Review

Efcacy of ultrasound therapy for the management of knee osteoarthritis:

a systematic review with meta-analysis
A. Loyola-Snchez
*
, J. Richardson
a
, N.J. MacIntyre
b
School of Rehabilitation Science, McMaster University, Canada
a r t i c l e i n f o
Article history:
Received 5 January 2010
Accepted 17 June 2010
Keywords:
Ultrasound
Osteoarthritis
Knee
Meta-analysis
s u m m a r y
Objective: To assess the efcacy of ultrasound therapy (US) for decreasing pain and improving physical
function, patient-perception of disease severity, and cartilage repair in people with knee osteoarthritis
(OA).
Methods: We conducted a systematic review (to February 2009) without language limits in MEDLINE,
EMBASE, Cochrane Library, LILACS, MEDCARIB, CINAHL, PEDro, SPORT-discus, REHABDATA, and World
Health Organization Clinical Trial Registry. We included randomized controlled trials of people with knee
OA comparing the outcomes of interest for those receiving US with those receiving no US. Two reviewers
independently selected studies, extracted relevant data and assessed quality. Pooled analyses were
conducted using inverse-variance random effects models.
Main results: Six small trials (378 patients) were included. US improves pain [Standardized Mean
Difference (SMD) (95% condence interval (CI)) 0.49 (0.79, 0.18), P 0.002], and tends to improve
self-reported physical function [SMD (CI) 0.54 (1.19, 0.12), P 0.11] along with walking performance
[SMD (CI) 0.81 (0.09, 1.72), P 0.08]. Results from two trials (128 patients), conducted by the same
group, show a positive effect of US on pain [SMD (CI) 0.77 (1.15, 0.39), P <0.001], self-reported
physical function [SMD (CI) 1.25 (1.69, 0.81), P <0.001], and walking performance [SMD (CI)
1.47 (1.06, 1.88), P <0.001] at 10 months after the intervention concluded. Heterogeneity observed
between studies regarding the effect of US on pain was explained by US dose, mode and intensity. The
quality of evidence supporting these effect estimates was rated as low.
Conclusions: US could be efcacious for decreasing pain and may improve physical function in patients
with knee OA. The ndings of this review should be conrmed using methodologically rigorous and
adequately powered clinical trials.
2010 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.
Introduction
Knee osteoarthritis (OA) is the most common presentation of
OA, with an estimated prevalence between 12% and 35% in the
general population
1,2
and is considered the leading cause of
musculoskeletal disability in the elderly population worldwide
3
.
Physiotherapy is recommended for the management of painful
knee OA
4
and ultrasound therapy (US) is one of the most
common physical agents used within physiotherapy practice in
several countries
5
. US is based on the application of high
frequency sound waves to the tissues of the body in order to
obtain mechanical or thermal effects
6
. These effects aim to
enhance soft tissue healing, decrease the inammatory response,
increase blood ow, increase metabolic activity, and decrease
pain
6
. Moreover, there is some evidence that ultrasonic energy
stimulates the repair of joint cartilage in animal models of
cartilage injury
7e9
. Therefore, US could be an effective inter-
vention in the management of pain and disability in people with
knee OA.
The Osteoarthritis Research Society International (OARSI)
struck a committee to complete a systematic review of existing
treatment guidelines (2007) in order to develop recommenda-
tions for the management of knee and hip OA
4
. Ultrasound was
not identied as a core treatment modality based on the results of
a systematic review published in 2001 by the Cochrane Collabo-
ration
10
. The Cochrane systematic review included studies
* Address correspondence and reprint requests to: Adalberto Loyola-Snchez,
School of Rehabilitation Science, IAHS e 403, 1400 Main Street West, Hamilton, ON,
Canada L8S 1C7. Tel: 1-905-525-9140x26410.
E-mail addresses: loyolaa@mcmaster.ca (A. Loyola-Snchez), jrichard@mcmaster.ca
(J. Richardson), macint@mcmaster.ca (N.J. MacIntyre).
a
Tel: 1-905-525-9140x27811.
b
Tel: 1-905-525-9140x21166.
1063-4584/$ e see front matter 2010 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.joca.2010.06.010
Osteoarthritis and Cartilage 18 (2010) 1117e1126
available prior to June 2000
10
while more recent trials evaluating
the effectiveness of US in the management of knee OA were not
reviewed. Since there is a need for effective conservative treat-
ment options for people with knee OA, it is important to conrm
or change current clinical practice guidelines based on best
available evidence. Therefore, the objective of this systematic
review and meta-analysis was to determine the efcacy of US in
decreasing pain and improving physical function in people with
knee OA. Further, we extended the scope of previous systematic
reviews on the topic by evaluating the efcacy of US on patient-
perception of disease severity, and cartilage repair. We followed
the Preferred Reporting Items for Systematic reviews and Meta-
Analyses (PRISMA) recommendations
11
in the creation of this
manuscript.
Methods
Search strategy
The Cochrane Central Register of Controlled Trials (fourth Quarter
2008), MEDLINE (1950 to January week 4, 2009), MEDLINE Daily
Update (Feb 5, 2009), MEDLINE In Process & Other non-indexed cita-
tions (February 5, 2009), EMBASE (1980e2009 week 6), LILACS
(February6, 2009), MEDCARIB(February6, 2009), CINAHL (February8,
2009), pre CINAHL (February 8, 2009), PEDro (last updated February 2,
2009), AMED(1985toJanuary2009), SPORTdiscus (1830toFebruary9,
2009), REHABDATA (1956 to February 9, 2009), and World Health
Organization Clinical Trial Registry (February 8, 2009) databases were
searched by one of the authors (AL). In addition, published literature
with restricted distribution was searched through the ISI Web of
Knowledge, Papers First, Proceedings First, and ProQuest for Disserta-
tions and Theses. Authors of an unpublished study reported as
a conference proceeding in the Proceedings First database were con-
tacted via e-mail. A detailed example of the full electronic search
strategy for Ovid MEDLINE is provided in Appendix A. Briey, the
following medical subject headings (MeSH) were used: Osteoarthritis,
Arthritis, Ultrasonic Therapy, Ultrasonics, Sonication, Diathermy,
Cartilage, and Wound Healing; keywords were osteoarthritis, arthritis
experimental, ultrasound therapy, low intensity pulsed ultrasound,
low intensity ultrasound, and cartilage repair. To increase the search
sensitivity, no date, language, or design limits were included. Dupli-
cates were removed after all databases were searched.
