609415

research-article2015
CRE0010.1177/0269215515609415Clinical RehabilitationZhang et al.

CLINICAL
Article REHABILITATION

Clinical Rehabilitation

Effects of therapeutic ultrasound 2016, Vol. 30(10) 960­–971

© The Author(s) 2015
Reprints and permissions:
on pain, physical functions and sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0269215515609415

safety outcomes in patients with cre.sagepub.com

knee osteoarthritis: a systematic

review and meta-analysis

Chi Zhang1,2,*, Yujie Xie2,*, Xiaotian Luo1,3, Qiaodan Ji1,3,

Chunlan Lu1,3, Chengqi He1,3 and Pu Wang1,3

Abstract
Objective: To explore the effects of therapeutic ultrasound with sham or no intervention on pain,
physical function and safety outcomes in patients with knee osteoarthritis.
Data sources: This systematic review was searched on CENTRAL, EMBASE, MEDLINE, CINAHL,
Physiotherapy Evidence Database, Open Gray on 4 September 2015. Trials included randomized
controlled trials that compared therapeutic ultrasound with a sham or no intervention in patients with
osteoarthritis of the knee.
Review methods: Eligible trials and extracted data were identified by two independent investigators.
Standardized mean differences (SMDs) and 95% confidence interval (CI) were calculated for pain and
physical function outcomes. Heterogeneity was assessed by the I2 test and inverse-variance random-
effects analysis was applied to all trials.
Results: Ten randomized controlled trials (645 patients) met the inclusion criteria. Therapeutic ultrasound
showed a positive effect on pain (SMD = −0.93, 95%, CI = −1.22 to −0.64, p < 0.01, p for heterogeneity = 0.12,
I2 = 42%). For physical function, therapeutic ultrasound was advantageous for reducingWestern Ontario
and McMaster Universities physical function score (SMD = −0.37, 95% CI = −0.73 to −0.01, p = 0.04, p for
heterogeneity = 0.94, I2 = 0%). In terms of safety, no occurrence of adverse events caused by therapeutic
ultrasound was reported in any trial.
Conclusion: The authors suggested that therapeutic ultrasound is beneficial for reducing knee pain and
improving physical functions in patients with knee osteoarthritis and could be a safe treatment.

*Both the authors (Chi Zhang, Yujie Xie) have equal
contribution in this article
1Rehabilitation Medicine Center, West China Hospital,
Sichuan University, Chengdu, Peoples’ Republic of China
2Rehabilitation Medicine Department, Luzhou Medical
Colleage Affiliated Hosipital, Luzhou, Peoples’ Republic of
China
3Department of Rehabilitation Medicine, West China Hospital,
Corresponding authors:
Chengqi He and Pu Wang, Rehabilitation Medicine Center,
West China Hospital, Sichuan University, Guoxue Road,
610041 Chengdu, Sichuan, People’s Republic of China.
Email: 1198455105@qq.com; wangpu0816@qq.com

