Summary
Objective
To assess the efficacy 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% confidence 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 efficacious for decreasing pain and may improve physical function in patients with knee OA. The findings of this review should be confirmed using methodologically rigorous and adequately powered clinical trials.
Keywords
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 worldwide3
. Physiotherapy is recommended for the management of painful knee OA4
and ultrasound therapy (US) is one of the most common physical agents used within physiotherapy practice in several countries5
. US is based on the application of high frequency sound waves to the tissues of the body in order to obtain mechanical or thermal effects6
. These effects aim to enhance soft tissue healing, decrease the inflammatory response, increase blood flow, increase metabolic activity, and decrease pain6
. Moreover, there is some evidence that ultrasonic energy stimulates the repair of joint cartilage in animal models of cartilage injury7
, 8
, 9
. Therefore, US could be an effective intervention 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 recommendations for the management of knee and hip OA
4
. Ultrasound was not identified as a core treatment modality based on the results of a systematic review published in 2001 by the Cochrane Collaboration10
. The Cochrane systematic review included studies available prior to June 200010
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 treatment options for people with knee OA, it is important to confirm 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 efficacy 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 efficacy of US on patient-perception of disease severity, and cartilage repair. We followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) recommendations11
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 citations (February 5, 2009), EMBASE (1980–2009 week 6), LILACS (February 6, 2009), MEDCARIB (February 6, 2009), CINAHL (February 8, 2009), pre CINAHL (February 8, 2009), PEDro (last updated February 2, 2009), AMED (1985 to January 2009), SPORTdiscus (1830 to February 9, 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 Dissertations and Theses. Authors of an unpublished study reported as a conference proceeding in the Proceedings First database were contacted via e-mail. A detailed example of the full electronic search strategy for Ovid MEDLINE is provided in Appendix A. Briefly, 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. Duplicates 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 from the review if phonophoresis was the only ultrasonic intervention, US was combined with another intervention not provided to the comparison group, and samples included subjects having other diagnoses and results for the subjects with knee OA were not reported separately. Cohen’s unweighted Kappa (κ) was used to measure agreement between reviewers
12
. Observed agreement was fair (κ=0.57)23
. Disagreement was solved by consensus including a third reviewer (NM). Colleagues translated non-English articles (n=9) written in their first language (a physical therapist, an engineer, an occupational therapist, a rheumatologist, and a physiatrist).Data extraction and management
Two reviewers, using a pre-tested data collection form, followed 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, intensity and dose), co-interventions, outcomes (pain, physical function, participant’s 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 defined cartilage repair as all measures that directly or indirectly assess the cartilage formation–degradation process (i.e., imaging, arthroscopy, fluid 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 therapist 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 identified 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 (κ=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 findings, 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 calculate 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 between groups and the standard deviation (SD) of the outcomes among participants (Hedges’ adjusted g). The 95% confidence interval (CI) was calculated and a Z test was performed with significance set at P<0.05. Heterogeneity among studies was assessed with a χ2 test with significance set at P<0.10 and an inconsistency test (I2) which represents the percentage variability in the effect estimates that is due to heterogeneity rather than chance. Inconsistency (I2)>40% was considered significantly high16
.Two studies
where X=mean for each group, SD=standard deviation, s2=variance, and n=number of participants in that group.
17
, 18
reported different subgroups of US, so the means and SDs of these subgroups were pooled using the following formulas19
:where X=mean for each group, SD=standard deviation, s2=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 measurements taken from the knees of the same patient.In order to express the effect estimates as percentage change relative to the control groups, the SMDs were back transformed to mean differences (MD) using SDs reported in observational studies
21
, - Villanueva I.
- del Mar Guzman M.
- Javier Toyos F.
- Ariza-Ariza R.
- Navarro F.
Relative efficiency 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: 225-231
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 identified, fulfilled the inclusion criteria and were retrieved for full text review (see Fig. 1). Of these, only six studies were included and considered in the final analysis. The remaining 17 studies were excluded because the samples included persons with diagnoses other than knee OA
24
, , 26
; compared US with short wave diathermy27
, 28
, 29
, diadynamic currents/magnetotherapy30
, or phonophoresis31
, without a control group with similar exposure; compared different frequencies/modes of US, 33
, 34
, or used US in combination with other physical modalities35
without a control group; four articles were case series reports36
, , 38
, 39
; and one study was performed in vitro40
.
Fig. 1Flowchart of study selection.
Description of included studies
The six small, English-language, randomized controlled trials
17
, 18
, 20
, 41
, 42
, 43
included in this review are summarized in Table I. In one study18
, unpublished data were provided by the primary author. All trials included people with knee OA who met the American College of Rheumatology diagnostic criteria44
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.- Altman R.