Study selection and eligibility criteria
Two reviewers (AL, JR) independently screened all citations
obtained and retrieved all parallel group randomized controlled
trials involving patients with knee OA that compared US with
placebo or no intervention. Trials that compared US in combination
with another intervention to which the comparison group was
exposed were also included. Studies were excluded fromthe review
if phonophoresis was the only ultrasonic intervention, US was
combined with another intervention not provided to the compar-
ison group, and samples included subjects having other diagnoses
and results for the subjects with knee OA were not reported sepa-
rately. Cohens unweighted Kappa (k) was used to measure agree-
ment between reviewers
12
. Observed agreement was fair
(k 0.57)
23
. Disagreement was solved by consensus including
a third reviewer (NM). Colleagues translated non-English articles
(n 9) written in their rst language (a physical therapist, an engi-
neer, an occupational therapist, a rheumatologist, and a physiatrist).
Data extraction and management
Two reviewers, using a pre-tested data collection form, fol-
lowed a double extraction method. The reviewers independently
extracted data related to the study population (clinical setting,
diagnostic criteria, joint involvement, sex, age and OA severity),
study design, US intervention (device, frequency, mode, inten-
sity and dose), co-interventions, outcomes (pain, physical func-
tion, participants perception of disease severity, and cartilage
repair), and the monitoring/reporting of adverse events. The US
dose was calculated using the following formula
5
:
We dened cartilage repair as all measures that directly or indi-
rectly assess the cartilage formationedegradation process (i.e.,
imaging, arthroscopy, uid biomarkers). Primary authors were
contacted in the case of missing data or unclear reporting. A
physiatrist with expertise in OA and US (AL) and a physical ther-
apist with expertise in research methodology and OA (NM)
reviewed the papers and extracted the data.
Risk of bias and quality assessment
The risk of bias was assessed for each study by evaluating the
rigor of the randomization process, the treatment allocation
concealment, the blinding process, the completeness of the data,
and the reporting of results following the Cochrane Collaboration
recommendations
13
. To judge the completeness of the data as
adequate, a dropout rate of less than 15% was required.
Given the objective of this systematic review, we identied
randomization, treatment allocation concealment, blinding, and
completeness of data as key domains for establishing risk of bias.
For each included study, risk of bias was determined to be low
when all key domains were performed adequately, unclear when
one or more key domains were not clearly described, and high
when one or more key domains were inadequate. Across studies,
the risk of bias was considered low if all studies had low risk,
unclear when more than 75% of the studies had unclear or low risk
and less than 25% of the studies had high risk of bias, and high
when more than 25% of the studies had high risk of bias. Observed
agreement was fair (k 0.56)
23
and disagreement was resolved by
consensus and inclusion of a third reviewer, who is a physical
therapist with research methodology expertise (JR).
The quality of the evidence for each outcome was determined by
considering the risk of bias, the heterogeneity of the ndings, the
use of surrogate measurements for outcome assessment, and the
precision of the effect estimates as recommended by the Grading of
Recommendations, Assessment Development and Evaluation
(GRADE) working group
14
.
Statistical analyses
An inverse-variance random effects model
15
was used to calcu-
late pooled standardized mean differences (SMDs) using Review
Manager (RevMan [computer program] Version 5.0 Copenhagen:
The Nordic Cochrane Centre, The Cochrane Collaboration, 2008,
Oxford, UK). The SMD is a ratio between the differences observed
Energy

J=cm
2

Average temporal intensity Time Effective radiating area
Treated surface area
A. Loyola-Snchez et al. / Osteoarthritis and Cartilage 18 (2010) 1117e1126 1118
between groups and the standard deviation (SD) of the outcomes
among participants (Hedges adjusted g). The 95% condence
interval (CI) was calculated and a Z test was performed with signif-
icance set at P <0.05. Heterogeneity among studies was assessed
with a c
2
test with signicance set at P <0.10 and an inconsistency
test (I
2
) which represents the percentage variability in the effect
estimates that is due to heterogeneity rather than chance. Incon-
sistency (I
2
) >40% was considered signicantly high
16
.
Two studies
17,18
reported different subgroups of US, so the
means and SDs of these subgroups were pooled using the following
formulas
19
:
Mean
pooled

X1 n1 X2 n2 X3 n3 X. n.
n1 n2 n3 n.
SD
pooled

s
2
1
n11

s
2
2
n21

s
2
3
n31

s
2
n
n1

n11n21n31n.1
s
where X mean for each group, SDstandard deviation,
s
2
variance, and n number of participants in that group.
Three studies
17,18,20
reported the results of pain outcomes by
knee rather than by patient. Therefore, we included the number of
patients (n 35 instead of n 70) in each group during the pooled
analysis, in order to account for the intercorrelation of measure-
ments taken from the knees of the same patient.
In order to express the effect estimates as percentage change rela-
tive to the control groups, the SMDs were back transformed to mean
differences (MD) using SDs reported in observational studies
21,22
(Appendix B). Heterogeneity was explored through subgroup
analyses following a priori hypotheses that considered factors such as
disease severity, US mode/intensity/dose, co-interventions, number of
sessions provided, and methodological adequacy.