Sichuan University, Chengdu, Peoples’ Republic of China

Zhang et al. 961
Keywords
Knee, osteoarthritis, ultrasonic therapy, systematic review, meta-analysis
Received: 14 May 2015; accepted: 5 September 2015
Introduction
Osteoarthritis is the primary cause of pain and
motor disability in adults worldwide.1 It is a multi-
factor disease characterized by degradation and ero-
sion of articular cartilage, inflammation of synovial
membrane, sclerosis of subchondral bone and for-
mation of osteophytes.2 Therapeutic ultrasound is a
deep heating physical therapy for patients with knee
osteoarthritis.3 The treatment is based on the appli-
cation of mechanical energy, which is produced by
sound waves at different frequencies. Therapeutic
ultrasound can induce biological responses, includ-
ing muscle relaxation, induction of tissue regenera-
tion and reduction of inflammation.4
A systematic review of Cochrane library
reported that therapeutic ultrasound may be benefi-
cial for patients with osteoarthritis of the knee, but
it was uncertain about the magnitude of the effects
on pain relief and function.5 This systematic review
also pointed out that the pooled effects may influ-
ence the final results with a limited number of
included trials, and the confidence level in the
effect will hence reduce. In the last five years, more
trials have been designed to evaluate the effective-
ness of therapeutic ultrasound on knee osteoarthri-
tis. Therefore, an increased number of high quality
clinical trials should be included in an updated sys-
tematic review. The Osteoarthritis Research
Society International guideline for the non-surgical
management of knee osteoarthritis in 2014 regarded
therapeutic ultrasound as an uncertain physical
therapy.6 It is important to review present evidence
to verify the result.
The aim of our systematic review is to explore
the therapeutic ultrasound with sham or no inter-
vention on pain, physical function and safety out-
comes in patients with knee osteoarthritis.
Methods
We searched six electronic databases, namely
CENTRAL (via The Cochrane Library), MEDLINE
(via Ovid platform), EMBASE (via Ovid plat-
form), CINAHL (via EBSCOhost), Physiotherapy
Evidence Database (via PEDro) and Open Gray
(http://www.opengrey.eu/) for the gray literature.
All data recorded from the implementation of these
databases to 4 September 2015 were searched.
Conference proceedings were searched manually
using keywords and text words related to therapeu-
tic ultrasound and osteoarthritis. A detailed exam-
ple of the full electronic search strategy for Ovid
MEDLINE is provided in Appendix A, available
online.
Study selection
The eligibility of all titles and abstracts were evalu-
ated by two reviewers independently, with disagree-
ments resolved through discussions. If necessary, a
third reviewer was involved. Only randomized con-
trolled trials that compare therapeutic ultrasound
with sham ultrasound or no intervention on pain and
physical function in patients with knee osteoarthritis
were included. There are no language limits. Titles
and abstracts obtained were screened to select rele-
vant articles. Subsequently, the full texts of the
remaining articles were read in detail to iden-
tify their eligibility. The inclusion criteria were as
follows.
1. Participate: The diagnosis was established by
American College of Rheumatology diagnostic
criteria for knee osteoarthritis;7 there were no
restrictions on age, gender, ethnicity and setting.
2. Intervention: There were no limits in type,
dose, frequency, mode of therapeutic ultra-
sound; Intervention groups of concurrent treat-
ment were excluded, such as Phonophoresis
ultrasound, short wave and magnetic therapy;
therapeutic ultrasound combining hot pack,
knee exercise therapy, manual therapy and
joint protection were included.
962 Clinical Rehabilitation 30(10)
3. Comparisons: The control group should include
sham group and/or no intervention group.
4. Outcome: Primary outcomes were pain and
physical function. Pain including visual ana-
logue scale,8 Western Ontario and McMaster
Universities (WOMAC) pain subscale,
Function measurement including WOMAC
physical function score;9 Lequesne Indexes10
and other functional scales; adverse events
including the number of patients experiencing
any adverse event caused by therapeutic
ultrasound.
Trials would be excluded if they were:
(1) reports published as conference proceedings;
(2) reports in books.
Data extraction
Data were extracted by two reviewers indepen-
dently. A test data form was used to extract the fol-
lowing data: study population (first author, year of
publication, clinical setting, diagnostic criteria,
sample size and age), therapeutic ultrasound inter-
vention (device, frequency, mode, intensity and
dose), combined intervention, outcome measures
(characteristics of pain, physical function, safety
outcomes, time points and follow-up). Extracted
data were entered into a database (Table 1) and
checked by another investigator.
Assessment of risk of bias
Two reviewers independently assessed the risk of
bias in included trials, with disagreements being
resolved through consensus. The risk of bias was
assessed in accordance with methods recommended
by the Cochrane Collaboration8 (risk of bias in
included trials). Moreover, figures generated by the
RevMan5.3 (http://ims.cochrane.org/revman) were
presented to provide summary assessments of the
risk of bias. In addition, the Physiotherapy Evidence
Database (PEDro) tool (http://www.pedro.fhs.usyd.
edu.au/scale_item.html) was also used to assess the
methodological quality of individual randomized
controlled trial.11 Scores of nine trials were pro-
vided by the PEDro website, only one of which was
assessed by our research group.12
Statistical analyses
For continuous outcomes, a mean change between
after and before treatment was calculated by stand-
ardized mean difference (SMD) with 95% confi-
dence intervals (CIs). The means and deviations of
trials, in which both groups of continuous and
pulsed modes existed, were pooled. Based on ultra-
sound mode and treatment duration, subgroup
analyses were conducted while sufficient trials
were taken into each subcategory. Sources of het-
erogeneity were analyzed by a I2 statistical test. I2
values of 25%, 50% and 75% may be interpreted as
low, moderate and high between-trial heterogene-
ity.12 A random-effect model was used regardless
of heterogeneity. If the I2 test was not significant
(p > 0.05), potential sources of heterogeneity would
be identified by sensitivity analyses that were con-
ducted by omitting one study and investigating the
influence of a single study on the overall pooled
estimate. P < 0.05 was considered statistically sig-
nificant. All statistical analyses were conducted by
RevMan5.3.
Results
A total of 2493 references were identified, 539 of
which were excluded for duplicated trials.
Furthermore, a total of 1870 trials were excluded
after screening the titles and abstracts. Another 71
trials were eliminated for various reasons (reviews,
letter, non-randomized trials, not relevant to our
analysis, or combined using other physical treat-
ment). Moreover, three were excluded owing to the
absence of sham group or blank control group.13–15
Finally, ten trials (645 patients) that met our eligibil-
ity criteria were selected for this systematic review.
Figure 1 shows the flow of information through the
different phases of our systematic review. This flow
figure followed the Preferred Reporting Items for
Systematic reviews and Meta-Analyses (PRISMA)
guidance.16 Two reports17,18 that described the same
trial were included in our systematic review.
Description of trials
Descriptive data for the trials in this systematic
review are shown in Table 1. Ten trials (N = 645)
Table 1.  Summary of included randomized controlled trials evaluating the effectiveness of therapeutic ultrasound in people with knee osteoarthritis.
Investigators Characteristics of Intervention group Comparison group Outcomes Follow-up
subjects

