If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Center for MicroElectroMechanical Systems (CMEMS‑UMINHO), University of Minho, Azurém Campus, 4800‑058 Guimarães, PortugalLABBELS – Associate Laboratory, Braga, Guimarães, Portugal
Clínica Espregueira - FIFA Medical Centre of Excellence, Porto, PortugalDom Henrique Research Centre, Porto, PortugalPorto Biomechanics Laboratory (LABIOMEP), Faculty of Sports, University of Porto, Porto, Portugal
Clínica Espregueira - FIFA Medical Centre of Excellence, Porto, PortugalDom Henrique Research Centre, Porto, PortugalICVS/3B's-PT Government Associate Laboratory, Braga/Guimarães, Portugal3B’s Research Group‑Biomaterials, Biodegradables and Biomimetics, Headquarters of the European Institute of Excellence on Tissue Engineering and Regenerative Medicine, University of Minho, AvePark, Parque de Ciência e Tecnologia, Zona Industrial da Gandra, Barco, 4805‑017 Guimarães, PortugalSchool of Medicine, University of Minho, Braga, Portugal
Center for MicroElectroMechanical Systems (CMEMS‑UMINHO), University of Minho, Azurém Campus, 4800‑058 Guimarães, PortugalLABBELS – Associate Laboratory, Braga, Guimarães, Portugal
Center for MicroElectroMechanical Systems (CMEMS‑UMINHO), University of Minho, Azurém Campus, 4800‑058 Guimarães, PortugalLABBELS – Associate Laboratory, Braga, Guimarães, Portugal
Center for MicroElectroMechanical Systems (CMEMS‑UMINHO), University of Minho, Azurém Campus, 4800‑058 Guimarães, PortugalLABBELS – Associate Laboratory, Braga, Guimarães, PortugalDom Henrique Research Centre, Porto, Portugal
Meta-analyses presented very-low certainty of evidence.
•
All studies were judged to have risk of bias in one or more domains.
•
15 of 42 (38%) pooled comparisons were superior in mechanical-based therapies.
•
These therapies may be cautiously applied in some patients with knee OA.
Abstract
Background
Mechanical-based therapies are not yet recommended to manage osteoarthritis (OA). This systematic review and meta-analysis aim to assess the effects of passive mechanical-based therapies (isolated or combined with other therapies) on patients with knee OA compared to placebo, other isolated or combined interventions.
Methods
Pubmed, Cochrane, Web of Science and EMBASE were searched up to December 2020. We included randomized and non-randomized trials using therapeutic ultrasound, phonophoresis, extracorporeal shockwave therapy (ESWT) and vibration (single or combined with other therapies) compared to placebo, and/or other physical therapies groups. Biochemical, patient-reported, physical and imaging outcome measures were retrieved. We judged risk of bias using the RoB2 tool for randomized studies, the ROBINS-I tool for non-randomized studies, and the GRADE to interpret certainty of results.
Results
We included 77 clinical studies. Ultrasound and ESWT statistically improved pain and disability comparing to placebo (combined or not with other therapies), and when added to other therapies versus other therapies alone. Ultrasound was statistically inferior to phonophoresis (combined or not with other therapies) in reducing pain and disability for specific therapeutic gels and/or combined therapies. Vibration plus exercise statistically improved pain relief and function versus exercise alone. All meta-analyses showed very-low certainty of evidence, with 15 of 42 (38%) pooled comparisons being statistically significant (weak to large effect).
Conclusions
Despite the inconsistent evidence with very-low certainty, the potential benefits of passive mechanical-based therapies should not be disregard and cautiously recommended that clinicians might use them in some patients with knee OA.
Knee osteoarthritis (OA) is a highly prevalent joint degenerative disease and one of the main causes of disability; with a global prevalence of 22.9% and affecting approximately 654 million individuals [
]. OA has a significant clinical and economic impact worldwide, involving a considerable amount of healthcare resources including costs from pharmacological treatments, rehabilitation and time off-work [
Non-operative treatment is the first-line of treatment to treat knee OA, and mechanical-based treatments (exercises-based physiotherapy and weight loss) are commonly used for pain and joint function improvement [
]. There is a growing interest in mechanical-based therapies, such as ultrasound or phonophoresis (use of ultrasound along with topical gels or drugs to enhance their delivery), extracorporeal shockwave therapy (ESWT) and vibration [
Effects of therapeutic ultrasound on pain, physical functions and safety outcomes in patients with knee osteoarthritis: A systematic review and meta-analysis.
] on patient-centered and physical outcomes measures. Most of these meta-analyses did not perform subgroup, sensitivity, publication bias analyses, as well as did not provide recommendations based on the certainty of evidence. Our aim is to perform a systematic review with meta-analysis that compares the aforementioned passive mechanical-based treatments to control, placebo, other therapies and/or in combination with other interventions, employing more rigorous procedures which were not applied in the previous referred systematic reviews. The purpose of the review was to summarize the characteristics of mechanical-based interventions; and to evaluate the effect of mechanical-based interventions in patients with knee OA.
2. Methods
This systematic review with meta-analysis was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement [
Implementing the 27 PRISMA 2020 Statement items for systematic reviews in the sport and exercise medicine, musculoskeletal rehabilitation and sports science fields : the PERSiST (implementing Prisma in Exercise, Rehabilitation, Sport medicine and Spor.
Oliveira S, Andrade R, Valente C, Espregueira-mendes J, Silva F, Hinckel BB, et al. Effects of mechanical stimulation on knee cartilage for osteoarthritis treatment : protocol for a systematic review with meta-analysis. 2021.
We conducted a comprehensive electronic database search on Pubmed, Cochrane, Web of Science and EMBASE to identify clinical studies that analysed the effect of passive mechanical-based treatments on individuals with knee OA. The search strategy presented in supplemental Appendix S1 was applied from database inception to December 18, 2020. The reference list of the most relevant original studies and reviews were also retrieved to identify other potential eligible studies.
All records were exported to an Excel® file (Microsoft Office®) and duplicated studies were removed using the software filter. Manual verification was performed to check for any missing duplicate records. Two authors (S.O. and R.A.) screened all titles and abstracts independently, followed by evaluation of the full texts for eligibility. Three other reviewers (B.B.H., A.L. and O.C.) were consulted in case of disagreement.
