2009;339:b4074.
18. Domenech J, Sanchis-Alfonso V, Espejo B. Changes in catastroph-
izing and kinesiophobia are predictive of changes in disability and
pain after treatment in patients with anterior knee pain. Knee Surg
Sports Traumatol Arthrosc. 2014;22(10):2295-2300.
19. Piva SR, Fitzgerald GK, Wisniewski S, Delitto A. Predictors of
pain and function outcome after rehabilitation in patients with pa-
tellofemoral pain syndrome. J Rehabil Med. 2009;41(8):604-612.
20. Laerum E, Indahl A, Skouen JS. What is "the good back-consul-
tation"? A combined qualitative and quantitative study of chronic
low back pain patients' interaction with and perceptions of consul-
tations with specialists. J Rehabil Med. 2006;38(4):255-262.
21. Hott A, Brox JI, Pripp AH, Juel NG, Paulsen G, Liavaag S.
Effectiveness of isolated hip exercise, knee exercise, or free phys-
ical activity for patellofemoral pain: a randomized controlled trial.
Am J Sports Med. 2019;47(6):1312-1322.
22. Rathleff MS, Rathleff CR, Olesen JL, Rasmussen S, Roos EM. Is
Knee pain during adolescence a self-limiting condition? Prognosis
of patellofemoral pain and other types of knee pain. Am J Sports
Med. 2016;44(5):1165-1171.
23. Lankhorst NE, van Middelkoop M, Crossley KM, et al. Factors
that predict a poor outcome 5–8 years after the diagnosis of
patellofemoral pain: a multicentre observational analysis. Br J
Sports Med. 2016;50(14):881-886.
24. Hott A, Liavaag S, Juel NG, Brox JI. Study protocol: a randomised
controlled trial comparing the long term effects of isolated hip
strengthening, quadriceps-based training and free physical activ-
ity for patellofemoral pain syndrome (anterior knee pain). BMC
Musculoskelet Disord. 2015;16:40.
25. Moher D, Hopewell S, Schulz KF, et al. CONSORT 2010 expla-
nation and elaboration: updated guidelines for reporting parallel
group randomised trials. BMJ. 2010;340:c869.
26. Earl JE, Hoch AZ. A proximal strengthening program improves
pain, function, and biomechanics in women with patellofemoral
pain syndrome. Am J Sports Med. 2011;39(1):154-163.
27. Khayambashi K, Mohammadkhani Z, Ghaznavi K, Lyle MA,
Powers CM. The effects of isolated hip abductor and external rota-
tor muscle strengthening on pain, health status, and hip strength in
females with patellofemoral pain: a randomized controlled trial. J
Orthop Sports Phys Ther. 2012;42(1):22-29.
28. Witvrouw E, Lysens R, Bellemans J, Peers K, Vanderstraeten G. Open
versus closed kinetic chain exercises for patellofemoral pain. A pro-
spective, randomized study. Am J Sports Med. 2000;28(5):687-694.
29. Lindstrom I, Ohlund C, Eek C, et al. The effect of graded activity
on patients with subacute low back pain: a randomized prospective
clinical study with an operant-conditioning behavioral approach.
Phys Ther. 1992;72(4):279-290; discussion 291–273.
30. Hott A, Liavaag S, Juel NG, Brox JI, Ekeberg OM. The reliabil-
ity, validity, interpretability, and responsiveness of the Norwegian
version of the Anterior Knee Pain Scale in patellofemoral pain.
Disabil Rehabil. 2019;1-10. https​ ://doi.org/10.1080/09638​
288.2019.1671499[Epub ahead of print]
31. Kujala UM, Jaakkola LH, Koskinen SK, Taimela S, Hurme M,
Nelimarkka O. Scoring of patellofemoral disorders. Arthroscopy.
1993;9(2):159-163.
32. Crossley KM, Bennell KL, Cowan SM, Green S. Analysis of out-
come measures for persons with patellofemoral pain: which are
reliable and valid? Arch Phys Med Rehabil. 2004;85(5):815-822.
33. Haugen AJ, Grovle L, Keller A, Grotle M. Cross-cultural adapta-
tion and validation of the Norwegian version of the Tampa scale for
kinesiophobia. Spine. 2008;33(17):E595-601.
34. Thomee P, Wahrborg P, Borjesson M, Thomee R, Eriksson BI,
Karlsson J. A new instrument for measuring self-efficacy in pa-
tients with an anterior cruciate ligament injury. Scand J Med Sci
Sports. 2006;16(3):181-187.
35. Rabin R, Oemar M, Oppe M, Janssen B, Herdman M. EQ-5D-5L
User Guide: basic information on how to use the EQ-5D-5L in-
strument. 2011; http://www.euroq​ol.org/filea​dmin/user_uploa​d/
Docum​enten/​PDF/Folde​rs_Flyer​s/UserG​uide_EQ-5D-5L.pdf.
36. Loudon JK, Wiesner D, Goist-Foley HL, Asjes C, Loudon KL.
Intrarater reliability of functional performance tests for subjects with
patellofemoral pain syndrome. J Athl Train. 2002;37(3):256-261.
37. Thorborg K, Bandholm T, Holmich P. Hip- and knee-strength as-
sessments using a hand-held dynamometer with external belt-fixa-
tion are inter-tester reliable. Knee Surg Sports Traumatol Arthrosc.
2013;21(3):550-555.
38. Cohen J. Statistical Power Analysis for Behavioral Sciences, (2nd
edn). Hillsdale, NJ: Lawrence Erlbaum Associates; 1988.
39. Chavance M. Handling missing items in quality of life studies.
Commun Stat Theory Methods. 2004;33(6):1371-1383.
40. Lydersen S. Last observation carried forward. Tidsskr Nor
Laegeforen. 2019;139(9):845.
HOTT et al.   
|
   753

41. Young JL, Rhon DI, Cleland JA, Snodgrass SJ. The influ-
46. Matthews M, Rathleff MS, Claus A, et al. Can we predict the out-
ence of exercise dosing on outcomes in patients with knee come for people with patellofemoral pain? A systematic review on
disorders: a systematic review. J Orthop Sports Phys Ther. prognostic factors and treatment effect modifiers. Br J Sports Med.
2018;48(3):146-161. 2017;51(23):1650-1660.
42. Osteras B, Osteras H, Torstensen TA, Vasseljen O. Dose-response
effects of medical exercise therapy in patients with patellofem-
oral pain syndrome: a randomised controlled clinical trial. SUPPORTING INFORMATION
Physiotherapy. 2013;99(2):126-131. Additional supporting information may be found online in
43. Santos TR, Oliveira BA, Ocarino JM, Holt KG, Fonseca ST. the Supporting Information section. 
Effectiveness of hip muscle strengthening in patellofemoral
pain syndrome patients: a systematic review. Braz J Phys Ther.
2015;19(3):167-176. How to cite this article: Hott A, Brox JI, Pripp AH,
44. Selfe J, Janssen J, Callaghan M, et al. Are there three main sub- Juel NG, Liavaag S. Patellofemoral pain: One year
groups within the patellofemoral pain population? A detailed
results of a randomized trial comparing hip exercise,
characterisation study of 127 patients to help develop targeted in-
tervention (TIPPs). Br J Sports Med. 2016;50(14):873-880.
knee exercise, or free activity. Scand J Med Sci Sports.
45. Collins N, Crossley K, Beller E, Darnell R, McPoil T, Vicenzino B. 2020;30:741–753. https​://doi.org/10.1111/sms.13613​
Foot orthoses and physiotherapy in the treatment of patellofemoral
pain syndrome: randomised clinical trial. BMJ. 2008;337:a1735.