1. Judith Setting: clinic Group 1, N = 34; Group 2, N = 35; VAS; 6 weeks

Zhang et al.

Falconer Age: 67.5 Pulsed US, 1 MHz, 1.7 W/cm2, 26 J/cm2,12 minute/session in Sham US 50 meters
1991 ±13 years 4–6 weeks (12 sessions) isometric exercise, mobilizing exercise, manual walking time
USA (Chattanooga Intellect 200 ultrasound unit) therapy and joint protection
2. Levent Setting: clinic Group 1, N = 34; Group 2, N = 33; 50 meters 2 weeks
Özgönenel Age: 45~65 years Continuous US, 1 MHz, 1 W/cm2, 5 minute/session in 2 weeks Sham US walking time;
2008 (54.8 ±7) (10 sessions) The physiotherapy program was conducted WOMAC;
Turkey (Petson®.250ultrasoundequipment,Petas,Turkey) five times a week for 2 weeks VAS
3. F tascioglu Setting: clinic Group 1, N = 27; Group 3, N = 27; WOMAC; 2 weeks
2010 Age: 54~70 years Continuous US, 1 MHz, 2 W/cm2, 5 minute/session in 2 weeks Sham US VAS
Turkey (10 sessions)
Group 2, N = 28;
pulsed US, 1 MHz, 2 W/cm2, pulsed mode duty cycle=1:4, 5 minute/
session in 2 weeks (10 sessions); (Sonopuls 434; EnrafNonius,
Delft, The Netherlands)
4. Peng-fei Yang Setting: clinic Group 1, N = 50; Group 2, N = 50; VAS; 4 weeks
2011 Age: 38~81 years Treatment method: unclear Sham US LI
China National Engineering Research Center of Ultrasound Medicine
(NERCUM, Chongqing, China).
5.Adalberto Setting: clinic. Group 1, N = 14; Group 2, N = 13; WOMAC; 8 weeks
Loyola-Sánchez Age: ⩾45 years pulsed US, 1 MHz, 0.2 W/cm2 Sham US 6MWT
2012 9.5 minute/session in 8 weeks (24 sessions) 6MWT pain
canada Lower Extremity
Functional Scale
6. Yasemin Ulus Setting: Inpatients Group 1, N = 20; Group 2, N = 20; 50 meters 3 weeks
2012 clinic Continuous US, 1 MHz,1.5 W/cm2, Sham US walking time;
Turkey 40~80 years 10 minute/session in 3 weeks (15 sessions) Both group received 20 minute hotpacks, knee range of
ASonopuls434USmachine (EnrafNonius, Rotterdam, The 10 minute interferential current, and 15 minute motion;
Netherlands) quadriceps isometric exercise VAS;
WOMAC;
LI;
Hospital Anxiety
and Depression
Scale
7. Cakir S Setting: Group 1, N = 20; Group 3, N = 20; Pain (VAS); 6 months
2014 Outpatients Continuous US, 1 MHz,1.5 W/cm2, Sham US 15 meter walking
Turkey 40~80 years 12 minute/session in 2 weeks (10 sessions) Quadriceps isometric exercise, muscle time;
Group 2, N = 20; strength exercises, stretching exercises of the WOMAC
Pulse US, 1 MHz, 1.5 W/cm2, pulser atios = 1:4 lower extremity muscles, 3 times per week
12 minutes/session in 2 weeks (10 sessions)
(Enraf Nonius Sonoplus
190)
963
 964

