Joint Bone Spine 76 (2009) 629–636

Review

Clinical practice guidelines for rest orthosis, knee sleeves, and unloading
knee braces in knee osteoarthritis
Johann Beaudreuil a , Samy Bendaya b , Marc Faucher c , Emmanuel Coudeyre d ,
Patricia Ribinik e , Michel Revel f , François Rannou f,∗
a Service de rhumatologie, hôpital Lariboisière, université Paris-7, Paris, France
b Service de rééducation neurologique, hôpital Robert-Ballanger, 93602 Aulnay-sous-Bois, France
c Service de rééducation neuro-orthopédique, hôpital Rothschild, Paris, France
d Centre de médecine physique et réadaptation Notre-Dame, BP 86, 4, avenue Joseph-Claussat, 63404 Chamalières cedex, France
e Service de médecine physique et de réadaptation, centre hospitalier de Gonesse, 95500 Gonesse, France
f Service de médecine physique et de réadaptation, hôpital Cochin, université Paris Descartes, 27, rue de Faubourg-Saint-Jacques, 75014 Paris, France

Accepted 24 February 2009

Available online 20 May 2009

Abstract
Objective: To develop clinical practice guidelines concerning the use of bracing – rest orthosis, knee sleeves and unloading knee braces – for knee
osteoarthritis.
Methods: The French Physical Medicine and Rehabilitation Society (SOFMER) methodology, associating a systematic literature review, collection
of everyday clinical practice, and external review by multidisciplinary expert panel, was used.
Results: Few high-level studies of bracing for knee osteoarthritis were found. No evidence exists for the effectiveness of rest orthosis. Evidence
for knee sleeves suggests that they decrease pain in knee osteoarthritis, and their use is associated with subjective improvement. These actions do
not appear to depend on a local thermal effect. The effectiveness of knee sleeves for disability is not demonstrated for knee osteoarthritis. Short-
and mid-term follow-up indicates that valgus knee bracing decreases pain and disability in medial knee osteoarthritis, appears to be more effective
than knee sleeves, and improves quality of life, knee proprioception, quadriceps strength, and gait symmetry, and decreases compressive loads
in the medial femoro-tibial compartment. However, results of response to valgus knee bracing remain inconsistent; discomfort and side effects
can result. Thrombophlebitis of the lower limbs has been reported with the braces. Braces, whatever kind, are infrequently prescribed in clinical
practice for osteoarthritis of the lower limbs.
Conclusion: Modest evidence exists for the effectiveness of bracing – rest orthosis, knee sleeves and unloading knee braces – for knee osteoarthritis,
with only low level recommendations for its use. Braces are prescribed infrequently in French clinical practice for osteoarthritis of the knee.
Randomized clinical trials concerning bracing in knee osteoarthritis are still necessary.
© 2009 Société française de rhumatologie. Published by Elsevier Masson SAS. All rights reserved.

Keywords: Osteoarthritis; Orthosis; Brace; Sleeve; Guidelines

1. Introduction joint immobilisation, which excludes any dynamic, correc-

Orthosis are external devices used for lower-limb osteoarthri-
tis (OA) [1] and are classified as rest orthosis, knee sleeves
and unloading knee braces. They are recommended by EULAR
(third recommendation) and ACR for the non-pharmacological
management of knee OA [2,3]. Rest orthosis are used for
∗ Corresponding author.
E-mail address: francois.rannou@cch.aphp.fr (F. Rannou).
tive or functional effect. Rest orthosis are created by a stiff
composite, by either casting or a line. Knee sleeves are elastic
non-adhesive orthosis associated with various devices aimed at
patellar alignment or frontal femoro-tibial stabilization. Unload-
ing knee braces are, like knee sleeves, functional devices [1].
They are composed of external stems, hinges and straps. Their
purpose is to decrease compressive loads transmitted to the
joint surfaces, either in the medial or lateral femoro-tibial
compartment, depending on the valgus or varus position of
device.

1297-319X/$ – see front matter © 2009 Société française de rhumatologie. Published by Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.jbspin.2009.02.002
630 J. Beaudreuil et al. / Joint Bone Spine 76 (2009) 629–636
The objective of this study was to develop clinical practice
guidelines concerning the use of bracing – rest orthosis, knee
sleeves and unloading knee braces – for knee OA.
2. Methods
The SOFMER methodology was used [4]. This three-stage
method, involving a systematic review of the literature, an anal-
ysis of professional practices and final scientific committee
advice, has been previously described.
2.1. Systematic review of literature
Literature search professionals systematically searched the
Pascal Biomed, Pubmed and Cochrane databases for articles
published between January 1966 and January 2006. Keywords
were proposed by the scientific committee, which consisted of
physicians in physical medicine and rehabilitation (PMR) as
well as rheumatologists and orthopedic surgeons from both aca-
demic and private practices. The keywords were osteoarthritis,
hip, knee, orthosis, bracing, brace, insole, tapping, strapping,
wrap and strip. We selected clinical trials and biomechanical
studies, controlled or not and randomized or not. Only studies
of human adults in English or French languages were considered.