2.2 Study selection
The eligibility criteria for the systematic review was framed according to the Participants, Intervention, Comparison, Outcomes and Study Design (PICOS) approach [
]. Only studies written in the English language were included.
2.2.1 Participants
We included individuals of both genders diagnosed with unilateral and/or bilateral knee OA, confirmed by imaging examination using the Kellgren–Lawrence grading system.
2.2.2 Intervention
We included studies that evaluated passive mechanical-based treatments for knee OA, the ones including therapeutic ultrasound and phonophoresis, ESWT and vibration, alone or in combination with other therapies. Studies including ultrasound only for diagnosis or as support for other therapies (i.e., ultrasound-guided) were excluded. Studies evaluating active mechanical-based treatments, such as exercises and manual therapy, were not considered.
2.2.3 Comparison
We included studies that compared the patients subjected to mechanical-based treatments with placebo, control (without intervention) and/or with other physical therapies.
2.2.4 Outcomes
We included any outcomes related to biochemical, clinical and imaging outcomes.
2.2.5 Study design
We included randomized and non-randomized clinical comparative studies. We excluded conference proceedings, reviews or meta-analysis, case studies and expert opinions.
2.3 Data extraction and quality assessment
All data related to study characteristics and outcomes from the included studies were extracted by one author (S.O.) and subsequently reviewed by four other authors (R.A., B.B.H., A.L. and O.C.). We retrieved data on study characteristics (year, study design), population characteristics (sample number, age, percentage of male/female, height, weight, body mass index [BMI], OA grade and respective classification criteria, and duration of symptoms) and interventions definition (definition used for intervention and comparisons groups). Intervention parameters were obtained depending on the intervention (e.g., operating mode, frequency, acceleration, amplitude, duty cycle, pulse duration, power density, energy density, number of pulses, stimulation time, treatment duration, stimulated area and the number of stimulated points). Measured variables and outcomes were collected and categorized into four different domains: biochemical (inflammation, cartilage degradation markers); patient-reported outcome measures (pain intensity, disability and quality of life); physical examination outcome measures (gait performance, knee range of motion, muscle strength); and imaging outcomes (cartilage and synovial thickness, muscle morphology).
Median, 25% and 75% percentiles, minimum and maximum values were calculated for each intervention parameter. Corresponding authors were contacted for additional data when needed. Missing raw data were also computed using the methods suggested by Cochrane Handbook [
]. This tool assesses five distinct bias domains: i) randomization, ii) deviations from intended interventions, iii) missing outcome data, iv) measurement of outcomes and v) selection of the reported result. Each of these domains were appraised and scored as “low risk”, “some concerns” and “high risk” of bias. The overall risk of bias judgement was based on the bias appraisal from the five domains.
The risk of bias of non-randomized studies was judged using the revised Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool version 1 [
]. This tool assesses seven distinct bias domains: i) confounding, ii) selection of patients, iii) classification of interventions, iv) deviations from intended interventions, v) missing outcome data, vi) measurement of outcomes, viii) selection of the reported result. Each of these bias domains were appraised and scored as “low risk”, “moderate risk”, “serious risk”, “critical risk” and “no information”. The overall risk of bias judgement was based on the bias appraisal from the seven domains.
The risk of financial bias, due to industry funding and/or conflict of interest from the authors of the study, was also judged in all studies as an additional bias criterion.
Three independent authors (S.O., R.A. and C.V.) appraised the risk of bias of all included studies and disagreements were resolved by consensus.
2.4 Data synthesis and analysis
Quantitative syntheses were conducted on RStudio 3.3.1 software (RStudio, Boston, MA, USA), using the “dmetar”, “meta” and “metafor” packages. Meta-analyses were conducted when there were three or more homogenous studies investigating the same intervention groups and reporting the same outcomes. Outcomes that were not eligible for meta-analysis were reported by narrative synthesis.
We used a random-effects model due to the small sample sizes and variations in study methods and control populations. Continuous outcomes were expressed as standardized mean differences (SMDs) with 95% confidence intervals (CI). The SMDs magnitude were interpreted as large (≥0.8), moderate (0.5–0.79) and weak (0.2–0.49) [
]. The level of statistical heterogeneity was established using I-squared (I2) statistics. We interpreted the statistical heterogeneity as not important (<50%), moderate (50–75%) and high (>75%) [
]. All meta-analyses were first stratified according to the type of intervention, and then by outcome metric, intervention type and follow-up - post-treatment evaluation time (immediately after treatment) and short-term evaluation (<13 weeks). Subgroup analyses were performed for each intervention according to the comparison group (type of exercises, type of therapeutic gels and other therapies). All main analyses were performed using the mean changes from baseline to follow-up, using a coefficient correlation of 0.5 to calculate the SD of the mean change. Sensitivity analyses were then conducted by calculating with coefficient correlation values of 0.3, 0.5 and 0.7. Additional sensitivity analyses were conducted by removing (1) studies with partial and/or total overlapping population, (2) studies with different characteristics, and (3) studies which SD values were calculated from p-values for differences in means, according to Cochrane handbook methods [
]. A minimum of two studies were required to remain after study removal to perform the sensitive analysis. Leave-one-out analysis and Baujat plots were conducted to evaluate the effect of any single study in the pooled SMD or heterogeneity. Publication bias was assessed by a funnel plot symmetry. We also carried out the trim-and-fill analysis to compute the publication bias corrected SMD and assess if it had an effect on the pooled outcome (either under- or over-estimated). Meta-regression analysis was conducted for each intervention according to stimulation parameters (i.e., number of sessions, power and energy density, frequency, stimulation time, number of pulses, amplitude), percentage of females and OA grade.
2.4.1 Certainty of evidence
The recommendations based on the strength of evidence were summarized using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach [
Schünemann H, Brozek J, Guyatt G, Oxman A, editors. GRADE handbook for grading quality of evidence and strength of recommendations. Updated October 2013. The GRADE Working Group, 2013. Available from guidelinedevelopment.org/handbook.; n.d.
]. Three authors (S.O., R.A. and C.V.) performed the GRADE appraisal.