Table 1. (Continued)
Investigators Characteristics of Intervention group Comparison group Outcomes Follow-up
subjects

8. Huang MH Setting: clinic; Group 1, N = 35 Group 1, N = 35 LI; 1 year

2005a Age: 40~77 year Isokinetic muscular strengthening exercises Isokinetic muscular strengthening exercises VAS;
Taiwan (65.0 ±6.4) 15 minutes/session in 8 weeks, 3 times a week 15 minutes/session in 8 weeks (24 sessions) muscle peak
Group 2, N = 35 Group 3, N = 35 torques;
Pulse US and isokinetic exercise Intra articular hyaluronan ambulation
1 MHz, 2.5 W/cm2 Group 4, N = 35 speed;
15 minutes/session in 8 weeks, 3 times a week Control group knee range of
(Sonopulus 590; Enraf Nonius, Only warm-up exercises motion
AL Delft, Netherlands) All the research groups receive Also: warm up
exercise (20 min) and passive ROM exercise
(electric stationary bike, 5 minutes)
9.Huang MH Setting: clinic; Group 3, N = 30 Group 1, N = 30 LI; 1 year
2005b Age: 40~77 year Pulse US and isokinetic exercise Isokinetic muscular strengthening exercises VAS;
Taiwan (65.0 ±6.4) 1 MHz, 1.5 W/cm2 15 minutes/session in 8 weeks (24 sessions) muscle peak
15 minutes/session in 8 weeks (24 sessions) Group 4, N = 30 torques;
Group 2, N = 30 Control group ambulation
Continuous US and isokinetic exercise Only warm-up exercises speed;
1 MHz, 2.5 W/cm2 All the research groups receive Also: warm knee range of
15 minutes/session in 8 weeks (24 sessions) up exercise (20 minutes) and passive ROM motion
exercise (electric stationary bike, 5 minutes)
10. Cetin N Setting: clinic; Group 3, N = 20 continuous ultrasound + hotpacks and isokinetic Group 1, N = 20 short-wave diathermy + hot Index of severity 8 weeks
2008 Age: 59.82 exercise; Continuous mode 1 MHz, 1.5 w/cm2 packs and isokinetic exercise; Group 2, for knee
Turkey ±9.05 years 10 minutes/session in 8 weeks (24 sessions) N = 20 transcutaneous electrical nerve osteoarthritis;
Sonopuls 590US machine (Enraf-NoniusB Delftechpark 39) stimulation + hot packs and isokinetic exercise; LI;
Group 4, N = 20 VAS;
hotpacks and isokinetic exercise; 50 meters
Group 5, N = 20 walking time
Control group
All the research groups had the hot packs
wrapped in toweling (20 minutes) and passive
ROM exercise (electric stationary bike,
5 minutes)