No restrictions were placed on the definition of OA.
The scientific committee first reviewed abstracts, ensuring
that the studies dealt with bracing before requesting the full-
length articles. Three expert readers from two medical domains
(J.B. [rheumatologist] and S.B. and M.F. [PMR physician])
selected the final set of full-length articles using the predefined
following selection criteria: clinical trials and biomechanical
studies, controlled or not and randomized or not, dealing with
bracing for knee OA. Each reviewer conducted a personal
research to complete the committee’s research. The bibliogra-
phy of each selected article was searched for additional articles.
Consensus between reviewers was obtained if necessary.
The quality of manuscripts was assessed according to the
grading scale of the French Agency for Accreditation and Eval-
uation in Healthcare (ANAES) including four levels of quality
from 1 (high quality) to 4 (low quality) leading to three levels of
recommendation: A, established scientific evidence; B, scien-
tific presumption; and C, low level of evidence (Table 1). Data
were analyzed by the three reviewers using a standardized ques-
tionnaire that included the kind of study, the main objectives,
the population studied, the type of intervention, the outcomes,
results and bias. Differences in assessment among reviewers
were resolved by discussion.
2.2. Study of daily clinical practice
Data on physicians’ daily clinical practice for the prescrip-
tion of bracing for lower-limb OA were collected at the National
Rehabilitation Conference (SOFMER, Rouen, France, Octo-
ber 18, 2006) and the National Rheumatology Congress (SFR
annual Congress, Paris, December 3, 2006), with the use of an
electronic voting device. One of the three medical experts (J.B.)
presented the results from the literature review. Notes were taken
Table 1
Grading system for evidence from published trials of the French Agency for
Accreditation and Evaluation in Healthcare (ANAES).
Grading of published trial results
Randomized controlled trials of high power
Meta-analysis of randomized controlled trials
Decision based on well designed trials
Randomized controlled trials of low power
Comparative non randomized trials well designed
Cohort studies
Case–control studies
Comparative studies with major biais
Retrospective studies
Cases series
Grading of recommendations
Established scientific evidence (level 1 of evidence)
Scientific presumption (level 2 of evidence)
Low level of evidence (levels 3 and 4 of evidence)
on the question-and-answer session that followed. Data on daily
clinical practice were then collected and analyzed (J.B.).
2.3. Elaboration of guidelines and external review by a
multidisciplinary reading committee
Clinical guidelines were established on the basis of system-
atic literature review and daily clinical practice. These guidelines
were reviewed by the scientific committee before validation by
a multidisciplinary expert reading committee including physi-
cians in PMR, rheumatologists, orthopedic surgeons, general
practitioners, physical therapists, social workers, podologists,
occupational therapists, nurses, and patients.
3. Results
3.1. Literature review
3.1.1. Selected articles
The scientific committee collected 52 articles (Biomed, 10;
Pubmed, 38; Cochrane, 4). Fourteen original articles (Table 2)
and one review article [5] were then selected by the reviewers.
Four original articles (Table 2) and one descriptive review [1]
were added to this first selection. After full-text examination,
20 articles were retained. Two descriptive reviews concerned
only rest orthosis, three original articles concerned knee sleeves
and 16 unloading knee bracing.
The results of the quality assessment of each article are in
Table 2. No article deserved a level-1 ANAES score. Seven were
given a level-2 ANAES score and 11 a level-4 ANAES score.
3.1.2. Effectiveness of rest orthosis
The effectiveness of rest orthosis for lower-limb OA has not
been studied in clinical trials. Only two descriptive reviews sug-
gested the usefulness of rest orthosis in knee OA [1,5] and
we found no evidence for the effectiveness of rest orthosis for
lower-limb OA.
 J. Beaudreuil et al. / Joint Bone Spine 76 (2009) 629–636 631

Table 2
Assessment of quality of selected studies.
Authors Experimental design Patients and interventions Outcomes Level of quality
(ANAES)a

Knee sleeves
Berry et al., 1992 [8]
Kirkley et al., 1999 [6]
Mazzuca et al., 2004 [7]b
Unloading knee braces
Barnes et al., 1992 [9]
Horlick et Loomer, 1993 [11]b
Lindenfeld et al., 1997 [15]
Matsuno et al., 1997 [13]
Davidson et al., 1998 [10]b
Hewett et al., 1998 [14]
Liu et al., 1998 [23]
Katsugawara et al., 1999 [18]
Kirkley et al., 1999 [6]
Komistek et al., 1999 [12]b
Draper et al., 2000 [21]
Self et al., 2000 [17]
Birmingham et al., 2001 [19]
Pollo et al., 2002 [16]
Giori 2004 [22]
Richards et al., 2005 [20]
a ANAES scale including four levels of quality from 1 (high quality) to 4 (low quality).
b References added to the preliminary selection of the scientific committee.