- Asch E.
- Bloch D.
- Bole G.
- Borenstein D.
- Brandt K.
- et al.
Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and therapeutic criteria committee of the american rheumatism association.
Arthritis Rheum. 1986; 29: 1039-1049
42
, tabulated results were extracted. Five studies reported sufficient information for the dose calculation17
, 18
, 20
, 42
, 43
. Three trials17
, 18
, 20
conducted by the same group reported a peak intensity 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 modified for each US application and the calculated dose would be inaccurate. Communication with the primary author confirmed that the intensity output was fixed and only the speed of the sound head varied during the US application. Thus, US dose could be calculated.Table ISummary 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; ∼58 (8) years. | n=18 Sonopulus 590 US device, continuous mode, 1.5 W/cm2 intensity, 180 J/cm2 dose, 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-VAS Physical function-LSI Walking speed-time to walk 50 m in seconds) | 8 Weeks | High |
| Falconer et al. 42 (USA) | ∼72% Women with severity not specified; ∼67.5 (11) years. | n=34 Chattanooga Intellect 200 US device, continuous mode, 1.7 W/cm2 intensity, 26 J/cm2 dose, 12 twelve-min sessions in 4–6 weeks. | n=35 Sham US (start button not pushed). | Pain-VAS Walking speed-50 ft walk time (converted to m/min) | 4–6 Weeks | High |
| Huang et al. 18 (Taiwan) | ∼33% Women with mild to severe bilateral knee OA; ∼60 (9) years. | n=30 Sonopulus 590 US device; pulsed mode (duty cycle: 25%), 2.5 W/cm2 intensity, 112 J/cm2 dose, 24 fifteen-min sessions in 8 weeks. | n=30 Sham US. | Pain-VAS Physical function-LSI Walking speed-50 ft walk time (converted to m/min) Arthritis Severity Index=ratio of 99mTechnetium uptake knee/99mTechnetium uptake middle third ipsilateral femur | 8 Weeks | Unclear |
| Huang et al. 17 (Taiwan) | ∼81% Women with mild knee OA; ∼62 (17) years. | Group 1: n=27; Continuous mode, 1.5 W/cm2 intensity, 270 J/cm2 dose. Group 2: n=30; Pulsed mode (duty cycle: 25%), 2.5 W/cm2 intensity, 112.5 J/cm2 dose. Both groups used a Sonopulus 590 US device, 24 fifteen-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-VAS Physical function-LSI Walking speed-50 ft walk time (converted to m/min) | 8 Weeks and 12 months | High |
| Huang et al. 20 (Taiwan) | ∼81% Female with mild knee OA, ∼65 (13) years. | n=32 Sonopulus 590 US device, pulsed mode (duty cycle: 25%), 2.5 W/cm2 intensity, 112 J/cm2 dose, 24 fifteen-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-VAS Physical function-LSI Walking speed-50 ft walk time (converted to m/min) | 8 Weeks and 12 months | High |
| Ozgonenel et al. 43 (Turkey) | ∼80% Women with mild to moderate knee OA; ∼55 (7.5) years. | n=34 Peterson .250 US device, continuous mode, 1 W/cm2 intensity, 150.72 J/cm2 dose, 10 five-min sessions in 2 weeks. | n=31 Sham US (applicator disconnected from device). | Pain-VAS Physical function-WOMAC Walking speed-time to walk 50 m (seconds) | 2 Weeks | High |
∗ Sample size of subjects whose data were included in the present meta−analysis.
† Treated surface area was not reported and US dose was calculated using a value of 25 cm2 reported in 1 similar trial
17
.‡ Visual Analogue Scale (VAS) from 0 cm (no pain) to 10 cm (most intense pain).
§ LSI scale from 0 (better) to 26 (worst).
‖ 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/cm2 to maximal tolerable dose not exceeding 2.5 W/cm2 ”).
# 99mTechnetium uptake was measured by bone scan (TOSHIBA GCA-90 γ-camera) 3 h after administrating the radioisotope.
∗∗ Same number of sessions and period of time as in the intervention group.
One trial
42
did not report the mode of US, however the author confirmed 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 cm2) reported in a trial that used the same US device17
. Only two trials17
, 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 five studies17
, 20
, 41
, 42
, 43
(Table II). Thus, the evidence included in this review has a high risk of bias overall.Table IIRisk 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-final 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 |
The quality of evidence is low for pain and physical function outcomes because of the high risk of bias (Table II), and the heterogeneity observed in results across trials [I2=51–92%, Fig. 2, Fig. 3, Fig. 4] (Appendix B). The quality of evidence is considered low for cartilage repair because the only trial
18
that reported this outcome had an unclear risk of bias (Table II) and used a surrogate measure (99mTechnetium uptake) to assess cartilage status (Appendix B).