6-minute walk test (6MWT); LI: Lequesne efficacy indexes; OA: osteoarthritis; US: ultrasound; VAS: visual analogue scale; WOMAC: Western Ontario and McMaster Universities physical function
score.
Clinical Rehabilitation 30(10)
Zhang et al. 965
Figure 1.  Flow of information through the different phases of our systematic review.
were included and participants in all trials were
enrolled from the clinical setting. The ages of
patients in nine trials were more than 40 years old,
except one study brought in a 38-year-old patient.12
Two trials were evaluated for pulsed ultrasound,17,19
three for continuous ultrasound20–22 and three for
both.18,23,24 For frequency and intensity of therapeu-
tic ultrasound, 1 MHz were applied in all trials, the
intensity of which ranged from 0.2–2.5 Watts cm-2.
We could not confirm the ultrasound mode
through the article in the trial Yang et al.,12 as we
failed to contact the author. The trial (Falconer
et al., 1992)25 did not provide sufficient information
for the therapeutic ultrasound treatment mode and
dose. What is more, the previous two systematic
review5,26 have different opinions in the treatment
mode of the trial. The perspective of the systematic
review written by Loyola-Sánchez et al.26 was cho-
sen because they confirmed the continuous mode
by personal communication.
Risk of bias within trials and across trials
Two risks of bias assessment tools were used in our
systematic review. The risk of bias assessment
results of the Cochrane library assessment tool
were shown in Figure 2 (available online), whereas
PEDro methodological quality assessment result
was displayed in Table 2 (available online).
Only one trial completed an adequate sequence
and group allocation through a simple-computer-
ized random-number generation and telephone.19
Cakir et al. used the coin randomization method,
but the detail information was unclear.24 Although
966 Clinical Rehabilitation 30(10)
Tascioglu et al. reported a security system of
opaque closed envelopes numbered 1–323 and
Huang MH reported number I–IV,17,18, the infor-
mation about who opens the envelopes and their
allocation were unclear. Five trials provided insuf-
ficient details on the randomization and
concealment.12,20–22,25
When it came to blinding, sham therapeutic
ultrasound in eight trials were administered by a
sham device without delivering any ultrasound
output. Loyola-Sánchez et al.26 took no sound-head
crystal into the sham device. Two trials (Özgönenel
et al.22 and Ulus et al.21) used an applicator discon-
nected from the back on a working ultrasound
machine. Cakir et al.24 applied the same ultrasound
device of treatment group, but with the power
switched off. Two trials (F Tascioglu et al.23 and
Yang et al.12) provided insufficient information to
ensure the sham device. Failure of patient blinding
was caused by the absence of a warm or stinging
sensation in the control group in the trial of Huang
MH18. Falconer et al.25 designed the study where
the patient could hear an audible beep at each sham
device increase in the dosage.
Three trials reporting more than six months fol-
low-up were considered as adequate.17,18,24 Only
one study used intention-to-treat analysis in clini-
cal outcomes analysis.26 Three trials did not receive
any support from commercial organizations,12,21,23
The treatment equipment of one trial was supplied
by commercial organizations. The source of fund-
ing was unclear in other six trials.18,20,22,23,25,26
Synthesis of results
Pain. Nine trials assessed pain by using a visual
analogue scale. Eight of all those trials were meas-
ured in centimeters, but two of the eight trials pro-
duced insufficient statistical data.12,25 At last, six
trials involving 224 patients contributed to the
pooled analysis of pain outcomes (Figure 3). The
aggregated results of these six trials suggested that
therapeutic ultrasound could reduce pain of knee
osteoarthritis patients (SMD  =  −0.93, 95%
CI = −1.22 to −0.64, p < 0.01, p for heterogene-
ity = 0.12, I2 = 42%). Subgroup analysis was con-
ducted based on different modes (continuous mode
and pulsed mode) and durations. For the continu-
ous mode, the aggregated results of five trials sug-
gested that therapeutic ultrasound would reduce
pain, with a SMD of −0.83 (95% CI = −1.16 to
−0.51) by analysis. As for the pulsed mode, the
aggregated results of three trials suggested that
therapeutic ultrasound may reduce pain, with anal-
ysis suggesting a SMD of −1.04 (95% CI = −1.34 to
−0.73). In addition, the three trials that lasted for
less than four weeks were pooled. The aggregated
results suggested that therapeutic ultrasound would
reduce pain, with analysis indicating a SMD of
−1.51 (95% CI = −1.93 to –1.08). Lastly, the aggre-
gate results of eight-week treatments were indica-
tive of the possibility of therapeutic ultrasound
reducing pain, with analysis suggesting a SMD of
−1.13 (95% CI = −1.50 to −0. 76). (Figure 4).
Physical function.  The WOMAC physical function
subscale was utilized in four trials to assess the
physical function. However, one of them reported
an unrealistic standard deviation.22 Therefore,
only three trials contributed in the pooled analy-
sis. The aggregated results of these trials sug-
gested that therapeutic ultrasound was associated
with significantly reduced disability in physical
function (SMD = −0.37, 95% CI = −0.73 to −0.01,
p = 0.04, p for heterogeneity = 0.94, I2 = 0%). Also,
five other trials used Lequesne Indexes to assess
the physical function. As one of them provided
insufficient data, only the remaining four were
conducted in the pooled analysis. Their aggre-
gated results suggested that therapeutic ultra-
sound was associated with reduced disability in
physical function (SMD = −0.88, 95% CI = −1.46
to 00.30, p < 0.0001, p for heterogeneity = 0.005,
I2 = 77%. Ultimately, three trials used 50 meters
walking time to assess physical function. The
aggregated results of these trials suggested that
therapeutic ultrasound could not decrease the
walking time (SMD = −0.84, 95% CI = −2.82 to
−1.14, p = 0.41, p for heterogeneity = 0.95, I2 = 0%)
(Figure 5).
Safety. Three patients complained about mental
stress, dizziness, palpitation or fatigue, but none of
them had adverse reactions in the treatment (Yang
Zhang et al. 967

Figure 3.  Forest plot of trials comparing the effects of therapeutic ultrasound and control on knee pain changes in
a visual analogue scale (VAS). (A) The aggregated results of VAS changes with combined two therapeutic ultrasound
mode (six trials). (B) Aggregated results of VAS changes with continuous mode (five trials). (C) The aggregated
results of VAS changes with pulsed mode (three trials).