Controlled, parallel, n = 170, knee OA Pain, global assessment 2
randomized Sleeve (83)
No sleeve (87)
Controlled, parallel, n = 119, medial femoro-tibial OA WOMAC, MACTAR, function tests 2
randomized Unloader-brace (41)
Neoprene sleeve (38)
No bracing (40)
Controlled, parallel, n = 52, knee OA WOMAC 2
randomized Heat-retaining sleeve (26)
Placebo sleeve (25)
Not controlled n = 30, medial femoro-tibial OA Femoro-tibial angle, AAOS arthritis 4
Unloader-brace (30) questionnaire, short-form SF-36,
compliance
Controlled, crossover, n = 39, medial femoro-tibial OA Femoro-tibial angle, pain and 2
randomized Unloader-brace (39) function
Neutral brace (39)
No brace (39)
Not controlled n = 11, medial femoro-tibial OA Adductor moment, walk analysis, 4
Unloader-brace (11) pain and function
Not controlled n = 19, medial femoro-tibial OA Femoro-tibial angle, postural control, 4
Unloader-brace (19) strength of quadriceps, JOA
Not controlled n = 12, medial femoro-tibial OA Femoro-tibial angle 4
Unloader-brace (12)
Not controlled n = 18, medial femoro-tibial OA Adductor moment, walk analysis, 4
Unloader-brace (18) pain and function, global assessment
Not controlled n = 11, knee OA Global assessment, orthosis 4
Unloader-brace (11) complaints
Not controlled n = 14, medial femoro-tibial OA Compression loads, pain and function 4
Unloader-brace (14)
Controlled, parallel, n = 119, medial femoro-tibial OA WOMAC, MACTAR, function tests 2
randomized Unloader-brace (41)
Neoprene sleeve (38)
No bracing (40)
Not controlled n = 20, medial femoro-tibial OA Femoro-tibial angle 4
No bracing (40)
Not controlled n = 30, medial femoro-tibial OA Walk analysis, pain and function 4
Unloader-brace (30)
Controlled, crossover, n = 5, medial femoro-tibial OA Adductor moment 2
randomized Unloader-brace (5)
No brace (5)
Controlled, crossover, n = 20, genu varum with OA Proprioception, postural control 2
randomized Unloader-brace (20)
No brace (20)
Not controlled n = 11, medial femoro-tibial OA Adductor moment, compressions 4
Unloader-brace (11) loads, walk analysis, pain and
function
Not controlled n = 49, medial (43) or lateral (6) Global assessment, complications 4
femoro-tibial OA
Unloader-brace (49)
Controlled, crossover, n = 12, genu varum with OA Walk analysis, pain and function 2
randomized Unloader-brace (12)
Simple hinged brace (12)

3.1.3. Effectiveness of knee sleeves found a decrease in pain after walking and climbing stairs at
The effectiveness of knee sleeves was evaluated in three level- 6-month follow-up with and without the sleeve. Stiffness was
2 ANAES studies (Table 2). Simple neoprene knee sleeves were also improved but not Western Ontario and MacMaster Univer-
evaluated in one study of medial femoro-tibial OA [6], which sities OA Index (WOMAC) physical disability score. However,
632 J. Beaudreuil et al. / Joint Bone Spine 76 (2009) 629–636
results with knee sleeves appeared not as good as those with
unloading valgus knee braces for pain at rest or effort. Heat-
retaining sleeves composed of lycra, polyester and aluminium
fibers were evaluated in one study [7] and compared with stan-
dard devices. After 4 weeks of use, 12 hours per day, WOMAC
scores for pain, stiffness and physical disability did not differ.
Sleeves with a silicon peripatellar device were evaluated in one
study [8], which found a decrease in pain at rest and subjective
improvement after a 6-week follow-up.
3.1.4. Effectiveness of unloading knee bracing
Data concerning unloading knee braces were more numerous
(Table 2). They can be divided in biomechanical and clinical
results.
3.1.4.1. Biomechanical criteria. The biomechanical effect of
unloading knee braces was evaluated according to three crite-
ria: femoro-tibial angle [9–13], adductor moment [14–17], and
femoro-tibial compressive loads [16,18].
3.1.4.1.1. Femoro-tibial angle. Results concerning
femoro-tibial angle are conflicting. Three level-4 ANAES
studies showed a valgus effect of unloading knee braces
for the medial femoro-tibial compartment [10,12,13]. One
level-2 ANAES study [11] and one level-4 study [9] failed to
demonstrate this effect.
Among studies with positive results, one involved the static
condition: an anteroposterior roentgenogram taken in the one-
foot-standing position with and without a valgus brace after 12-
month use in 20 patients [13]. The study found a decrease in
varus of 1.4◦ . Two other studies involved dynamic conditions
[10,12]:
• one found a decrease in varus of 2.2◦ and an increase in medial
femoro-tibial space of 1.3 mm during the foot-strike of gait
on fluoroscopic examination in 20 patients [12];
• another study of videoscopy found a decrease in varus at the
beginning of the swing phase of gait depending on tibia and
not femur position variation in 12 patients [10].