Fig. 2Meta-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).
Fig. 3Meta-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. 4Meta-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).
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 (χ2=10.26, P=0.07, I2=51%), therefore predefined subgroup analyses were conducted. Ultrasound mode, intensity, and therapeutic dose completely explained the inconsistency between the groups [Fig. 2(B)]. In all subgroups, effect estimates favored US therapy; however, the differences were statistically significant only in the low intensity/pulsed US and US dose<150 J/cm2 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 score43
(Table I). The effect estimate observed favored the US intervention, however the difference between groups was not statistically significant [SMD (CI)=−0.54 (−1.19, 0.12), P=0.11] [Fig. 3(A)]. Heterogeneity was high (χ2=29.11, P<0.001, I2=86%), and predefined subgroup analyses were performed. Inconsistency was not explained satisfactorily by any of the subgroup analyses performed. Data pooled from two trials17
, 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 significant 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 (χ2=52.2, P<0.001, I2=92%), and the subgroup analyses did not explain this inconsistency. Pooling the results of two studies17
, 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 identified 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 99mTechnetium uptake on bone scans in order to determine an ‘index of arthritis severity’, 99mTechnetium uptake in the knee divided by the 99mTechnetium uptake in the middle third of the ipsilateral femur. The authors validated this outcome measure on animal models of cartilage injury45
. The results of this study showed a significant decrease in the ‘index of arthritis severity’ measured after an 8 week US intervention for patients in the lowest and middle 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 study43
reported that no adverse events occurred either during or after the interventions. Since the number of adverse events reported in these two trials was 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 efficacy 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 finding 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 sample47
. Using this approach, the NNT is 7. It also seems that US applied using low intensity (<1 W/cm2), pulsed mode, and a therapeutic dose<150 J/cm2 could be more effective at reducing pain than US applied using high intensity (≥1 W/cm2), continuous mode and a therapeutic dose>150 J/cm2. In general, a non-significant 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 findings from two trials by one research group17
, 20
, beneficial effects may last for 10 months after the US treatment is completed [Fig. 2, Fig. 3, Fig. 4].Limited evidence found in this review prevents conclusive statements regarding dose effects and the effectiveness of US on cartilage repair. Our findings which suggest a threshold US dosage for pain reduction in persons with knee OA are consistent with previous findings that higher US doses are less effective for tissue repair in humans
48
and even harmful for the growth plate of rabbits49
. 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 findings together with animal studies which permit direct measurement of cartilage tissue response to US7
, 8
, 9
are consistent with the “mechanotransduction theory”. This theory proposes that mechanical stimuli increase the chondrocyte production of proteoglycans and anti-inflammatory proteins50
. Further studies are warranted to determine if low dose pulsed US is optimal for pain relief and stimulation of cartilage repair.Our findings 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 fibers which attenuates nociception5
. However, our findings suggest a non-thermal mechanism for pain reduction which may be related to a threshold dose 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 performance outcomes. Heterogeneity was reduced for self-reported physical function (I2 reduced from 86% to 30%) and walking performance (I2 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 findings 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 confirmed that it was continuous. Our search yielded one unpublished trial18
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 trials17
, 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 findings are based on evidence which has a high risk of bias and low quality. When we synthesized only the trials that adequately blinded the participants
42
, 43
the effect size decreased considerably and was no longer statistically significant [SMD (CI)=−0.24 (−0.63, 0.14)]. This suggests that the effect sizes for pain found in this review could be partly inflated 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 beneficial 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 identified, 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 limitations described reduce the confidence in the effect estimates observed in the present meta-analyses.Implications for practice
US (10–24 sessions) appears to be efficacious 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 influence 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 definitive trials are needed.
Implications for research
Trials that are methodologically rigorous and adequately powered are needed to confirm 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 benefit 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 conceptualization, design, data collection, analysis, interpretation, drafting and revisions; and approved the final version.