Figure 4.  Forest plot of trials comparing the effects of therapeutic ultrasound and control on knee pain (visual
analogue scale) in different duration. (A) SMDs changes in eight weeks (three trials). (B) SMDs changes in less than
four weeks (three trials).
968 Clinical Rehabilitation 30(10)

Figure 5.  Forest plot of trials comparing the effects of therapeutic ultrasound and control on physical function
changes. (A) SMDs at the end of the intervention (WOMAC physical function score). (B) SMDs at the end of the
intervention (Lequesne efficacy indexes). (C) SMDs at the end of the intervention (50 meters walking time).

et al.12). No systemic or local side-effects were
reported during or after the treatment in the trial
(Tascioglu et al.23). No adverse events were reported
at any time during the application of the interven-
tions period in two trials either.21,26 Two patients
dropped out owing to the increased pain in the trial
(Özgönenel et al.22), but none withdrew because of
adverse events. There were no adverse events dis-
covered in the rest of the included trials. Safety
results showed that therapeutic ultrasound could be
a safe treatment, and the relative risks could not be
calculated.
Discussion
This systematic review suggested that therapeutic
ultrasound in favor of pain relief may improve the
physical function of patients with knee osteoarthritis.
No adverse events were found in the therapeutic
ultrasound treatment of any trial, proving it a safe
treatment.
The main results of this article supported the
main findings of the previous two reviews.5,26 In
contrast to the previous two systematic review, four
more randomized trials were supplemented to sup-
port our conclusion. The two systematic reviews
used the recognized Cochrane library assessment to
assess the risk of bias. PEDro was a professional
risk of bias assessment tool for physical therapy.
Because this scale consists of 10 ratings, which can
provide detail assessment information, we adopted
it in our systematic review.
The important information about how long the
therapeutic ultrasound could relieve pain for knee
osteoarthritis patients was not provided in the two
reviews. We considered that this is important for
Zhang et al. 969
physical therapist in clinical practice, and hence we
pooled the results of different therapeutic ultra-
sound duration. The results showed that both four-
week and eight-week ultrasound treatments can
relieve pain. What is more, the subgroup analysis
found that there is no subgroup differences between
the two different treatment durations. Therefore,
physical therapists could choose either durations.
There are two modes of therapeutic ultrasound, but
figuring out which one is better for patients with
osteoarthritis is important for physical therapists.
We found that both continuous and pulsed ultra-
sound could relieve pain, which was not reported
by the previous systematic review.
Potential biases in the review process should be
explained. We combined the results of trials that
had a continuous ultrasound group and pulsed
ultrasound group in the same trial. All the formula
for the calculation of changes in means and devia-
tions followed the Cochrane handbook for system-
atic reviews of interventions (version 5.3.0). In
pain assessment, eight trials used a visual analogue
scale measurement consisting of a 10-centimeter
horizontal line, whereas one trail24 utilized millim-
eter as the measurement unit for horizontal line of
pain, which will cause the bias in pooled analysis.
We excluded this trial24 in the pooled analysis,
because there would be a high degree of heteroge-
neity if it was brought into the analysis. The timing
of pain assessment involved rest pain and move-
ment pain. Only two trials provided the detailed
information to confirm the timing, and the first one
was excluded for measurement reason.24. The sec-
ond one had both rest and movement pain measure-
ment,21 and the change of the latter was more
obvious than the former. In order to get a more con-
servative result, we applied the rest timing data. In
addition, as two trials reported the assessment
results by numbers of knees rather than by patients,
we included the number of patients in our analysis,
which was the same method of the former system-
atic review.21
In the physical function assessment, four trials
applied the WOMAC physical function subscale.
We excluded one of the four trials that reported a
mean ± standard deviation (30.6 ±511.4) in the pla-
cebo group,22 because the sensitivity analysis
found that it would change the result of therapeutic
effects. Four trials used Lequesne Indexes, but the