Two studies with negative results involved radiography of the
static condition and a two-foot standing position in one of the
studies, in 20 and 40 patients each [9,11].
3.1.4.1.2. Adductor moment. The knee adductor moment
serves as an indirect measure of medial femoro-tibial compart-
ment loading [16]. A change in the adductor moment signifies
a change in load distribution across the knee joint. Adductor
moment depends on mechanical alignment of the knee as well
as on the ground reaction force. Greater is the moment arm dis-
tance between force line and knee axis, greater is the adductor
moment (Fig. 1). Adductor moment was evaluated by video anal-
ysis of gait and forceplates in four studies, one level-2 ANAES
study [17] and three level-4 studies [14–17], in patients with
medial femoro-tibial OA. Two studies showed a decrease in
adductor moment with an unloading valgus knee brace [15,17].
One indicated a 10% variation in conducting to values that were
close to those of healthy controls, in 11 patients with medial
femoro-tibial OA that was previously treated with arthroscopy
Fig. 1. Shematic representation of adductor moment. Adductor moment is the
product of the ground reaction force and the distance between the line of action
of this force and the center of the axis of the knee.
[15]. The second showed a decrease of adductor moment dur-
ing the stance phase in patients with and without an unloading
valgus knee brace. However, two other studies found negative
results [14,16].
3.1.4.1.3. Femoro-tibial compressive loads. Compressive
loads in the medial femoro-tibial compartment were evaluated
by gait analysis in one level-4 ANAES study of 11 patients with
medial femoro-tibial OA [16]. This study, of patients with and
without the valgus brace, indicated a decrease in loads in the
calibrated 4◦ valgus brace setting. The unloading effect could
be improved to −15% by increasing the strap tension.
One study evaluated medial femoro-tibial loads by indirect
criteria: ratio of bone mineral density between lateral and medial
tibial plates [18]. After 3 months of valgus knee brace use in
14 patients with medial femoro-tibial OA, the ratio of lateral to
medial bone mineral density was increased.
Braces might not change the varus alignment of the leg,
prevent the increase of varus at mid-stance. Unloading could
be caused by gait adaptations, slower velocity, increasing out-
toeing of the foot.
3.1.4.2. Clinical criteria.
3.1.4.2.1. Proprioception. One level-2 ANAES study of a
crossover design in 20 patients with genu varum and OA indi-
cated that valgus knee braces improved the ability to replicate
knee joint position as a proprioception marker [19].
3.1.4.2.2. Postural control. Postural control in static con-
ditions was evaluated by plateforce in two ANAES studies, one
level 2 [19] and one level 4 [13]. Results are conflicting. One
crossover study of 20 patients with genu varum and OA wearing
4◦ valgus knee braces found no effect [19], and the other, of
20 patients with medial femoro-tibial OA, showed an improve-
ment over baseline values of joint stability with valgus knee
bracing after 12 months [13].
3.1.4.2.3. Strength of quadriceps. One level-4 ANAES
study of 20 patients with medial femoro-tibial OA showed an
 J. Beaudreuil et al. / Joint Bone Spine 76 (2009) 629–636 633

increase over baseline values in isokinetic strength of quadriceps satisfactory results at 2.5 years in 46 patients with medial or
with valgus knee bracing after 12 months [13]. lateral femoro-tibial OA treated with unloading knee braces
3.1.4.2.4. Walk analysis. During normal gait, the medial [22].
compartment of the knee is loaded more than the lateral com- 3.1.4.2.7. Observance and safety. Data on effectiveness of
partment. It is estimated that 60 to 80% of the load during the unloading knee bracing should be weighed with observance
mid-stance phase of gait is distributed to the medial compart- and safety study results. These are available from four level-
ment in normal knee. This is due to the adductor moment. Gait 4 ANAES studies [9,11,22,23]. Observation follow-up ranged
measurements have been analyzed by dynamic systems, with from 6 months to 3 years and observance varied from 41 to
plateforce and video acquisition. 93%. Reasons for the 20% discontinued use of orthosis were
3.1.4.2.4.1. Force at stance.3.1.4.2.4.1. Force at stance detailed in one 6-month study as being absence of effect and side
Results of force at stance are conflicting. One level-2 ANAES effects (2 knee swellings, 1 venous thromboembolic accident)
study showed an increase in peak force at stance with valgus [22]. Reasons for discomfort were from results of a 1.75-year
knee bracing as compared with no brace in 12 patients with study of 11 patients and were heaviness (3), heat (9), volume (4)
genu varum and medial femoro-tibial OA [20]. However, no or mobility (10) of devices, and persistent joint instability (5)
difference was found between bracing in the valgus or neutral [23]. At 3 years, nine of 22 patients (41%) claimed skin irrita-
position. Two level-4 studies, of 11 and 13 patients with medial tion (6), mechanical problems with the device (2) and attachment
femoro-tibial OA, did not find any such effect [14,15]. failure (1).