3. Results
3.1 Search results and studies characteristics
The initial database and hand search generated a total of 10,689 records, of which 821 full texts were screened to assess their eligibility, from which 77 studies met the eligibility criteria and were included in this study (Figure 1). A total of 69 studies were randomized [
Therapeutic ultrasound versus sham ultrasound for the management of patients with knee osteoarthritis: a randomized double-blind controlled clinical study.
Does adding transcutaneous electrical nerve stimulation to therapeutic ultrasound affect pain or function in people with osteoarthritis of the knee? A randomized controlled trial.
Phonophoresis of Phyllanthus amarus nanoparticle gel improves functional capacity in individuals with knee osteoarthritis: A randomized controlled trial.
Effect of low-intensity long-duration ultrasound on the symptomatic relief of knee osteoarthritis: a randomized, placebo-controlled double-blind study.
Short-term effects of neuromuscular electrical stimulation and ultrasound therapies on muscle architecture and functional capacity in knee osteoarthritis: a randomized study.
Efficacy of focused low-intensity pulsed ultrasound therapy for the management of knee osteoarthritis: a randomized, double blind, placebo-controlled trial.
Phonopheresis associated with nanoparticle gel from phyllanthus amarus relieves pain by reducing oxidative stress and proinflammatory markers in adults with knee osteoarthritis.
The effectiveness of pulsed ultrasound treatment on pain, function, synovial sac thickness and femoral cartilage thickness in patients with knee osteoarthritis: a randomized, double-blind clinical, controlled study.
Loyola-sánchez A, Richardson J, Beattie KA, Otero-fuentes C, A AL, Richardson J, et al. Effect of low-intensity pulsed ultrasound on the cartilage repair in people with mild to moderate knee osteoarthritis: a double-blinded, randomized, placebo-controlled pilot study. Arch Phys Med Rehabil 2012;93:35–42. https://doi.org/10.1016/j.apmr.2011.07.196.
Efficacy and safety of a stimulator using low-intensity pulsed ultrasound combined with transcutaneous electrical nerve stimulation in patients with painful knee osteoarthritis.
Therapeutic effects of low-frequency phonophoresis with a Chinese herbal medicine versus sodium diclofenac for treatment of knee osteoarthritis: a double-blind, randomized, placebo-controlled clinical trial.
Efficacy of extracorporeal shock wave therapy versus mobilization with movement on pain, disability and range of motion in patients with knee osteoarthritis.
The influence of extracorporeal shockwave therapy and kinesiotherapy on health status in females with knee osteoarthritis: a randomized controlled trial.
Effects of radial extracorporeal shock wave therapy on clinical variables and isokinetic performance in patients with knee osteoarthritis: a prospective, randomized, single-blind and controlled trial.
Comparison of the effects between low- versus medium-energy radial extracorporeal shock wave therapy on knee osteoarthritis: A randomised controlled trial.
Effects of combination therapy and infrared radiation on pain, physical function, and quality of life in subjects with knee osteoarthritis: A randomized controlled study.
Whole body vibration training on muscle strength and brain-derived neurotrophic factor levels in elderly woman with knee osteoarthritis: A randomized clinical trial study.
The effect of adding whole-body vibration to squat training on the functional performance and self-report of disease status in elderly patients with knee osteoarthritis: a randomized, controlled clinical study.
Effects of Whole Body Vibration Exercise associated with Quadriceps Resistance Exercise on functioning and quality of life in patients with knee osteoarthritis: A randomized controlled trial.
Effects of whole-body vibration training with quadriceps strengthening exercise on functioning and gait parameters in patients with medial compartment knee osteoarthritis: a randomised controlled preliminary study.
Li W, Pan Y, Yang Q, Guo Z-G, Yue Q, Meng Q-G. Extracorporeal shockwave therapy for the treatment of knee osteoarthritis. Med (United States) 2018;97. https://doi.org/10.1097/MD.0000000000011418.
Effect of the combined intervention with passive whole-body vibration and auriculotherapy on the quality of life of individuals with knee osteoarthritis assessed by the WHOQOL-Bref: A multi-arm clinical trial.
Phonophoresis of Phyllanthus amarus nanoparticle gel improves functional capacity in individuals with knee osteoarthritis: A randomized controlled trial.
Phonopheresis associated with nanoparticle gel from phyllanthus amarus relieves pain by reducing oxidative stress and proinflammatory markers in adults with knee osteoarthritis.
The influence of extracorporeal shockwave therapy and kinesiotherapy on health status in females with knee osteoarthritis: a randomized controlled trial.
The effect of adding whole-body vibration to squat training on the functional performance and self-report of disease status in elderly patients with knee osteoarthritis: a randomized, controlled clinical study.
A total of 4798 participants were included with a weighted mean age of 60 ± 7.3 years old; 77% females. When reported, most patients displayed grade II and III (Kellgren–Lawrence grading system) and symptoms persistence for a median of 4.5 years (2.8–6.0). The population characteristics is displayed on supplemental Table S1 and the demographic data of each study is present on supplemental Tables S2 to S4.
3.2 Risk of bias
Most of the randomized clinical studies were judged as high risk of bias (70%, k = 48), with twenty studies showed some concerns (29%), and only one study (1%) had low risk of bias (Figure 2). More than one-third of the studies (36%, k = 25) were judged with high or unclear risk of randomization bias arising from the lack of information regarding allocation concealment. Bias due to deviation from intended interventions and measurement of the outcomes was observed in most studies (65%, k = 45) and about 28% of the studies (k = 19) were judged high or unclear risk of bias owing to missing data. Almost all studies (94%, k = 65) presented some concerns to high risk of selective reporting bias.
Figure 2Risk of bias plot for ROB-2 tool. Traffic lights and weight summary plots for randomized clinical studies.
Seven non-randomized studies presented serious risk of bias, while one study had critical risk of bias due to selective reporting (Figure 3). Bias due to confounding was found in five studies (63%) since most studies did not report the OA grade of the participants. High risk of selection bias, due to participant selection, was judged to be present in half of the studies (50%, k = 4) since those studies did not present any information whether all the eligible participants were selected for the clinical trial. Seventy five percent of the studies (k = 6) were judged with serious risk of bias caused by the classification of interventions because important parameters, such as stimulation time, were not reported. Six studies (75%) were judged with no information due to missing data. Bias due to deviations from intended interventions was detected in only one study (12%), with moderate risk, while five other studies (63%) did not provide any information. Two studies (25%) were judged with serious risk of detection bias because the measurement of outcomes was not blinded.