pooled analysis result showed a high heterogeneity,
which maybe came from a diverse therapeutic
ultrasound mode and treatment duration. The
measurement of 50  meters walking time was
recorded only by three trials, and the pooled analy-
sis result did not support therapeutic ultrasound.
We suggested that more trials are needed to test
whether therapeutic ultrasound can change the
time that patients with osteoarthritis require for
walking a distance of 50 meters. The evidence
quality of our systematic review is limited by the
quality and low number of trials. Only one trial
completed an adequate sequence and the group
allocation,26 whereas the rest all described insuffi-
cient randomization as well as allocation proce-
dure. We found that it is difficult to blind the patient
in clinical trials. Despite that, some blinding meas-
ures were utilized, such as pressing the button or
using a non-output probe; the patient can discover
the treatment differences easily when the machine
audible beep and thermal effects are absent. In
order to maintain patient blinding, Falconer et al.25
added a more detailed procedure in. Such a proce-
dure involved that the physical therapist gave
scripted instructions to all patients, which stated
that they may not feel any warmth and patients
were not asked directly whether they felt it.25 For
other aspects of evidence quality, only three trials
reporting more than six months follow-up,17,18,24
and only one study used intention-to-treat analysis
in clinical outcomes analysis.26
There were three limitations in our systematic
review. First, it was only dedicated to compare
therapeutic ultrasound with sham ultrasound or no
intervention. However, it is important for physical
therapist to know whether therapeutic ultrasound
is more effective than any other physical therapy.
Therefore, we considered that a net meta-analysis
may be needed. Second, hot packs, passive range
of motion exercise, quadriceps isometric exercise
and stretching exercises of the lower extremity
muscles were the concurrent treatment in six trials,
which may enhance biological effects of therapeu-
tic ultrasound. We proposed a hypothesis that if all
the groups received the same concurrent therapy,
970 Clinical Rehabilitation 30(10)
the effects of ultrasound treatment could be tested.
Nevertheless, to pool the results of trials that com-
bined those concurrent therapies may influence
the final analysis results. Third, two reports of
Huang MH that described the same trial were
included in our systematic review. Therefore, this
analysis method may increase the total sample
size.
Clinical messages
•• Therapeutic ultrasound is beneficial for
reducing pain and improving physical
function of knee osteoarthritis.
•• Both the continuous and pulsed modes
may relieve pain for patients with
osteoarthritis.
•• Treatment durations should not be lim-
ited to four or eight weeks, but ought to
be tailored to the specific needs of indi-
vidual patients.
Acknowledgements
Authors are grateful to the National Science Foundation of
China for providing funds for this research. C Zhang and
XT Luo would like to express sincere appreciation to the
Chinese Cochran Center for systematic review training.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest
with respect to the research, authorship, and/or publica-
tion of this article.
Contributors
C Zhang, YJ Xie and XT Luo were responsible for the
literature search and data extraction. CL Lu and QD Ji
were responsible for the blinding literature search and
the data extraction. P Wang was in charge of data inter-
pretation and technical guidance. CQ He made important
intellectual contribution in research design and manu-
script revision.
Funding
The author(s) disclosed receipt of the following financial
support for the research, authorship, and/or publication
of this article: This research was supported by National
Natural Science Foundation of China [Grant Number
81372110]; and National Natural Science Foundation of
China for Young Scholar [Grant Number 81401858].
Reference
1. Zhang W, Nuki G, Moskowitz RW, et al. OARSI recom-
mendations for the management of hip and knee osteoar-
thritis: part III: Changes in evidence following systematic
cumulative update of research published through January
2009. Osteoarthritis Cartilage 2010; 18: 476–499.
2. Weinstein SL, Jacobs JJ and Goldberg MJ. Osteoarthritis
of the knee. New Engl J Med 2006; 354: 2508–2509;
author reply 2509.
3. Kotlarz H, Gunnarsson CL, Fang H and Rizzo JA. Insurer
and out-of-pocket costs of osteoarthritis in the US:
Evidence from national survey data. Arthritis Rheumat
2009; 60: 3546–3553.
4. Paliwal S and Mitragotri S. Therapeutic opportunities in