3.1.4.2.4.2. Vertical propulsive force.3.1.4.2.4.2. Vertical
propulsive force One level-2 ANAES study found an increase 3.2. Clinical daily practice
in vertical propulsive peak force with bracing in the valgus or
neutral position as compared with no bracing in 12 patients with 3.2.1. Rest orthosis
genu varum and medial femoro-tibial OA [20] but no difference Prescription of rest orthosis for lower-limb OA was in general
in effect between the two positions. limited among PMR physicians and rheumatologists (Table 3).
3.1.4.2.4.3. Gait cadence and velocity.3.1.4.2.4.3. Gait It was systematically proposed only in 0 to 1% of cases of knee
cadence and velocity Three level-4 ANAES studies [14–16] or ankle OA but often prescribed in 7 to 10% of cases. The
found no effect of valgus knee bracing on gait cadence [14,15] prescription of this kind of device was relatively more frequent
or gait velocity for patients with medial femoro-tibial OA [16]. among PMR physicians (rarely 33–60% and never 33–56%) than
3.1.4.2.4.4. Gait symmetry.3.1.4.2.4.4. Gait symmetry among rheumatologists (rarely 19–27% and never 63–72%). The
One level-4 ANAES study described an improvement of gait prescription seemed also more frequent for knee than ankle OA.
symmetry during stance and swing phase with unloading valgus
knee bracing versus no bracing in 30 patients with medial 3.2.2. Knee sleeves
femoro-tibial OA [21]. Variations in ratio of stance to swing PMR physicians and rheumatologists more similarly pre-
phase suggested lengthening the stance phase as the swing scribed knee sleeves (Table 3). Systematic prescription was
phase is shortened.
3.1.4.2.5. Pain and disability. Three level-2 ANAES stud- Table 3
ies [6,11,20] and seven level-4 studies [9,13–16,18,21] evaluated Daily clinical practice in physical medicine and readaptation (PMR) and in
rheumatology.
the effect of unloading valgus knee bracing on pain and disabil-
ity in patients with medial femoro-tibial OA with or without PMR (%) Rheumatology (%)
genu varum. Results suggested the effectiveness of such devices Do you prescribe rest orthosis in knee osteoarthritis (OA)?
as compared with no orthosis. The follow-up periods ranged Always 0 1
from 6 weeks to 6 months in level-2 studies and from 2 weeks to Often 7 9
1 year in level-4 studies. A level-2 study showed a better effect Rarely 60 27
Never 33 63
with unloading knee bracing as compared with neoprene knee
sleeves on pain and walk capacity (6-min walk test) at 6 months Do you prescribe rest orthosis in ankle OA?
[6]. Always 1 1
Often 10 8
3.1.4.2.5.1. Quality of life.3.1.4.2.5.1. Quality of life One Rarely 33 19
level-4 ANAES study used the SF36 quality-of-life score to eval- Never 56 72
uate the effect of unloading valgus knee bracing in 30 patients
Do you prescribe knee sleeves in knee OA?
with medial femoro-tibial OA [9] and found an improvement Always 1 3
after 8-week use as compared to baseline values. Often 25 32
3.1.4.2.6. Global assessment of final results. Three level-4 Rarely 53 49
ANAES studies described their results in terms of global assess- Never 21 16
ment [14,22,23]: 39 and 54% good, very good or excellent results Do you prescribe unloading knee braces in knee OA?
at 9 and 46 weeks in 18 and 13 patients with medial femoro- Always 4 0
tibial OA wearing unloading valgus knee braces [14]; 72% Often 19 9
Rarely 51 32
good results with unloading valgus knee braces at 1.75 years Never 26 58
in 11 patients with knee OA [23]; and 56% satisfactory or very
634 J. Beaudreuil et al. / Joint Bone Spine 76 (2009) 629–636
extremely low. The prescription of knee sleeves was frequent
among one-quarter to one-third of physicians and was rarely or
never proposed by 65 to 74%.
3.2.3. Unloading knee bracing
Unloading knee bracing was more frequently prescribed by
PMR physicians than rheumatologists (Table 3). Indeed, 23% of
PMR physicians proposed it systematically, as compared with
only 9% of rheumatologists. Prescription of unloading knee
bracing remained low, however, for both groups. Such devices
were rarely prescribed among 32 to 51% of physicians and never
among 26 to 58%.
4. Discussion
In contrast to other studies [3,24], this systematic review of
the literature, concerning bracing in knee OA – rest orthosis,
knee sleeves and unloading knee braces – included clinical or
biomechanical trials in the field, whatever their quality assess-
ment score. Indeed, non-randomized studies were also being
considered for analysis. This analysis allows for an extensive
coverage of scientific publications on the topic, as a sensitive
synthesis and recommendations according to level of evidence.
These results are further weighed by study of clinical prac-
tice. The orthoses available for lower-limb OA, namely for knee
OA, are numerous and varied [1], as are probably their mecha-
nisms of action. Their potential disadvantages and prices vary.