Figure 3Risk of bias plot for ROBIS-I tool. Traffic lights and weight summary plots for nonrandomized clinical studies.
Considering the risk of financial bias (Supplemental Figure S1), most of the studies stated no conflict of interests (77%, k = 59), others did not provide any information and, thus, were judged with unclear risk of bias (22%, k = 17). Only one study showed high risk of financial bias (1%).
3.3 Frequency of interventions
From the 77 included studies, 42 studies used ultrasound and/or phonophoresis [
Therapeutic ultrasound versus sham ultrasound for the management of patients with knee osteoarthritis: a randomized double-blind controlled clinical study.
Does adding transcutaneous electrical nerve stimulation to therapeutic ultrasound affect pain or function in people with osteoarthritis of the knee? A randomized controlled trial.
Phonophoresis of Phyllanthus amarus nanoparticle gel improves functional capacity in individuals with knee osteoarthritis: A randomized controlled trial.
Effect of low-intensity long-duration ultrasound on the symptomatic relief of knee osteoarthritis: a randomized, placebo-controlled double-blind study.
Short-term effects of neuromuscular electrical stimulation and ultrasound therapies on muscle architecture and functional capacity in knee osteoarthritis: a randomized study.
Efficacy of focused low-intensity pulsed ultrasound therapy for the management of knee osteoarthritis: a randomized, double blind, placebo-controlled trial.
Phonopheresis associated with nanoparticle gel from phyllanthus amarus relieves pain by reducing oxidative stress and proinflammatory markers in adults with knee osteoarthritis.
The effectiveness of pulsed ultrasound treatment on pain, function, synovial sac thickness and femoral cartilage thickness in patients with knee osteoarthritis: a randomized, double-blind clinical, controlled study.
Loyola-sánchez A, Richardson J, Beattie KA, Otero-fuentes C, A AL, Richardson J, et al. Effect of low-intensity pulsed ultrasound on the cartilage repair in people with mild to moderate knee osteoarthritis: a double-blinded, randomized, placebo-controlled pilot study. Arch Phys Med Rehabil 2012;93:35–42. https://doi.org/10.1016/j.apmr.2011.07.196.
Efficacy and safety of a stimulator using low-intensity pulsed ultrasound combined with transcutaneous electrical nerve stimulation in patients with painful knee osteoarthritis.
Therapeutic effects of low-frequency phonophoresis with a Chinese herbal medicine versus sodium diclofenac for treatment of knee osteoarthritis: a double-blind, randomized, placebo-controlled clinical trial.
Effects of combination therapy and infrared radiation on pain, physical function, and quality of life in subjects with knee osteoarthritis: A randomized controlled study.
Efficacy of extracorporeal shock wave therapy versus mobilization with movement on pain, disability and range of motion in patients with knee osteoarthritis.
The influence of extracorporeal shockwave therapy and kinesiotherapy on health status in females with knee osteoarthritis: a randomized controlled trial.
Effects of radial extracorporeal shock wave therapy on clinical variables and isokinetic performance in patients with knee osteoarthritis: a prospective, randomized, single-blind and controlled trial.
Comparison of the effects between low- versus medium-energy radial extracorporeal shock wave therapy on knee osteoarthritis: A randomised controlled trial.
Li W, Pan Y, Yang Q, Guo Z-G, Yue Q, Meng Q-G. Extracorporeal shockwave therapy for the treatment of knee osteoarthritis. Med (United States) 2018;97. https://doi.org/10.1097/MD.0000000000011418.
Whole body vibration training on muscle strength and brain-derived neurotrophic factor levels in elderly woman with knee osteoarthritis: A randomized clinical trial study.
The effect of adding whole-body vibration to squat training on the functional performance and self-report of disease status in elderly patients with knee osteoarthritis: a randomized, controlled clinical study.
Effects of Whole Body Vibration Exercise associated with Quadriceps Resistance Exercise on functioning and quality of life in patients with knee osteoarthritis: A randomized controlled trial.
Effects of whole-body vibration training with quadriceps strengthening exercise on functioning and gait parameters in patients with medial compartment knee osteoarthritis: a randomised controlled preliminary study.
Effect of the combined intervention with passive whole-body vibration and auriculotherapy on the quality of life of individuals with knee osteoarthritis assessed by the WHOQOL-Bref: A multi-arm clinical trial.
Placebo (k = 6) and phonophoresis (k = 6) were the most used comparison groups for ultrasound, while phonophoresis was compared to placebo (k = 1) and compared with different therapeutic gels (k = 3). Both interventions were often combined with other therapies (supplemental Table S5). Most of the studies (79%, k = 33) used ultrasound and phonophoresis as continuous (supplemental Table S6), with median resonance frequency of 1 MHz (range, 0.04 to 3.00) at a median power density of 1 W/cm2 (range, 0.1 to 2.5) for a median stimulation time of 9.5 min (range, 3 to 240) and a median number of sessions of 10 (range, 9 to 80). Supplemental Table S7 shows the intervention parameters of ultrasound and phonophoresis used in each study.
Therapeutic ultrasound versus sham ultrasound for the management of patients with knee osteoarthritis: a randomized double-blind controlled clinical study.
Phonophoresis of Phyllanthus amarus nanoparticle gel improves functional capacity in individuals with knee osteoarthritis: A randomized controlled trial.
Effect of low-intensity long-duration ultrasound on the symptomatic relief of knee osteoarthritis: a randomized, placebo-controlled double-blind study.
Short-term effects of neuromuscular electrical stimulation and ultrasound therapies on muscle architecture and functional capacity in knee osteoarthritis: a randomized study.
Efficacy of focused low-intensity pulsed ultrasound therapy for the management of knee osteoarthritis: a randomized, double blind, placebo-controlled trial.
The effectiveness of pulsed ultrasound treatment on pain, function, synovial sac thickness and femoral cartilage thickness in patients with knee osteoarthritis: a randomized, double-blind clinical, controlled study.
Loyola-sánchez A, Richardson J, Beattie KA, Otero-fuentes C, A AL, Richardson J, et al. Effect of low-intensity pulsed ultrasound on the cartilage repair in people with mild to moderate knee osteoarthritis: a double-blinded, randomized, placebo-controlled pilot study. Arch Phys Med Rehabil 2012;93:35–42. https://doi.org/10.1016/j.apmr.2011.07.196.