biological responses of ultrasound. Ultrasonics 2008; 48:
271–278.
5. Rutjes AW, Nuesch E, Sterchi R and Juni P. Therapeutic
ultrasound for osteoarthritis of the knee or hip. Cochrane
Database Syst Rev 2010: CD003132.
6. McAlindon TE, Bannuru RR, Sullivan MC, et al. OARSI
guidelines for the non-surgical management of knee oste-
oarthritis. Osteoarthritis Cartilage 2014; 22: 363–388.
7. Altman R, Asch E, Bloch D, et al. Development of cri-
teria for the classification and reporting of osteoarthritis.
Classification of osteoarthritis of the knee. Diagnostic
and Therapeutic Criteria Committee of the American
Rheumatism Association. Arthritis Rheumat 1986; 29:
1039–1049.
8. Green S. Cochrane handbook for systematic reviews of
interventions version 5.1.0. The Cochrane Collaboration
2011.
9. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J
and Stitt LW. Validation study of WOMAC: A health sta-
tus instrument for measuring clinically important patient
relevant outcomes to antirheumatic drug therapy in
patients with osteoarthritis of the hip or knee. J Rheumatol
1988; 15: 1833–1840.
10. Lequesne MG, Mery C, Samson M and Gerard P.
Indexes of severity for osteoarthritis of the hip and knee.
Validation—value in comparison with other assessment
tests. Scand J Rheumatol Supplement 1987; 65: 85–89.
11. Maher CG, Sherrington C, Herbert RD, Moseley AM and
Elkins M. Reliability of the PEDro scale for rating quality of
randomized controlled trials. Phys Ther 2003; 83: 713–721.
12. Yang PF, Li D, Zhang SM, et al. Efficacy of ultrasound
in the treatment of osteoarthritis of the knee. Orthop Surg
2011; 3: 181–187.
13. Boyaci A, Tutoglu A, Boyaci N, Aridici R and Koca I.
Comparison of the efficacy of ketoprofen phonophoresis,
ultrasound, and short-wave diathermy in knee osteoarthri-
tis. Rheumatol Int 2013; 33: 2811–2818.
14. Mascarin NC, Vancini RL, Andrade ML, Magalhaes
Ede P, de Lira CA and Coimbra IB. Effects of kinesio-
therapy, ultrasound and electrotherapy in management
Zhang et al. 971
of bilateral knee osteoarthritis: prospective clinical trial.
BMC Musculoskel Disord 2012; 13: 182.
15. Luksurapan W and Boonhong J. Effects of phonophoresis
of piroxicam and ultrasound on symptomatic knee osteo-
arthritis. Arch Phys Med Rehabil 2013; 94: 250–255.
16. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA state-
ment for reporting systematic reviews and meta-analyses of
studies that evaluate health care interventions: Explanation
and elaboration. J Clin Epidemiol 2009; 62: e1–34.
17. Huang MH, Yang RC, Lee CL, Chen TW and Wang MC.
Preliminary results of integrated therapy for patients with
knee osteoarthritis. Arthritis Rheumat 2005; 53: 812–820.
18. Loyola-Sanchez A, Richardson J, Beattie KA, Otero-
Fuentes C, Adachi JD and MacIntyre NJ. Effect of low-
intensity pulsed ultrasound on the cartilage repair in
people with mild to moderate knee osteoarthritis: A dou-
ble-blinded, randomized, placebo-controlled pilot study.
Arch Phys Med Rehabil 2012; 93: 35–42.
19. Cetin N, Aytar A, Atalay A and Akman MN. Comparing
hot pack, short-wave diathermy, ultrasound, and TENS on
isokinetic strength, pain, and functional status of women
with osteoarthritic knees: A single-blind, randomized, con-
trolled trial. Am J PhysMed Rehabil 2008; 87: 443–451.
20. Ulus Y, Tander B, Akyol Y, et al. Therapeutic ultrasound
versus sham ultrasound for the management of patients
with knee osteoarthritis: A randomized double-blind con-
trolled clinical study. Int J Rheumat Dis 2012; 15: 197–206.
21. Özgönenel L, Aytekin E and Durmusoglu G. A double-
blind trial of clinical effects of therapeutic ultrasound
in knee osteoarthritis. Ultrasound Med Biol 2009; 35:
44–49.
22. Huang MH, Lin YS, Lee CL and Yang RC. Use of ultra-
sound to increase effectiveness of isokinetic exercise for
knee osteoarthritis. Arch Phys Med Rehabil 2005; 86:
1545–1551.
23. Tascioglu F, Kuzgun S, Armagan O and Ogutler G. Short-
term effectiveness of ultrasound therapy in knee osteoar-
thritis. J Int Med Res 2010; 38: 1233–1242.
24. Cakir S, Hepguler S, Ozturk C, Korkmaz M, Isleten B
and Atamaz FC. Efficacy of therapeutic ultrasound for the
management of knee osteoarthritis: A randomized, con-
trolled, and double-blind study. Am J PhysMed Rehabil
2014; 93: 405–412.
25. Falconer J, Hayes KW and Chang RW. Effect of ultra-
sound on mobility in osteoarthritis of the knee. A rand-
omized clinical trial. Arthritis Care Res 1992; 5: 29–35.
26. Loyola-Sanchez A, Richardson J and MacIntyre NJ.
Efficacy of ultrasound therapy for the management of
knee osteoarthritis: A systematic review with meta-anal-
ysis. Osteoarthritis Cartilage 2010; 18: 1117–1126.