Establishing an objective balance between the advantages and
disadvantages of these kinds of devices appears therefore of
strong interest.
The experimental basis for the use of rest orthosis in lower-
limb OA is lacking. Our results of clinical practice study among
PMR physicians and rheumatologists agree with this conclusion.
Rest orthosis is infrequently prescribed for knee and ankle OA.
However, this kind of device could be interesting for transient
immobilization for congestive OA. This hypothesis remains to
be tested in a high-level clinical trial with, as a control, for
example, simple rest, balneotherapy, physiotherapy, and arthritic
crutches or sticks [25].
Results concerning knee sleeves remain sparse [6–8]. How-
ever, knee sleeves are the most-used device among orthoses for
this indication, among PMR physicians and rheumatologists.
The simplicity of knee sleeves in daily practice could be an
explanation for this observation. One-quarter to one-third of
physicians often prescribed knee sleeves for knee OA. However,
many did not. Finally, the frequency of prescription remained
relatively low. Objectively, the use of knee sleeves decreases
pain and is associated with short-term subjective improvement
[6,8]. These actions do not appear to depend on any local ther-
mic effect [7] and are not associated with reduced physical
disability.
A high number of studies have involved unloading knee brac-
ing for knee OA. The studies mainly evaluated the effectiveness
of valgus devices in medial femoro-tibial OA with or without
genu varum. Results concerning the biomechanical effect, with
the femoro-tibial angle [9–13] and adductor moment [14–17]
used as criteria, remain conflicting. Therefore, the effectiveness
of unloading knee bracing is not established. However, unload-
ing knee bracing appears to decrease compressive loads in the
medial femoro-tibial compartment [16,18].
Unloading knee bracing also improves knee propriocep-
tion [19]. Results concerning joint stability remain conflicting
[13,19]. Unloading valgus knee bracing improves isokinetic
strength of quadriceps [13], but results concerning strength
of foot strike during walking are conflicting [14,15,20]. This
kind of device does not improve gait velocity [14–16] but does
improve vertical propulsive force [20] and gait symmetry [21].
However, the effect is not better than with articulated orthosis
in neutral (no valgus) position [20].
Unloading valgus knee bracing has short- and mid-term ben-
efits for pain and disability [6,9,11,13–16,18,20,21], with the
effectiveness higher than with neoprene knee sleeves [6]. The
bracing also improves quality of life in the short-term [9].
However, the responsiveness to unloading knee braces remains
inconsistent; 39 to 93% of patients with knee OA appear to
show improvement with such devices [14,22,23]. Improvement
on observation also varies [9,11,22,23], with 41 to 93% effect at
6-month to 3-year follow-up. Unloading knee bracing can gen-
erate various side effects [22,23], the most serious being venous
thromboembolic events [22].
These findings may be responsible for a low level of pres-
cription of rest orthosis, as noticed by the daily clinical practice
of PMR physicians and rheumatologists. The cost of unload-
ing knee braces and other therapeutics for knee OA could be an
explanation. Differences in practice between PMR physicians
and rheumatologists could be explained by the resort, command,
availability or demand for other therapeutic options depending
on physician and patient specificity.
5. Recommendations
Twenty-three treatment guidelines for the management of
hip and knee OA have been identified up to January 2006
[26]. Combination of non-pharmacological and pharmaco-
logical modalities appeared as optimal management. Among
non-pharmacological treatments, knee bracing is recommended
for knee OA. However, non-pharmacological recommenda-
tions are less precise than pharmacological recommendations.
We propose separate recommendations for rest orthosis, knee
sleeves and unloading knee braces. These recommendations sup-
plement those for knee taping that have recently been published
[27].
5.1. Rest orthosis
The effectiveness of rest orthosis in lower-limb OA is not
clearly established on the basis of scientific evidence (grade C).
Indeed, the literature search uncovered no clinical trial. Rest
orthosis has been proposed for knee OA in descriptive reviews,
without more evidence (grade C evidence). The daily clinical
practice of French PMR physicians and rheumatologists agrees
with this lack of evidence: rest orthosis remains little prescribed
for lower-limb OA.
 J. Beaudreuil et al. / Joint Bone Spine 76 (2009) 629–636
High quality clinical trials – level-1 or -2 ANAES score – of
congestive knee OA are recommended (grade C evidence).
5.2. Knee sleeves
Knee sleeves appear to have an antalgic effect on knee OA,
independent of any local heating action (grade B evidence).
Their use is associated with subjective improvement (grade B
evidence) and their effect on physical disability is not demon-
strated (grade B evidence).
High quality clinical trials – level-1 or -2 ANAES score, with
particular attention to potential conflict of interest – are still nec-
essary. The following design is suggested: with and without knee
sleeves and knee sleeves with and without peripatellar alignment
or frontal stabilizing devices (grade C evidence).
5.3. Unloading knee braces
Unloading valgus knee braces can be used for symptomatic
medial femoro-tibial OA because of short- and mid-term reduc-
tion of pain and disability (grade B evidence). They appear to be
more effective than neoprene knee sleeves (grade B evidence)
and improve quality of life in the short-term (grade C evidence).