Results
Twenty-three studies, out of 1119 citations identied, fullled
the inclusion criteria and were retrieved for full text review (see
Fig. 1). Of these, onlysix studies were included andconsidered inthe
nal analysis. The remaining 17 studies were excluded because the
samples included persons with diagnoses other than knee OA
24e26
;
compared US with short wave diathermy
27e29
, diadynamic
currents/magnetotherapy
30
, or phonophoresis
31
, without a control
group with similar exposure; compared different frequencies/
modes of US
32e34
, or used US in combination with other physical
modalities
35
without a control group; four articles were case series
reports
36e39
; and one study was performed in vitro
40
.
Description of included studies
The six small, English-language, randomized controlled tri-
als
17,18,20,41e43
included in this review are summarized in Table I. In
one study
18
, unpublished data were provided by the primary
author. All trials included people with knee OA who met the
American College of Rheumatology diagnostic criteria
44
and the
mean age of participants was over 53 years. All trials delivered
ultrasonic energy at a frequency of 1 MHz. For the study by Falconer
et al.
42
, tabulated results were extracted. Five studies reported
sufcient information for the dose calculation
17,18,20,42,43
. Three
trials
17,18,20
conducted by the same group reported a peak intensity
Fig. 1. Flowchart of study selection.
A. Loyola-Snchez et al. / Osteoarthritis and Cartilage 18 (2010) 1117e1126 1119
Table I
Summary of included randomized controlled trials evaluating the effectiveness of US in people with knee OA
Source Characteristics of subjects Intervention group Comparison group Outcomes Follow-up Risk of bias
Cetin et al.
41
(Turkey)
100% Women with mild
to severe bilateral knee
OA; w58 (8) years.
n* 18
Sonopulus 590 US device, continuous
mode, 1.5 W/cm
2
intensity, 180 J/cm
2
dosey,
24 ten-min sessions in 8 weeks.
Also: standardized warm up, isokinetic
exercises, and hotpack.
n* 17
Standardized warm up, isokinetic
exercises and hotpack.**
Pain-VASz
Physical function-LSIx
Walking speed-time to walk
50 m in seconds)
8 Weeks High
Falconer et al.
42
(USA)
w72% Women with
severity not specied;
w67.5 (11) years.
n* 34
Chattanooga Intellect 200 US device,
continuous mode{, 1.7 W/cm
2
intensity,
26 J/cm
2
dose, 12 twelve-min sessions
in 4e6 weeks.
n* 35
Sham US (start button not pushed).
**
Pain-VAS
Walking speed-50 ft walk time
(converted to m/min)
4e6 Weeks High
Huang et al.
18
(Taiwan)
w33% Women with
mild to severe bilateral knee
OA; w60 (9) years.
n* 30
Sonopulus 590 US device; pulsed mode
(duty cycle: 25%), 2.5 W/cm
2
intensity, 112 J/cm
2
dose, 24 fteen-min
sessions in 8 weeks.
n* 30
Sham US.
**
Pain-VASz
Physical function-LSIx
Walking speed-50 ft walk time
(converted to m/min)
Arthritis Severity Index ratio
of
99m
Technetium uptake#
knee/
99m
Technetium uptake middle
third ipsilateral femur
8 Weeks Unclear
Huang et al.
17
(Taiwan)
w81% Women with
mild knee OA;
w62 (17) years.
Group 1: n* 27; Continuous mode,
1.5 W/cm
2
intensity, 270 J/cm
2
dose.
Group 2: n* 30; Pulsed mode
(duty cycle: 25%), 2.5 W/cm
2
intensity, 112.5 J/cm
2
dose.
Both groups used a Sonopulus
590 US device, 24 fteen-min
sessions in 8 weeks.
Also: standardized warm up, isokinetic
exercises and hotpack
(home exercise program
following 8 weeks).
n* 25
Standardized warm up, isokinetic
exercises and hotpack
**
(home exercise program following 8 weeks).
Pain-VASz
Physical function-LSIx
Walking speed-50 ft walk time
(converted to m/min)
8 Weeks and
12 months
High
Huang et al.
20
(Taiwan)
w81% Female with
mild knee OA,
w65 (13) years.
n* 32
Sonopulus 590 US device, pulsed mode
(duty cycle: 25%),
2.5 W/cm
2
intensity, 112 J/cm
2
dose,
24 fteen-min sessions in 8 weeks.
Also: standardized warm up, isokinetic
exercises and hotpack
(home exercise program following 8 weeks).
n* 30
Standardized warm up, isokinetic
exercises and hotpack
**
(home exercise program following 8 weeks).
Pain-VASz
Physical function-LSIx
Walking speed-50 ft walk time
(converted to m/min)
8 Weeks and
12 months
High
Ozgonenel et al.
43
(Turkey)
w80% Women with
mild to moderate knee
OA; w55 (7.5) years.
n* 34
Peterson .250 US device, continuous mode,
1 W/cm
2
intensity,
150.72 J/cm
2
dose, 10 ve-min
sessions in 2 weeks.
n* 31
Sham US (applicator disconnected from device).
**
Pain-VASz
Physical function-WOMAC
k
Walking speed-time to walk 50 m
(seconds)
2 Weeks High
* Sample size of subjects whose data were included in the present metaanalysis.
y
Treated surface area was not reported and US dose was calculated using a value of 25 cm
2
reported in 1 similar trial
17
.
z
Visual Analogue Scale (VAS) from 0 cm (no pain) to 10 cm (most intense pain).
x
LSI scale from 0 (better) to 26 (worst).
k
Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) physical function subscale from 0 (better) to 68 (worst).
{
The ultrasonic mode was not reported, we assumed it was continuous based on the way the intensity of the energy was delivered (from 0 W/cm
2
to maximal tolerable dose not exceeding 2.5 W/cm
2
).