Efficacy and safety of a stimulator using low-intensity pulsed ultrasound combined with transcutaneous electrical nerve stimulation in patients with painful knee osteoarthritis.
Effects of combination therapy and infrared radiation on pain, physical function, and quality of life in subjects with knee osteoarthritis: A randomized controlled study.
] with ultrasound and/or phonophoresis interventions were included in the meta-analyses (Table 1). Six out of twenty-three (26%) meta-analyses showed the superiority of ultrasound over other therapies, whereas phonophoresis was superior to ultrasound in one meta-analysis (4%). All meta-analyses were judged as very-low certainty of evidence. Forest plots are available in supplemental Figure S2.
Table 1Quantitative meta-analysis results for each outcome measure of ultrasound and phonophoresis interventions.
* – serious study limitations due to risk of bias;
** – very serious study limitations due to risk of bias; ¶ – serious imprecision due to wide 95% CIs (cross the 0.5 or −0.5 SMD);¶¶ – very serious imprecision due to very wide 95% CIs (cross both the −0.5 & 0.5 SMD);§ – serious inconsistency due to substantial & significant heterogeneity &/or 95% CI do not overlap;§§ – very serious inconsistency due to substantial & significant heterogeneity & 95% CI do overlap with at least 2 studies;
ƚ – serious indirectness due to indirect comparisons;
ƚƚ – very serious indirectness due to indirect comparisons & differences in intervention;& – serious publication bias (studies added &/or asymmetric funnel plots);
&& – very serious publication bias (studies added, asymmetric funnel plots & different pooled effects).
Ultrasound was statistically significantly superior to placebo in WOMAC pain (SMD = −0.47, 95% CI −0.79 to −0.14, weak effect), function (SMD = −0.46, 95% CI −0.70 to −0.23, weak effect), total score (SMD = −0.42, 95% CI to −0.76 to −0.09, weak effect), Lequesne pain (SMD = −0.71, 95% CI −1.25 to −0.16, moderate effect) and pain at rest (SMD = −0.38, 95% CI −0.66 to −0.10, weak effect), regardless the combination with other therapies. Pain at activity (SMD = −0.96, 95% CI −1.37 to −0.56, large effect), WOMAC total (SMD = −0.75, 95% CI −1.15 to −0.36, moderate effect) and Lequesne pain (SMD = −0.75, 95% CI −1.15 to −0.36, moderate effect) showed statistically significant improvements with ultrasound plus diclofenac versus placebo plus diclofenac. By adding ultrasound to other therapies, the pain at rest (SMD = −0.91, 95% CI −1.79 to −0.04, large effect) was statistically significantly reduced compared to other therapies alone, except when compared to exercises alone. By contrast, ultrasound with exercise plus hot packs was statistically inferior to kinesio taping combined with exercise plus hot packs in decreasing the pain at rest (SMD = 3.90, 95% CI 2.38 to 5.43, large effect).
Ultrasound was statistically significantly inferior to phonophoresis for pain at rest when phyllanthus amarus gel (SMD = 3.03, 95% CI 2.10 to 3.97, large effect) or piroxicam (SMD = 0.93, 95% CI 0.32 to 1.54, large effect) were used during phonophoresis. However, when combining all types of gel, the overall pooled effect was not statistically significant. Combining ultrasound or phonophoresis with exercise and TENS, ultrasound was statistically inferior to phonophoresis in significantly improving the pain at rest (SMD = 1.64, 95% 0.95 to 2.33, large effect), WOMAC pain (SMD = 1.36, 95% CI 0.70 to 2.02, large effect), function (SMD = 1.44, 95% CI 0.77 to 2.11, large effect), stiffness (SMD = 1.26, 95% CI 0.61 to 1.92, large effect) and total score (SMD = 1.69, 95% CI 0.99 to 2.39, large effect). Combined with exercise plus hot packs, ultrasound was also statistically inferior to phonophoresis for WOMAC function (SMD = 0.46, 95% CI 0.07 to 0.85, weak effect) and total score (SMD = 0.48, 95% CI 0.15 to 0.80, weak effect). The application of diclofenac or dexamethasone gels during phonophoresis promoted statistically significant improvements in WOMAC index (moderate to large effect). However, the overall pooled effect for WOMAC stiffness outcome was the only one that changed significantly (moderate effect).
3.5 Extracorporeal shockwave therapy
The ESWT was more commonly compared to other therapies (k = 7) and combined with exercises (k = 5), while only two studies comparing the ESWT to placebo (supplemental Table S5). Nearly half of the studies (45%, k = 9) used radial ESWT (supplemental Table S6). ESWT was applied with a median frequency of 8 Hz (range, 4 to 16) with a median number of 2,000 pulses (range, 1,000 to 4,000) at a median energy density of 0.178 mJ/mm2 (range, 0.020 to 4.000), completing a median number of 4.5 sessions (range, 3 to 12). Supplemental Table S8 shows the ESWT parameters for each study.
Efficacy of extracorporeal shock wave therapy versus mobilization with movement on pain, disability and range of motion in patients with knee osteoarthritis.
The influence of extracorporeal shockwave therapy and kinesiotherapy on health status in females with knee osteoarthritis: a randomized controlled trial.
Effects of radial extracorporeal shock wave therapy on clinical variables and isokinetic performance in patients with knee osteoarthritis: a prospective, randomized, single-blind and controlled trial.
Comparison of the effects between low- versus medium-energy radial extracorporeal shock wave therapy on knee osteoarthritis: A randomised controlled trial.