They favorably modify compressive loads in the medial femoro-
tibial compartment, joint proprioception, isokinetic strength of
quadriceps, gait symmetry and perhaps vertical propulsive force
(grade C evidence).
The results of observation and responsiveness to unload-
ing knee braces remain inconsistent (grade C evidence). Side
effects are various (grade C evidence). The most serious are
venous thromboembolic events (grade C evidence). Safety,
observance, and responsiveness are therefore important consid-
erations before proposing unloading knee bracing for knee OA
(grade C evidence).
Most studies of unloading orthosis are of low quality, namely
level-4 ANAES score. High-quality trials are therefore nec-
essary (grade C evidence). The following fields should be
investigated:
• clinical predictive factors for safety, observance and respon-
siveness;
• optimal duration of orthosis wear;
• biomechanical ways of actions, considering especially the
previous conflicting results and criteria such as femoro-tibial
angle, adductor moment, foot strike force and vertical propul-
sive force (grade C evidence).
6. Limits
Our study has several limits. The literature search we did
focused on the Pascal Biomed, Pubmed and Cochrane databases.
ClinicalTrials.gov and Physiotherapy Evidence database were
not searched. Missing articles are therefore possible. Further-
more, publication bias was not considered. Methodology quality
assessment was performed on the basis of the ANAES grading
scale, which may be less informative than a Delphi list or the
PEDro score. However, it clearly demonstrated the low qual-
635
ity of most of the selected articles. The limiting factors of the
present guidelines are also linked to the low number of clinical
trials evaluating the effectiveness of orthosis for knee OA and
the diversity of devices that have been tested or, further, that are
now on the market.
Studies have been published since our search. Few are
randomized clinical trials [28,29]. They agree with clinical effec-
tiveness but low observance rate for unloading knee bracing in
knee OA. Biomechanical effect remained controversial or to be
further investigated. Knee sleeves may indeed increase static
and dynamic balance [30]. Medial condylar separation [31],
decreased adductor moment [32] and reduced muscles cocon-
tractions, namely vastus medialis-medial hamstrings [33], have
recently been suggested for unloading knee bracing. Evidence
for the effectiveness of orthosis – rest orthosis, knee sleeves and
unloading knee braces –for knee OA remain sparse. No evi-
dence has a grade of recommendation higher than B or C. The
usefulness of these devices among therapists for lower-limb OA
appears limited.
Conflict of interests
None of the authors has any conflicts of interest to declare.
Appendix A. Supplemantary data
The supplemantary data can be found at the following doi:
doi:10.1016/j.jbspin.2009.02.002.
References
[1] Beaudreuil J, Bernard JL. Orthèses du genou. In: de Sèze S, Ryckwaert A,
Kahn MF, et al., editors. L’actualité rhumatologique 2006. Paris: Elsevier;
2006. p. 393–409.
[2] Recommendations for the medical management of osteoarthritis of the hip
and knee: 2000 update. American College of Rheumatology Subcommittee
on Osteoarthritis Guidelines. Arthritis Rheum 2000;43:1905–15.
[3] Jordan KM, Arden NK, Doherty M, et al. EULAR recommendations
2003: an evidence based approach to the management of knee osteoarthri-
tis: report of a Task Force of the standing committee for international
clinical studies including therapeutic trials (ESCISIT). Ann Rheum Dis
2003;62:1145–55.
[4] Rannou F, Coudeyre E, Ribinik P, et al. Establishing recommandations for
rehabilitation interventions: the SOFMER methodology. Ann Readapt Med
Phys 2007;50:100–10.
[5] Thoumie P, Sautreuil P, Mevellec E. Orthèses de genou. Évaluation de
l’efficacité clinique à partir d’une revue de la littérature. Ann Readapt Med
Phys 2002;45:1–11.
[6] Kirkley A, Webster-Bogaert S, Litchfield R, et al. The effect of bracing on
varus gonarthrosis. J Bone Joint Surg 1999;81-A:539–48.
[7] Mazzuca SA, Page MC, Meldrum RD, et al. Pilot study of the effects of a
heat-retaining knee sleeve on joint pain, stiffness, and function in patients
with knee osteoarthritis. Arthritis Rheum 2004;51:716–21.
[8] Berry H, Black C, Fernandes L, et al. Controlled trial of a knee support
(Genutrain) in patients with osteoarthritis of the knee. Eur J Rheumatol
Inflamm 1992;12:30–4.
[9] Barnes CL, Cawley PW, Hederman B. Effect of CounterForceTM Brace
on symptomatic relief in a group of patients with symptomatic unicom-
partimental osteoarthritis: a prospective 2-year investigation. Am J Orthop
2002;31:396–401.
[10] Davidson PL, Sanderson DJ, Loomer RL. Kinematics of valgus bracing for
medial gonarthrosis: technical report. Clin Biomech 1998;13:414–9.