# 99m
Technetium uptake was measured by bone scan (TOSHIBA GCA-90 g-camera) 3 h after administrating the radioisotope.
** Same number of sessions and period of time as in the intervention group.
A
.
L
o
y
o
l
a
-
S

n
c
h
e
z
e
t
a
l
.
/
O
s
t
e
o
a
r
t
h
r
i
t
i
s
a
n
d
C
a
r
t
i
l
a
g
e
1
8
(
2
0
1
0
)
1
1
1
7
e
1
1
2
6
1
1
2
0
value followed by the statement: The intensity of sonication was
adjusted to the level at which the patient felt a warm sensation or
a mild sting. It can be inferred from this statement that the output
intensity was modied for each US application and the calculated
dose would be inaccurate. Communication with the primary author
conrmed that the intensity output was xed and only the speed of
the sound head varied during the US application. Thus, US dose
could be calculated.
One trial
42
did not report the mode of US, however the author
conrmed the use of continuous US (personal communication).
For the trial by Cetin et al.
41
, the therapeutic dose was calculated
using the size of the treated surface area (25 cm
2
) reported in
a trial that used the same US device
17
. Only two trials
17,20
, which
were conducted by the same research group, reported outcomes of
pain and physical function at 12 months (10 months after
completing the interventions) and this information was analyzed
separately.
Risk of bias and quality assessment
The risk of bias was unclear for one study
18
and high for ve
studies
17,20,41e43
(Table II). Thus, the evidence included in this
review has a high risk of bias overall.
The quality of evidence is low for pain and physical function
outcomes because of the high risk of bias (Table II), and the hetero-
geneity observed in results across trials [I
2
51e92%, Figs. 2(A), 3(A)
and4(A)] (Appendix B). The quality of evidence is considered lowfor
cartilage repair because the only trial
18
that reported this outcome
had an unclear risk of bias (Table II) and used a surrogate measure
(
99m
Technetium uptake) to assess cartilage status (Appendix B).
Effect of US on pain
All included trials assessed pain using a Visual Analog Scale
(VAS) measured in centimeters. Overall, the application of US
resulted in decreased pain [SMD (CI) 0.49 (0.79, 0.18),
P 0.002] [Fig. 2(A)]. However, high heterogeneity was found
(c
2
10.26, P 0.07, I
2
51%), therefore predened subgroup
analyses were conducted. Ultrasound mode, intensity, and thera-
peutic dose completely explained the inconsistency between the
groups [Fig. 2(B)]. In all subgroups, effect estimates favored US
therapy; however, the differences were statistically signicant only
in the low intensity/pulsed US and US dose <150 J/cm
2
subgroups
[SMD (CI) 0.85 (1.16, 0.54)]. Overall, trials
17,20
that reported
VAS at 12 months (10 months after completing the interventions)
favored US [SMD (CI) 0.77 (1.15, 0.39), P <0.001] [Fig. 2(C)].
Effect of US on physical function
Self-reported physical function
Five studies included self-reported physical function
measurements: four used the Lequesne Severity Index (LSI)
score
17,18,20,41
; one used the Western Ontario and McMaster
Universities Osteoarthritis Index (WOMAC) physical function
subscale score
43
(Table I). The effect estimate observed favored
the US intervention, however the difference between groups was
not statistically signicant [SMD (CI) 0.54 (1.19, 0.12),
P 0.11] [Fig. 3(A)]. Heterogeneity was high (c
2
29.11, P <0.001,
I
2
86%), and predened subgroup analyses were performed.
Inconsistency was not explained satisfactorily by any of the
subgroup analyses performed. Data pooled from two trials
17,20
showed an improvement at 12 months (10 months after
completing the interventions) in the self-reported physical
function of people who received US [SMD (CI) 1.25 (1.69,
0.81), P <0.001] [Fig. 3(B)].
Walking performance
Five studies reported walking performance: two studies
measured the time taken to walk 50 m in minutes
41,43
; three
studies measured walking speed (m/min)
17,18,20
(Table I). There was
no signicant improvement in the walking performance in the US
group [SMD (CI) 0.81 (0.09, 1.72), P 0.08] [Fig. 4(A)]. High
heterogeneity was observed (c
2
52.2, P <0.001, I
2
92%), and the
subgroup analyses did not explain this inconsistency. Pooling the
results of two studies
17,20
, by the same research group, showed that
walking speed at 12 months (10 months after completing the
interventions) was improved in the US group [SMD(CI) 1.47 (1.06,
1.88), P <0.001] [Fig. 4(B)].
Effect of US on patient-perception of disease severity
No studies were identied which reported an outcome related
to patient-perception of disease severity.
Effect of US on cartilage repair
Only one study
18
reported an outcome related to knee joint
structure. The authors measured
99m
Technetium uptake on bone
scans in order to determine an index of arthritis severity,
99m
Technetium uptake in the knee divided by the
99m
Techne-
tium uptake in the middle third of the ipsilateral femur. The
authors validated this outcome measure on animal models of
cartilage injury
45
. The results of this study showed a signicant
decrease in the index of arthritis severity measured after an 8
week US intervention for patients in the lowest and middle
Table II
Risk of bias assessment of the included randomized controlled trials
Trials Key domains Reporting of outcomes
complete
Risk of bias
Group allocation
concealment
Randomization Blinding of treatment
provider
Blinding of
participants
Completeness of data
Cetin et al.
41
Unclear Unclear No No Yes Yes High
Falconer et al.
42
Unclear Unclear No Yes Yes No-nal walking
performance not
reported by group
High
Huang et al.
18
Unclear Unclear Unclear Unclear Yes Yes Unclear
Huang et al.
17
No No No No Yes (8 weeks)
No (12 months)
Yes High
Huang et al.
20
No No No No Yes Yes High
Ozgonenel et al.