] assessing the effects of ESWT were included in the meta-analysis (Table 2). Five out of eleven (45%) meta-analyses showed the superiority of ESWT over other therapies. All meta-analyses were judged as very-low certainty of evidence. The ESWT was consistently superior to other therapies in the improvement of pain at rest (SMD = −1.20, 95% CI −2.32 to −0.09, large effect) and WOMAC total (SMD = −2.46, 95% CI −3.42 to −1.51, large effect) outcomes, except when compared to ultrasound. ESWT combined with exercise statistically significantly reduced pain at rest versus interferential current plus exercise (SMD = −1.67, 95% CI −2.61 to −0.74, large effect). ESWT plus standard of care (i.e., heat packs, interference current therapy and ultrasound) was statistically superior to placebo plus usual care in significantly decreasing pain at rest (SMD = −0.69, 95% CI, −1.09 to −0.29, moderate effect), WOMAC pain (SMD = −0.68, 95% CI −1.08 to −0.29, moderate effect) and function (SMD = −0.66, 95% CI −1.05 to −0.26, moderate effect). The combination of ESWT with TENS also reduced the WOMAC pain (SMD = −0.64, 95% CI −1.11 to −0.17, moderate effect) and function (SMD = −0.80, 95% CI −1.28 to −0.32, moderate effect) as compared to placebo plus TENS. ESWT plus hot packs statistically significantly decreased WOMAC pain (SMD = −0.65, 95% CI −1.05 to −0.26, moderate effect) versus placebo plus hot packs. At short follow-up, adding ESWT to exercise, hot packs or standard of care, statistically improved the WOMAC pain (SMD = −1.01, 95% CI −1.70 to −0.33, large effect) and function (SMD = −1.06, 95% CI −1.99 to −0.13, large effect) scores. WOMAC stiffness (SMD = −1.00, 95% CI −1.53 to −0.47, large effect) was only statistically improved after ESWT plus exercise compared to placebo plus exercise. Forest plots are available in supplemental Figure S3.
Table 2Quantitative meta−analysis results for each outcome measure of extracorporeal shockwave intervention.
* – serious study limitations due to risk of bias;
** – very serious study limitations due to risk of bias; ¶ – serious imprecision due to wide 95% CIs (cross the 0.5 or −0.5 SMD);¶¶ – very serious imprecision due to very wide 95% CIs (cross both the −0.5 & 0.5 SMD);§ – serious inconsistency due to substantial & significant heterogeneity &/or 95% CI do not overlap;§§ – very serious inconsistency due to substantial & significant heterogeneity & 95% CI do overlap with at least 2 studies;
ƚ – serious indirectness due to indirect comparisons;
ƚƚ – very serious indirectness due to indirect comparisons & differences in intervention;& – serious publication bias (studies added &/or asymmetric funnel plots);
&& – very serious publication bias (studies added, asymmetric funnel plots & different pooled effects).
Nearly half of the studies explored the vibration in combination with exercises (k = 7). Placebo, control and other therapies were also used as comparisons groups to vibration (supplemental Table S5). Vertical vibration was the most commonly mode employed (65%, k = 11) targeting the whole body (supplemental Table S6). A median frequency of 35 Hz (range, 5 to 150) with a median amplitude of 3.5 mm (range, 0.2 to 7.5) for a median duration of 20 min (range, 9 to 50) and a median number of 24 sessions (range, 9 to 120) were used. Supplemental Table S9 reports the vibration parameters for each study.
The effect of adding whole-body vibration to squat training on the functional performance and self-report of disease status in elderly patients with knee osteoarthritis: a randomized, controlled clinical study.
Effects of Whole Body Vibration Exercise associated with Quadriceps Resistance Exercise on functioning and quality of life in patients with knee osteoarthritis: A randomized controlled trial.
Effects of whole-body vibration training with quadriceps strengthening exercise on functioning and gait parameters in patients with medial compartment knee osteoarthritis: a randomised controlled preliminary study.
] applying vibration plus exercise were included in the meta-analysis (Table 3). Three out of eight (38%) meta-analyses showed the superiority of vibration over other therapies. All meta-analyses were judged as very-low certainty of evidence. Vibration with exercise was superior to exercise alone in statistically improving pain at rest (SMD = −0.46, 95% CI −0.68 to −0.25, weak effect), WOMAC function (SMD = −0.43, 95% CI −0.56 to −0.29, weak effect) and timed up and go test (SMD = −0.80, 95% CI −1.14 to −0.46, large effect). Forest plots are available in supplemental Figure S4.
Table 3Quantitative meta−analysis results for each outcome measure of vibration intervention.
Outcome
Intervention
Control
Subgroup
Main findings
GRADE
Supplemental Figure S3
k, n
I2, P value
SMD (95% CI)
Pain at rest After treatment
Vibration + Ex
Ex
Home-based
1, 36
NA
−0.42 (−1.08 to 0.24)
⊕◯◯◯ [*, ¶, ƚ ƚ]
3.1
Squat
1, 41
NA
−0.40 (−1.02 to 0.22)
Quadriceps
1, 39
NA
−0.56 (−1.20 to 0.08)
OVERALL
3, 116
0%, 0.93
−0.46 (−0.68 to −0.25)
WOMAC Pain After treatment
Vibration + Ex
Ex
Squat
1, 21
NA
−0.42 (−1.29 to 0.45)
⊕◯◯◯ [*, ¶, ƚ ƚ, &&]
Quadriceps
2, 138
47.9%, 0.17
−0.59 (−3.89 to 2.72)
3.2
OVERALL
3, 159
6.0%, 0.35
−0.57 (−1.31 to 0.16)
WOMAC Function After treatment
Vibration + Ex
Ex
Squat
1, 21
NA
−0.32 (−1.18 to 0.55)
⊕◯◯◯ [*, ¶, ƚ ƚ]
Quadriceps
2, 138
0%, 0.96
−0.44 (−0.55 to −0.34)
3.3
OVERALL
3, 159
0%, 0.96
−0.43 (−0.56 to −0.29)
WOMAC Stiffness After treatment
Vibration + Ex
Ex
Squat
1, 21
NA
−0.23 (−1.09 to 0.63)
⊕◯◯◯ [*, ¶¶, §, ƚ ƚ, &]
Quadriceps
2, 138
89.4%, 0.002
−0.10 (−7.63 to 7.43)
3.4
OVERALL
3, 159
79.0%, 0.009
−0.15 (−1.70 to 1.39)
6 min Walk After treatment
Vibration + Ex
Ex
Squat
2, 62
0%, 0.81
−0.11 (−0.87 to 0.66)
⊕◯◯◯ [*, ¶¶, §§, ƚ ƚ, &]
Quadriceps
2, 138
95.3%, <0.0001
2.00 (−14.21 to 18.21)
3.5
OVERALL
4, 200
92%, <0.0001
0.95 (−1.58 to 3.48)
TUG After treatment
Vibration + Ex
Ex
Home-based
1, 38
NA
−1.03 (−1.81 to −0.24)
⊕◯ [*, ¶, ƚ ƚ, &]
3.6
Squat
2, 62
0%, 0.80
−0.60 (−1.43 to 0.24)
Quadriceps
2, 138
49.2%, 0.16
−0.83 (−4.24 to 2.58)
OVERALL
5, 238
0%, 0.53
−0.80 (−1.14 to −0.46)
Muscle strength extensors After treatment
Vibration + Ex
Ex
Home-based
2, 74
94.5%, <0.001
1.45 (−15.21 to 18.10)
⊕◯ [*, ¶¶, §§, ƚ ƚ, &]
3.7
Squat
2, 80
83.5%, 0.01
0.68 (−6.69 to 8.06)
Quadriceps
1, 99
NA
0.22 (−0.18 to 0.61)
OVERALL
5, 253
87.2%, <0.0001
0.87 (−0.53 to 2.26)
Muscle strength flexors After treatment
Vibration + Ex
Ex
Home-based
1, 38
NA
0.44 (−0.31 to 1.20)
⊕◯ [*, ¶¶, ƚ ƚ]
3.8
Squat
2, 80
0%, 0.35
−0.58 (−3.28 to 2.12)
Quadriceps
1, 99
NA
−0.15 (−0.54 to 0.25)
OVERALL
3, 217
53%, 0.09
−0.23 (−1.01 to 0.54)
Ex: Exercise; NA: Not applicable.