636 J. Beaudreuil et al. / Joint Bone Spine 76 (2009) 629–636
[11] Horlick SG, Loomer RL. Valgus knee bracing for medial gonarthrosis. Clin
J Sport Med 1993;3:251–5.
[12] Komistek RD, Dennis DA, Northcut EJ, et al. An in vivo analysis of the
effectiveness of the osteoarthritic knee brace during heel-strike of gait. J
Arthroplasty 1999;14:738–42.
[13] Matsuno H, Kadowaki KM, Tsuji H. Generation II knee bracing for severe
medial compartment osteoarthritis of the knee. Arch Phys Med Rehabil
1997;78:745–9.
[14] Hewett TE, Noyes FR, Barber-Westin SD, et al. Decrease in knee joint pain
and increase in function in patients with medial compartment arthrosis: a
prospective analysis of valgus bracing. Orthopedics 1998;21:131–8.
[15] Lindenfeld TN, Hewett TE, Andriacchi TP. Joint loading with val-
gus bracing in patients with varus gonarthrosis. Clin Orthop Relat Res
1997;344:290–7.
[16] Pollo FE, Otis JC, Backus SI, et al. Reduction of medial compartment
loads with valgus bracing of the osteoarthritic knee. Am J Sports Med
2002;30:414–21.
[17] Self BP, Greenwald RM, Pflaster DS. A biomechanical analysis of a
medial unloading brace for osteoarthritis in the knee. Arthritis Care Res
2000;13:191–7.
[18] Katsuragawa Y, Fukui N, Nakamura K. Change of bone mineral density
with valgus knee bracing. Int Orthop 1999;23:164–7.
[19] Birmingham TB, Kramer JF, Kirkley A, et al. Knee bracing for medial
compartment osteoarthritis: effects on proprioception and postural control.
Rheumatology 2001;40:285–9.
[20] Richards JD, Sanchez-Ballester J, Jones RK, et al. A comparison of knee
braces during walking for the treatment of osteoarthritis of the medial
compartment of the knee. J Bone Joint Surg 2005;87-B:937–9.
[21] Draper ERC, Cable JM, Sanchez-Ballester J, et al. Improvement in function
after valgus bracing of the knee. An analysis of gait symmetry. J Bone joint
Surg 2000;82-B:1001–5.
[22] Giori NJ. Load-shifting brace treatment for osteoarthritis of the knee: a
minimum 2 1/2-year follow-up study. J Rehabil Res Dev 2004;41:187–93.
[23] Liu K, Lao L, Asami T, et al. Clinical care of osteoarthritic knee with knee
orthoses. Fukuoka Igaku Zasshi 1998;89:298–302.
[24] Brouwer RW, Jakma TSC, Verhagen AP, et al. Braces and orthoses
for treating osteoarthritis of the knee. Cochrane Database Syst Rev
2005;1:CD004020.
[25] Beaudreuil J, Haddad A, Caroit M. Traitement non médicamenteux et non
chirurgical de la gonarthrose (arthroscopie exclue). In: de Sèze S, Ryck-
waert A, Kahn MF, et al., editors. L’actualité rhumatologique 2002. Paris:
Elsevier; 2002. p. 353–69.
[26] Zhang W, Moskowitz RW, Nuki G, et al. OARSI recommendations for the
management of hip and knee osteoarthritis, Part II: OARSI evidence-based,
expert consensus guidelines. Osteoarthritis Cartilage 2008;16:137–62.
[27] Richette P, Sautreuil P, Coudeyre E, et al. Usefulness of taping in lower
limb osteoarthritis. French clinical practice guidelines. Joint Bone Spine
2008;75:475–8.
[28] Brouwer RW, van Raaij TM, Verhaar JA, et al. Brace treatment for
osteoarthritis of the knee: a prospective randomized multi-centre trial.
Osteoarthritis Cartilage 2006;14:777–83.
[29] Draganich L, Reider B, Rimington T, et al. The effectiveness of self-
adjustable custom and off-the-shelf bracing in the treatment of varus
gonarthrosis. J Bone Joint Surg 2006;88-A:2645–52.
[30] Chuang SH, Huang MH, Chen TW, et al. Effect of knee sleeve on static
and dynamic balance in patients with knee osteoarthritis. Kaohsiung J Med
Sci 2007;23:405–11.
[31] Dennis DA, Komistek RD, Nadaud MC, et al. Evaluation of off-loading
braces for treatment of unicompartmental knee arthrosis. J Arthroplasty
2006;21(Suppl. 1):2–8.
[32] Gaasbeek RD, Groen BE, Hampsink B, et al. Valgus bracing in patients
with medial compartment osteoarthritis of the knee. A gait analysis study
of a new brace. Gait Posture 2007;26:3–10.
[33] Ramsey DK, Briem K, Axe MJ, et al. A mechanical theory for the effec-
tiveness of bracing for medial compartment osteoarthritis of the knee. J
Bone Joint Surg 2007;89-A:2398–407.