43
Unclear Unclear No Yes Yes Yes High
A. Loyola-Snchez et al. / Osteoarthritis and Cartilage 18 (2010) 1117e1126 1121
tertile for baseline measures of index of arthritis severity [MD
(CI) 0.8 (0.32, 1.28), P <0.001 and 1.8 (0.85, 2.75), P <0.001,
respectively] but not for those in the highest tertile at baseline
[MD (CI) 0.10 (1.06, 1.26), P 0.87]. We considered this
outcome an indirect measurement of cartilage status/repair. Our
search did not yield any trial reporting a direct measurement of
cartilage repair.
Adverse events
One study
42
described the intention to monitor the incidence of
adverse events related to the application of US and reported the
absence of major complications. Another study
43
reported that no
adverse events occurred either during or after the interventions.
Since the number of adverse events reported in these two trials was
Fig. 2. Meta-analyses of ultrasound effect on pain (cm-VAS). A: SMDs at the end of the intervention. B: Mode/intensity and dose subgroup analysis. C: SMDs at 12 months
(10 months after completing US).
A. Loyola-Snchez et al. / Osteoarthritis and Cartilage 18 (2010) 1117e1126 1122
zero, and adverse events were not reported in the rest of the
included trials, an estimate of US safety could not be calculated.
Discussion
This systematic review provides a meta-analysis of the efcacy
of US for decreasing pain and improving physical function in people
with knee OA. New evidence was found (Table I) which shows that
US can reduce pain by 21%, compared to a control group (Appendix
B). The clinical importance of this nding can be appreciated in
terms of the number of patients it is necessary to treat in order to
observe improvement in pain in one patient. To calculate the
number needed to treat (NNT), we used the formula proposed by
Chinn et al.
46
. The SMD was transformed to an odds ratio and the
proportion of subjects in the control group that will experience
improvement in pain (29%) was determined from a prospective
study involving a comparable patient sample
47
. Using this
approach, the NNT is 7. It also seems that US applied using low
intensity (<1 W/cm
2
), pulsed mode, and a therapeutic dose <150 J/
cm
2
could be more effective at reducing pain than US applied using
high intensity (1 W/cm
2
), continuous mode and a therapeutic
dose >150 J/cm
2
. In general, a non-signicant positive effect
(19.68% lower Lequesne Severity Index (LSI) score than the control
group) was observed on physical function with the use of US
(Appendix B). Based on the ndings fromtwo trials by one research
group
17,20
, benecial effects may last for 10 months after the US
treatment is completed [Figs. 2(C), 3(B) and 4(B)].
Limited evidence found in this review prevents conclusive
statements regarding dose effects and the effectiveness of US on
cartilage repair. Our ndings which suggest a threshold US dosage
for pain reduction in persons with knee OA are consistent with
previous ndings that higher US doses are less effective for tissue
repair in humans
48
and even harmful for the growth plate of
rabbits
49
. Only one small trial with unclear risk of bias assessed the
cartilage repair process. This study used an indirect measurement
of cartilage status in people with knee OA and reported that low
intensity pulsed US may help to enhance the cartilage repair
process on this population. These ndings together with animal
studies which permit direct measurement of cartilage tissue
response to US
7,8,9
are consistent with the mechanotransduction
theory. This theory proposes that mechanical stimuli increase the
chondrocyte production of proteoglycans and anti-inammatory
Fig. 3. Meta-analyses of ultrasound effect on self-reported physical function (LSI scores and WOMAC physical function subscale scores). A: SMDs at the end of the intervention. B:
SMDs of LSI scores at 12 months (10 months after completing US).
Fig. 4. Meta-analyses of ultrasound effect on walking performance (time to walk 50 m in minutes and walking speed in m/min). A: SMDs at the end of the intervention. B: SMDs of
walking speed (m/min) at 12 months (10 months after completing US).
A. Loyola-Snchez et al. / Osteoarthritis and Cartilage 18 (2010) 1117e1126 1123
proteins
50
. Further studies are warranted to determine if low dose
pulsed US is optimal for pain relief and stimulation of cartilage
repair.
Our ndings suggest that US delivered using a pulsed mode and
at low intensities have a pronounced effect on pain reduction. In
theory, the thermal effect of US is proposed to reduce pain
6
. The
thermal effect is achieved by applying high intensity US in
a continuous mode to heat the nerve bers which attenuates noci-
ception
5
. However, our ndings suggest a non-thermal mechanism
for painreductionwhichmaybe relatedtoa thresholddose effect, as
stated previously, or attenuation of the nociceptive signals via
mechanical stimuli. Nevertheless, the limited number and quality of
the trials included in the subgroup analyses limit further inferences.
None of the subgroup analyses that we conducted explained the
heterogeneity observed for physical function and walking perfor-
mance outcomes. Heterogeneity was reduced for self-reported
physical function (I
2
reduced from 86% to 30%) and walking
performance (I
2
reduced from 94% to 0%) when groups with
a similar severity of knee OA were compared and the US group was
favoured. However, these subgroups were comprised of subjects
with mild disease severity (Altman II), so it is unclear if these
observations are related to the comparison of homogeneous groups
or to the mild disease severity. Thus, consideration of disease status
may be important when assessing the effects of US in the knee OA
population.
The main results of this review support the ndings of
a recently updated Cochrane review
51
. However some differences
warrant comment. The authors of the Cochrane review assumed
that the US provided in the study by Falconer et al.
42
was pulsed.
For the current review, however, the authors conrmed that it
was continuous. Our search yielded one unpublished trial
18
related to our objective to evaluate the effectiveness of US on
cartilage repair. This trial was not considered in the updated
Cochrane review. Finally, we decided to analyze the data from
two trials
17,20
that reported outcomes at 12 months (10 months
after the interventions were completed) to explore the longer
term US effects. These long-term effects were not considered in
the updated Cochrane review. Despite these differences, our
results are compatible.