Conflict of interest
None of the authors has any financial and personal relationships with other people or organizations that could potentially and inappropriately influence 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)
Tabled
1
| 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 |
Appendix B. Summary of findings and quality of evidence assessment
Tabled
1
| Outcomes | Control group mean | Ultrasound group mean with respect to the control group (CI) | Relative change (CI) | Quality of the evidence, † GRADE Working Group grades of evidence15. High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. | Comments |
|---|---|---|---|---|---|
| Pain VAS from 0 (no pain) to 10 (intense pain) Follow up: 2–8 weeks | 3.95 | 0.84 Lower (1.2–0.48 lower) | 21% Lower (30%, 12%) | Low | High risk of bias of the included studies and considerable results’ heterogeneity. |
| Follow-up: 12 months | 3.95 | 1.22 Lower (1.62–0.82 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: 2–8 weeks | 5.65 | 1.11 Lower (2.45 lower to 0.24 higher) Mild severity (grade II) subgroup=2.7 lower (4.03–1.38 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 months LSI | 5.65 | 2.07 Lower (2.91–1.22 lower) | 36% Lower (51%, 21%) | Low | |
| Walking performance Walking speed (m/min) Follow-up: 2–8 weeks | 82.47 | 4.64 Higher (0.5 lower to 9.85 higher) Mild severity (grade II) subgroup=11.17 higher (14.32–8.07 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 months Walking speed (m/min) | 82.47 | 10.89 Higher (7.07–14.71 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 (2.47–1.13 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.
† GRADE Working Group grades of evidence
15
. High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.‡ The final score mean for the control groups was calculated by pooling the means and standard errors through a generic inverse-variance method.
§ 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.
‖ .
¶ 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.- Villanueva I.
- del Mar Guzman M.
- Javier Toyos F.
- Ariza-Ariza R.
- Navarro F.
Relative efficiency 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: 225-231
22
[walking speed 5.73].References
- Epidemiology and cost of conservative treatment of patients with degenerative joint disease of the hip and knee.Ortop Traumatol Rehabil. 2007; 9: 405-412
- Prevalence of knee and hip osteoarthritis and the appropriateness of joint replacement in an older population.Arch Intern Med. 2008; 168: 1576-1584
- Epidemiology of osteoarthritis.Rheum Dis Clin North Am. 2008; 34: 515-529
- OARSI recommendations for the management of hip and knee osteoarthritis, part II: OARSI evidence-based, expert consensus guidelines.Osteoarthritis Cartilage. 2008; 16: 137-162
- Ultrasound.in: Evidence-based Guide to Therapeutic Physical Agents. Lippincott Williams & Wilkins, Philadelphia PA, USA2003: 223-261
- A review of therapeutic ultrasound: biophysical effects.Phys Ther. 2001; 81: 1351-1358
- The effect of low-intensity pulsed ultrasound on autologous osteochondral plugs in a canine model.Am J Sports Med. 2008; 36: 1733-1741
- Effects of sonication on articular cartilage in experimental osteoarthritis.J Rheumatol. 1997; 24: 1978-1984
- Gross and histomorphological effects of therapeutic ultrasound (1 watt/cm2) in experimental acute traumatic arthritis in donkeys.J Equine Vet Sci. 1997; 17: 150-155
- Therapeutic ultrasound for osteoarthritis of the knee.Cochrane Database Syst Rev. 2001; : 003132
- 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: e1-e34
- A coefficient of agreement for nominal scales.Educ Psychol Meas. 1960; 20: 37-46
- Chapter 8: assessing risk of bias in included studies.in: Higgins J.P.T. Green S. Cochrane Handbook for Systematic Reviews of Interventions. 2008 (Version 5.0.1 (updated September 2008). 5.0.1 ed) (The Cochrane Collaboration, 2008)
- GRADE: an emerging consensus on rating quality of evidence and strength of recommendations.BMJ. 2008; 336 (7650:ate of Pubaton: 26 Ar 2008)
- Meta-analysis in clinical trials.Control Clin Trials. 1986; 7: 177-188
- Measuring inconsistency in meta-analyses.BMJ. 2003; 327: 557-560
- Use of ultrasound to increase effectiveness of isokinetic exercise for knee osteoarthritis.Arch Phys Med Rehabil. 2005; 86: 1545
- Peek W.J. Lankhorst G.J. Effects of Pulse Sonication on Functional Status of Patients with Knee Osteoarthritis. International Society of Physical and Rehabilitation Medicine. Monduzzi, Amsterdam, The Netherlands2001 Jul: 297-300