* – serious study limitations due to risk of bias;
** – very serious study limitations due to risk of bias; ¶ – serious imprecision due to wide 95% CIs (cross the 0.5 or −0.5 SMD); ¶¶ – very serious imprecision due to very wide 95% CIs (cross both the −0.5 & 0.5 SMD);§ – serious inconsistency due to substantial & significant heterogeneity &/or 95% CI do not overlap;§§ – very serious inconsistency due to substantial & significant heterogeneity & 95% CI do overlap with at least 2 studies;
ƚ – serious indirectness due to indirect comparisons;
ƚƚ – very serious indirectness due to indirect comparisons & differences in intervention;& – serious publication bias (studies added &/or asymmetric funnel plots);
&& – very serious publication bias (studies added, asymmetric funnel plots & different pooled effects).
3.7 Sensitivity analysis, publication bias and meta-regression
More than half of the meta-analyses (52%) displayed moderate to high heterogeneity. After excluding the study with the highest contribution to the pooled heterogeneity in the leave-one-out analysis, 19% of the meta-analyses remained with moderate to high heterogeneity, while 64% of the meta-analyses showed an I2 of 0%. By analysing the Baujat statistics, the most influencing study contributed between 0.01 to 29.64 of the pooled heterogeneity and generated an influence on the pooled effect size ranging from 0.01 to 26.11 (supplemental Table S10). The leave-one-out analyses for the highest and lowest effect size revealed that only 17% of the meta-analyses significantly changed in pooled effect after excluding the study with the lowest effect (supplemental Table S11). The baujat and leave-one-out analyses plots are available in OSF repository [
Oliveira S, Andrade R, Valente C, Espregueira-mendes J, Silva F, Hinckel BB, et al. Effects of mechanical stimulation on knee cartilage for osteoarthritis treatment : protocol for a systematic review with meta-analysis. 2021.
The sensitivity analysis for the studies with coefficient correlation value of 0.3 demonstrated that only one of 44 meta-analyses changed the pooled effect, while calculating the improvement with a coefficient correlation value of 0.7, seven of 44 meta-analyses varied significantly (supplemental Table S12). These differences were only observed in the subgroups analyses and not in the overall pooled meta-analysis. The forest plots for the sensitivity analysis are available in OSF repository [
Oliveira S, Andrade R, Valente C, Espregueira-mendes J, Silva F, Hinckel BB, et al. Effects of mechanical stimulation on knee cartilage for osteoarthritis treatment : protocol for a systematic review with meta-analysis. 2021.
]. By removing studies with specific characteristics and/or with partial overlapping population, there were significant alterations in the pooled effect in four of 18 meta-analyses (supplemental Table S13).
Publication bias was detected in 74% of the meta-analyses owing to funnel plot asymmetry. The trim-and-fill analyses corrected the pooled effect for publication bias, indicating that two of 44 meta-analyses changed the SMDs to non-significative, while three of 44 meta-analyses changed to significative (supplemental Table S11).
Meta-regression was implemented to assess the influence of potential covariates in the pooled effect size (supplemental Table S14). For the meta-analyses comparing ultrasound to placebo (combining or not with other therapies), the number of sessions, power density, stimulation time and percentage of females influenced the pooled effect size for pain at rest, WOMAC stiffness and Lequesne pain. Comparing ultrasound plus other therapies versus other therapies alone, the percentage of females induced significant effects on the pooled effect size for pain at rest. The number of pulses and energy density were significant covariates that led to significant differences on pain at rest when adding exercise to ESWT. Percentage of females affected also the 6-minute walk distance when comparing vibration plus exercise to only exercise. The meta-regression plots are available in OSF repository [
Oliveira S, Andrade R, Valente C, Espregueira-mendes J, Silva F, Hinckel BB, et al. Effects of mechanical stimulation on knee cartilage for osteoarthritis treatment : protocol for a systematic review with meta-analysis. 2021.
The biochemical outcomes were only assessed in the studies (k = 5) applying ultrasound, phonophoresis and vibration (supplemental Table S15), revealing that ultrasound did not affect the inflammation and antioxidant markers, but vibration reduced the cartilage degradation markers content (supplemental Appendix S2). Patient-reported and physical examination outcome measures were reported for all intervention studies (supplemental Table S15), but the improvement as compared to comparison groups was inconsistent and conflicting across studies (supplemental Appendix S2). The imaging outcomes were only analysed in the studies (k = 5) applying ultrasound and ESWT (supplemental Table S15), but no significant improvement was reported following those interventions (supplemental Appendix S2).
4. Discussion
The main findings of our systematic review with meta-analysis are that mechanical-based treatments may significantly reduce pain intensity and disability in some of the patients with knee OA with weak to large effects, although with very-low certainty of evidence.