A limitation of our review is that the ndings are based on
evidence which has a high risk of bias and low quality. When we
synthesized only the trials that adequately blinded the partici-
pants
42,43
the effect size decreased considerably and was no longer
statistically signicant [SMD (CI) 0.24 (0.63, 0.14)]. This
suggests that the effect sizes for pain found in this review could be
partly inated by the methodological limitations of the included
studies.
A further limitation is the decision to pool the results of the
studies that compared US and placebo with studies that included
co-interventions. We assume that no interaction between US and
the isokinetic exercises occurred, however we cannot test this
assumption. It is well known that exercise is benecial in relieving
pain and improving physical function for people with knee OA
52
.
Therefore, a positive interaction between US and exercise cannot be
ruled out. Finally, because of the limited number of small trials
identied, the risk of publication bias could not be determined.
Hence, the possibility exists that only positive trials were published
while negative trials were not. Overall, the methodological limita-
tions described reduce the condence in the effect estimates
observed in the present meta-analyses.
Implications for practice
US (10e24 sessions) appears to be efcacious for decreasing
pain, and may improve physical function in patients with knee OA.
It is possible that the mode, intensity, and dose of US all inuence
the effect on pain. It is also possible that pain reduction may be
sustained for 10 months after US is discontinued. However, these
results are currently supported by low quality evidence and
denitive trials are needed.
Implications for research
Trials that are methodologically rigorous and adequately
powered are needed to conrm the effectiveness of US to reduce
pain, and improve physical function in people with knee OA.
Outcome measures in trials should include cartilage repair and
patient-perception of knee OA severity to provide insight into
potential synergistic action mechanisms. Careful consideration of
ultrasound prescription and disease stage is required to assess the
optimal therapeutic parameters and the subgroup(s) of people
who will benet most. Long-term effects of the US intervention
should be assessed as well. Finally, the mechanism by which
therapeutic ultrasound reduces pain in knee OA needs to be
explored further.
Author contributions
All authors made substantial contributions to the conceptuali-
zation, design, data collection, analysis, interpretation, drafting and
revisions; and approved the nal version.
Conict of interest
None of the authors has any nancial and personal relationships
with other people or organizations that could potentially and
inappropriately inuence this work and its conclusions.
Acknowledgement
Role of funding sources: AL was supported by the Consejo
Nacional de Ciencia y Tecnologia (CONACYT) of Mexico scholarship
(number 209621), and by a McMaster University School of
Graduate Studies International Excellence Award. These sponsors
had no involvement in the design, conduct or publication of this
study.
Appendix A. Medical Subject Headings and keyword search
strategy performed in Ovid MEDLINE (1950 to January week 4,
2009)
1 Exp Osteoarthritis/
2 Osteoarthritis.mp.
3 Exp Arthritis, Experimental/
4 Exp Arthritis/
5 Arthritis experimental.mp.
6 Exp Wound Healing/
7 Ultrasonic Therapy/
8 Exp Ultrasonics/
9 Ultrasound therapy.mp.
10 Ultrasonic therapy.mp.
11 Exp Sonication/
12 Sonication.mp.
13 Low intensity pulsed ultrasound.mp.
14 Low intensity ultrasound.mp.
15 Exp Diathermy/
16 Cartilage, Articular/
17 Exp Cartilage/
18 Cartilage repair.mp.
19 6 or 1 or 18 or 3 or 16 or 17 or 2 or 5
20 11 or 7 or 9 or 12 or 15 or 14 or 8 or 10 or 13
22 20 and 19
A. Loyola-Snchez et al. / Osteoarthritis and Cartilage 18 (2010) 1117e1126 1124
References
1. Pop T, Szczygielska D, Druzbicki M. Epidemiology and cost of
conservative treatment of patients with degenerative joint
disease of the hip and knee. Ortop Traumatol Rehabil
2007;9:405e12.
2. Quintana JM, Arostegui I, Escobar A, Azkarate J, Goenaga JI,
Lafuente I. Prevalence of knee and hip osteoarthritis and the
appropriateness of joint replacement in an older population.
Arch Intern Med 2008;168:1576e84.
3. Zhang Y, Jordan JM. Epidemiology of osteoarthritis. Rheum Dis
Clin North Am 2008;34:515e29.
4. Zhang W, Moskowitz RW, Nuki G, Abramson S, Altman RD,
Arden N, et al. OARSI recommendations for the management of
hipandkneeosteoarthritis, part II: OARSI evidence-based, expert
consensus guidelines. Osteoarthritis Cartilage 2008;16:137e62.
5. Blanger AY. Ultrasound. In: Evidence-based Guide to
Therapeutic Physical Agents. Philadelphia PA, USA: Lippincott
Williams & Wilkins; 2003:223e61.
6. Baker KG, Robertson VJ, Duck FA. A review of therapeutic
ultrasound: biophysical effects. Phys Ther 2001;81:1351e8.
7. Cook SD, Salkeld SL, Patron LP, Doughty ES, Jones DG. The effect
of low-intensitypulsedultrasoundonautologous osteochondral
plugs in a canine model. Am J Sports Med 2008;36:1733e41.
8. Huang MH, Ding HJ, Chai CY, Huang YF, Yang RC. Effects of
sonication on articular cartilage in experimental osteoarthritis.
J Rheumatol 1997;24:1978e84.
9. Singh KI, Sobti VK, Roy KS. Gross and histomorphological
effects of therapeutic ultrasound (1 watt/cm
2
) in experimental
acute traumatic arthritis in donkeys. J Equine Vet Sci
1997;17:150e5.
10. Welch V, Brosseau L, Peterson J, Shea B, Tugwell P, Wells G.
Therapeutic ultrasound for osteoarthritis of the knee.
Cochrane Database Syst Rev 2001;003132.
11. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC,
Ioannidis JP, et al. The PRISMA statement for reporting
systematic reviews and meta-analyses of studies that evaluate
health care interventions: explanation and elaboration. J Clin
Epidemiol 2009;62:e1ee34.
12. Cohen JA. A coefcient of agreement for nominal scales. Educ
Psychol Meas 1960;20:37e46.
13. Higgins JPT, Altman DG. Chapter 8: assessing risk of bias in
included studies. In: Higgins JPT, Green S, Eds. Cochrane
Handbook for Systematic Reviews of Interventions. Version
5.0.1 (updated September 2008). 5.0.1 ed, www.cochrane-
handbook.org; 2008. The Cochrane Collaboration, 2008.
14. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-
Coello P, et al. GRADE: an emerging consensus on rating
quality of evidence and strength of recommendations. BMJ
2008;336 (7650:ate of Pubaton: 26 Ar 2008).
15. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control
Clin Trials 1986;7:177e88.
16. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring
inconsistency in meta-analyses. BMJ 2003;327:557e60.
17. Huang M, Lin Y, Lee C, Yang R. Use of ultrasound to increase
effectiveness of isokinetic exercise for knee osteoarthritis.
Arch Phys Med Rehabil 2005;86:1545.
18. Huang M, Chen T, Weng M, Wang Y. In: Peek WJ, Lankhorst GJ,
Eds. Effects of Pulse Sonication on Functional Status of Patients
with Knee Osteoarthritis. International Society of Physical and
Rehabilitation Medicine. Amsterdam, The Netherlands:
Monduzzi; 2001 Jul:297e300.
Appendix B. Summary of ndings and quality of evidence assessment
Outcomes Control group
mean
Ultrasound group mean with respect to
the control group (CI)
Relative changek (CI) Quality of
the evidence*,y
Comments
Pain
VAS from 0 (no pain) to 10
(intense pain)
Follow up: 2e8 weeks
3.95z 0.84 Lower (1.2e0.48 lower) 21% Lower
(30%, 12%)
Low High risk of bias of the included
studies and considerable results
heterogeneity.
Follow-up: 12 monthsx 3.95z 1.22 Lower (1.62e0.82 lower) 30% Lower
(40%, 20%)
Low Blinding issues and lack of
completeness in the follow-up
data (<85% of participants) detected.
Physical function
LSI scale from 0 (better) to
26 (worst)
Follow-up: 2e8 weeks
5.65z 1.11{ Lower (2.45 lower to 0.24 higher)
Mild severity (grade II) subgroup
2.7{ lower (4.03e1.38 lower)
19.68% Lower (89.32%
lower to 8% higher)
47% Lower
(71%, 24%)
Low High risk of bias of the included
studies and considerable results
heterogeneity.
Follow-up: 12 monthsx
LSI
5.65z 2.07{ Lower (2.91e1.22 lower) 36% Lower
(51%, 21%)
Low
Walking performance
Walking speed (m/min)
Follow-up: 2e8 weeks
82.47z 4.64{ Higher (0.5 lower to 9.85 higher)
Mild severity (grade II) subgroup
11.17{ higher (14.32e8.07 higher)
5% Higher (0.6% lower
to 12% higher)
13% Higher
(17%, 10%)
Low High risk of bias of the included
studies and considerable results
heterogeneity.
Follow-up: 12 monthsx
Walking speed (m/min)
82.47z 10.89{ Higher (7.07e14.71 higher) 13% Higher
(8%, 18%)
Low
Cartilage repair
Follow-up: 8 weeks
Arthritis Severity Index
(smaller values mean
less severity)
4.9 1.8 Lower (2.47e1.13 lower) 36% Lower
(50%, 23%)
Low High risk of bias and indirectness
of the outcome measure
(we consider bone scan as a
surrogate measurement of
cartilage repair).
* We considered that the probability of having a publication bias in this review is low.
y
GRADE Working Group grades of evidence
15
. High quality: Further research is very unlikely to change our condence in the estimate of effect. Moderate quality: Further
research is likely to have an important impact on our condence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an
important impact on our condence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
z
The nal score mean for the control groups was calculated by pooling the means and standard errors through a generic inverse-variance method.
x
All patients included in these studies had mild knee OA (Altman Grade II) and completed a home-exercise program after 2 months of treatment/sham US.
k
Relative change %

Mean difference
Final control mean

100.
{
MD were calculated through a back transformation of the SMD using the SD reported in Villanueva et al.
21
[Lequesne severity index 2.06] and Huang et al.
22
[walking
speed 5.73].
A. Loyola-Snchez et al. / Osteoarthritis and Cartilage 18 (2010) 1117e1126 1125
19. Mendenhall W, Beaver RJ, Beaver BM. Introduction to Proba-
bility and Statistics. 10th edn. Pacic Grove, Calif.: Duxbury
Press; 1999.
20. Huang MH, Yang RC, Lee CL, Chen TW, Wang MC. Preliminary
results of integrated therapy for patients with knee osteoar-
thritis. Arthritis Rheum 2005;53:812e20.
21. Villanueva I, del Mar Guzman M, Javier Toyos F, Ariza-Ariza R,
Navarro F. Relative efciency and validity properties of a visual
analogue vs a categorical scaled version of the western ontario
and McMaster universities osteoarthritis (WOMAC) index:
Spanish versions. Osteoarthritis Cartilage 2004;12:225e31.
22. Huang MH, Lin YS, Yang RC, Lee CL. A comparison of various
therapeutic exercises on the functional status of patients with
knee osteoarthritis. Semin Arthritis Rheum 2003;32:398e406.
23. Orwin RG. Evaluating coding decisions. In: Cooper H,
Hedges LV, Eds. NewYork (NY): Russell Sage Foundation; 1994.
24. Esmat N. Treatment of arthrosis deformans by simultaneous
application of interferential current and ultrasonic waves.
J Egypt Med Assoc 1975;58:328e33.