- Introduction to Probability and Statistics.10th edn. Duxbury Press, Pacific Grove, Calif.1999
- Preliminary results of integrated therapy for patients with knee osteoarthritis.Arthritis Rheum. 2005; 53: 812-820
- Relative efficiency 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: 225-231
- A comparison of various therapeutic exercises on the functional status of patients with knee osteoarthritis.Semin Arthritis Rheum. 2003; 32: 398-406
- Evaluating coding decisions.in: Cooper H. Hedges L.V. Russell Sage Foundation, New York (NY)1994
- Treatment of arthrosis deformans by simultaneous application of interferential current and ultrasonic waves.J Egypt Med Assoc. 1975; 58: 328-333
- Revmatologiia (Mosk). 1988; : 37-39
- Does physiotherapy affect the pain in activated arthrosis?.Z Physiother. 1988; 40: 333-336
- Effectiveness of shortwave diathermy and ultrasound in the treatment of osteo-arthritis of the knee joint.Med J Zambia. 1975; 9: 138-139
- The effects of physiotherapy on osteoarthritic knees of females.J Formos Med Assoc. 1991; 90: 1008-1013
- Comparison of ultrasound and short wave diathermy therapy in knee osteoarthritis.Turk J Phys Med Rehab. 2006; 52: 168-173
- Physical agents for pain management in patients with gonarthrosis.Ortop Traumatol Rehabil. 2005; 7: 317-321
- Short term efficacy of ibuprofen phonophoresis versus continuous ultrasound therapy in knee osteoarthritis.Swiss Med Wkly. 2003; 133: 333-338
- Vopr Kurortol Fizioter Lech Fiz Kult. 1991; : 36-39
- Results of frequency differences in ultrasonic therapy.Phys Ther. 1970; 50: 481-486
- Clinical comparison of the effect of ultrasound using direct and alternating current in the treatment of gonarthrosis.Z Physiother. 1981; 33: 159-163
- Vopr Kurortol Fizioter Lech Fiz Kult. 1996; : 18-21
- Z Orthop Ihre Grenzgeb. 1981; 119: 126-137
- Radioter Radiobiol Fis Medica. 1954; 9: 182-190
- Ultrasonics in osteoarthritis.J Med Assoc State Ala. 1953; 22: 182-184
- Treatment of musculoskeletal disorders with ultrasound.J Occup Med. 1965; 7: 434-438
- Bersani F. Mechanical Stimulation of Cartilage by Ultrasound. Electricity and Magnetism in Biology and Medicine. Kluwer, Bologna, Italy1997 Jun: 947-950
- 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, controlled trial.Am J Phys Med Rehabil. 2008; 87: 443-451
- Effect of ultrasound on mobility in osteoarthritis of the knee. A randomized clinical trial.Arthritis Care Res. 1992; 5: 29-35
- A double-blind trial of clinical effects of therapeutic ultrasound in knee osteoarthritis.Ultrasound Med Biol. 2009; 35: 44-49
- Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and therapeutic criteria committee of the american rheumatism association.Arthritis Rheum. 1986; 29: 1039-1049
- The early evaluation of induced osteoarthritis in rats with 99Tcm-pertechnetate scans.Nucl Med Commun. 1996; 17: 529-535
- A simple method for converting an odds ratio to effect size for use in meta-analysis.Stat Med. 2000; 19: 3127-3131
- Factors associated with change in pain and disability over time: a community-based prospective observational study of hip and knee osteoarthritis.Br J Gen Pract. 2005; 55: 205-211
- Incisional wound healing: a controlled study of low and high dose ultrasound.J Orthop Sports Phys Ther. 1993; 18: 619-628
- The effects of therapeutic vs. high-intensity ultrasound on the rabbit growth plate.J Orthop Res. 2003; 21: 865-871
- Mechanotransduction pathways of low-intensity ultrasound in C-28/I2 human chondrocyte cell line.Proc Inst Mech Eng [H]. 2007; 221: 527-535
- Therapeutic ultrasound for osteoarthritis of the knee or hip.Cochrane Database Syst Rev. 2010; : CD003132https://doi.org/10.1002/14651858.CD003132.pub2
- Exercise for osteoarthritis of the knee.Cochrane Database Syst Rev. 2008; : 4
Article Info
Publication History
Published online: July 15, 2010
Accepted:
June 17,
2010
Received:
January 5,
2010
Identification
Copyright
© 2010 Osteoarthritis Research Society International. Published by Elsevier Inc.
User License
Elsevier user license | How you can reuse 
Elsevier's open access license policy
Elsevier user license
Permitted
For non-commercial purposes:
- Read, print & download
- Text & data mine
- Translate the article
Not Permitted
- Reuse portions or extracts from the article in other works
- Redistribute or republish the final article
- Sell or re-use for commercial purposes
Elsevier's open access license policy
00212-8&id=gr1.gif){kind=link}
00212-8&id=gr2.jpg){kind=link}
00212-8&id=gr3.jpg){kind=link}
00212-8&id=gr4.jpg){kind=link}