25. Sabadyshin RA, Rudyk BI, Filchagin NM. Effectiveness of the
method of ultraphonophoresis of sex hormones in osteo-
arthrosis deformans. Revmatologiia (Mosk) 1988;37e9.
26. Svarcova J, Zvarova J, Kouba A, Trnavsky K. Does physiotherapy
affect thepaininactivatedarthrosis?ZPhysiother 1988;40:333e6.
27. Bansil CK, Joshi JB. Effectiveness of shortwave diathermy and
ultrasound in the treatment of osteo-arthritis of the knee joint.
Med J Zambia 1975;9:138e9.
28. Jan MH, Lai JS. The effects of physiotherapy on osteoarthritic
knees of females. J Formos Med Assoc 1991;90:1008e13.
29. Kalpakcioglu BA, Cakmak B, Bahadir C. Comparison of ultra-
sound andshort wave diathermy therapy inknee osteoarthritis.
Turk J Phys Med Rehab 2006;52:168e73.
30. Lisinski P, Zapalski W, Stryla W. Physical agents for pain
management in patients with gonarthrosis. Ortop Traumatol
Rehabil 2005;7:317e21.
31. KozanogluE, BasaranS, Guzel R, Guler-Uysal F. Short termefcacy
of ibuprofenphonophoresis versuscontinuous ultrasoundtherapy
in knee osteoarthritis. Swiss Med Wkly 2003;133:333e8.
32. Golubenko TA. Low-frequency ultrasound in the treatment of
osteoarthrosis patients. Vopr Kurortol Fizioter Lech Fiz Kult
1991;36e9.
33. Grifn JE, Echternach JL, Bowmaker KL. Results of frequency
differences in ultrasonic therapy. Phys Ther 1970;50:481e6.
34. Winterfeld HJ, Conradi E. Clinical comparison of the effect of
ultrasound using direct and alternating current in the treat-
ment of gonarthrosis. Z Physiother 1981;33:159e63.
35. Grigoreva VD, Fedorova NE, Kiselev VI. The combined use of
cryogenic exposure and ultrasound in patients with arthrosis
of the joints of the legs. Vopr Kurortol Fizioter Lech Fiz Kult
1996;18e21.
36. Bernau A, Kruppa G. Low frequency electro-stimulation and
ultrasonic therapy (authors transl). Z Orthop Ihre Grenzgeb
1981;119:126e37.
37. Scala DA, Trinchieri P. Ultrasonics combined with other
physical therapy in osteoarthrosis; statistical study. Radioter
Radiobiol Fis Medica 1954;9:182e90.
38. Schwartz FF. Ultrasonics in osteoarthritis. J Med Assoc State
Ala 1953;22:182e4.
39. Soren A. Treatment of musculoskeletal disorders with ultra-
sound. J Occup Med 1965;7:434e8.
40. Ryaby JT, Cai FF, Culley PL, Kaufman JJ, Lippiello L. In: Bersani F,
Ed. Mechanical Stimulation of Cartilage by Ultrasound. Elec-
tricity and Magnetism in Biology and Medicine. Bologna, Italy:
Kluwer; 1997 Jun:947e50.
41. Cetin N, Aytar A, Atalay A, Akman MN. Comparing hot pack,
short-wave diathermy, ultrasound, and TENS on isokinetic
strength, pain, and functional status of women with osteoar-
thritic knees: a single-blind, randomized, controlled trial. Am J
Phys Med Rehabil 2008;87:443e51.
42. Falconer J, Hayes KW, Chang RW. Effect of ultrasound on
mobility in osteoarthritis of the knee. A randomized clinical
trial. Arthritis Care Res 1992;5:29e35.
43. Ozgonenel L, Aytekin E, Durmusoglu G. A double-blind trial of
clinical effects of therapeutic ultrasound in knee osteoarthritis.
Ultrasound Med Biol 2009;35:44e9.
44. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K,
et al. Development of criteria for the classication and
reporting of osteoarthritis. Classication of osteoarthritis of
the knee. Diagnostic and therapeutic criteria committee of the
american rheumatism association. Arthritis Rheum 1986;29:
1039e49.
45. Huang MH, Ding HJ, Yang CC, Chai CY, Yang RC. The early
evaluation of induced osteoarthritis in rats with 99Tcm-
pertechnetate scans. Nucl Med Commun 1996;17:529e35.
46. Chinn S. A simple method for converting an odds ratio to
effect size for use in meta-analysis. Stat Med 2000;19:
3127e31.
47. Peters TJ, Sanders C, Dieppe P, Donovan J. Factors associated
with change in pain and disability over time: a community-
based prospective observational study of hip and knee osteo-
arthritis. Br J Gen Pract 2005;55:205e11.
48. Byl NN, McKenzie A, Wong T, West J, Hunt TK. Incisional
wound healing: a controlled study of low and high dose
ultrasound. J Orthop Sports Phys Ther 1993;18:619e28.
49. Lyon R, Liu XC, Meier J. The effects of therapeutic vs. high-
intensity ultrasound on the rabbit growth plate. J Orthop Res
2003;21:865e71.
50. Choi BH, Choi MH, Kwak MG, Min BH, Woo ZH, Park SR.
Mechanotransduction pathways of low-intensity ultrasound in
C-28/I2 human chondrocyte cell line. Proc Inst Mech Eng [H]
2007;221:527e35.
51. Rutjes AWS, Nesch E, Sterchi R, Jni P. Therapeutic ultrasound
for osteoarthritis of the knee or hip. Cochrane Database Syst Rev
2010;(1):CD003132:doi:10.1002/14651858.CD003132.pub2.
52. Fransen M, McConnell S. Exercise for osteoarthritis of the knee.
Cochrane Database Syst Rev 2008;4.
A. Loyola-Snchez et al. / Osteoarthritis and Cartilage 18 (2010) 1117e1126 1126