Despite the statistically significant effects, our findings should be interpreted with some caution. Most of the meta-analyses displayed moderate to high heterogeneity (<50%), small sample sizes (<400), the subgroups analyses were performed with small number of studies and the certainty of evidence was very low for all meta-analyses. Some quantitative syntheses exhibited imprecise results with studies presenting pooled SMD effects size crossing both the −0.5 and 0.5 of the 95% CIs. Meta-regression demonstrated that some intervention parameters and percentage of females influenced the pooled effect size. Interestingly, no effect was observed for OA grade which may be due to the lack of studies reporting the OA grade. There were, though, a small number of studies in the meta-regression analyses, limiting these conclusions. The narrative synthesis also presented inconsistent and conflicting results, which may be also explained by the small sample size of the studies.
Ultrasound statistically significantly improved the pain and WOMAC outcomes versus placebo, combined or not with other therapies, potentiating, particularly, the effects of diclofenac in comparison to placebo plus diclofenac. Ultrasound plus other therapies significantly improved the pain at rest versus other therapies alone, while it did not induce statistically significant improvements versus other therapies when combined with exercise or hot packs. Prior systematic reviews with meta-analysis have also showed that ultrasound reduced the patient́s pain and improved physical performance, but those meta-analyses only considered the comparison group placebo, exercises and no intervention as one [
Effects of therapeutic ultrasound on pain, physical functions and safety outcomes in patients with knee osteoarthritis: A systematic review and meta-analysis.
]. We conducted meticulous meta-analyses for each control group to establish more reliable comparisons, which confirmed that ultrasound was superior to placebo and may be combined with other therapies to improve the OA symptoms.
Ultrasound and phonophoresis were compared in only one systematic review [
]. Our meta-analyses showed that when comparing ultrasound versus phonophoresis, both in adjunct to other therapies, phonophoresis led to statistically significant effects on pain intensity and WOMAC subscales, but those effects were only observed in the subgroup analyses according to the type of combined intervention or therapeutic gel used. On the other hand, only one of six meta-analyses showed statistically significant effects of phonophoresis in reducing WOMAC stiffness versus ultrasound. This suggests that the type of gel or the combined therapy strongly affected the clinical outcomes of phonophoresis when compared to ultrasound.
ESWT statistically improved the pain at rest and WOMAC total score with large effects versus other physical therapies such as kinesio taping, intra-articular injection or standard of care. Several systematic reviews with meta-analysis [
Extracorporeal shockwave therapy improves pain and function in subjects with knee osteoarthritis: A systematic review and meta-analysis of randomized clinical trials.
] indicated that ESWT was superior to placebo and to other physical therapies, being in line with our findings. We also studied the effects of ESWT in combination with other therapies. ESWT plus other physical therapies resulted in statistically significant effects in combination with exercise, standard of care, TENS or hot packs versus placebo plus those therapies after treatment or at short-term follow-up, suggesting its potential use in combination with other therapies to improve the pain and disability outcomes.
Our meta-analyses displayed statistically significant effects in improving pain intensity, WOMAC function and time up and go test following vibration plus exercise versus exercise, which is in agreement with two previous systematic reviews with meta-analysis [
] did not report any difference after vibration. Our results indicate a potential benefit of adding vibration to exercise in improving pain, disability and functional performance.
The stimulation parameters influence deeply the level of improvement in the outcome measures, being extremely relevant the understanding of the dosages applied in these clinical studies. One trial demonstrated that ultrasound for 8 minutes were superior than for 4 minutes [
Comparison of the effects between low- versus medium-energy radial extracorporeal shock wave therapy on knee osteoarthritis: A randomised controlled trial.
]. Most of studies did not fully report all stimulation parameters, precluding further conclusions on the influence of stimulation dosage. We have also not performed subgroup analysis by intervention parameters as it would often result in carrying those analyses with only one study.
The inconsistent results reported in our review can be elucidated by the design of clinical trials and the intervention parameters. Most of the randomized or non-randomized studies displayed high and serious risk of bias, respectively, mainly due to lack of information about the methodology employed in the clinical trials and intervention parameters, missing data and selective reporting. These results highlight the need to implement more rigorous and transparent methodology in future clinical trials. The intervention parameters may also induce conflicting results and contribute to the heterogeneity of the meta-analyses where studies differed slightly in duration, intensity, operating mode and frequency of interventions. Our meta-regression results for ultrasound studies indicate that the pooled effect size increased with higher number of sessions and stimulation time and lower intensities. The studies applying ESWT versus other therapies showed significant effect size when analysing as covariates the number of pulses and energy density (or intensity), suggesting that lower energy density and lower number of pulses resulted in higher contribution to pooled effect sizes. Vibration studies did not show any impact on the effect size for any intervention parameters.
Some limitations of our systematic review with meta-analysis must be acknowledged. There were some deviations to the protocol, whose justifications can be found in detail in the supplemental Appendix S3. The majority of meta-analyses assessed the effects only after treatment due to the lack studies with follow-up time points, with only three meta-analyses focusing on short follow-up and none for mid- and long-term follow-up. Some studies were not included in the meta-analyses due to missing data, and partial or total overlapping samples, but we might have missed other potential studies with overlapping population. We contacted the authors of the included trials for missing data, but most of them did not provide any information. There were also other possible meta-analyses that could be conducted, but the available studies were not homogeneous or in enough number. These limitations hampered us to perform further quantitative comparisons among the intervention studies and to assess their long-term effects. Other limitation is using combination of therapies in the meta-analyses. Despite statistically significant effects were observed, it is not possible to ascertain if the effects arose from the isolated addition of the mechanical-based intervention or from the combination of both therapies.
In one single systematic review with meta-analysis, three mechanical-based therapies were analysed (isolated or combined with other therapies) compared to placebo, other therapies and combination of therapies; in terms of pain, disability and function outcomes, evidencing also the range of parameters used. This comprehensive review summarizes and highlights the effects of using such therapies in knee OA management, which information is of utmost importance for patients and to guide clinicians and physiotherapists for clinical reasoned decisions. These interventions are not currently recommended to treat knee OA due to low-quality evidence of the available clinical trials [
]. Our findings confirm the statistically significant effects of using these passive therapies to reduce OA symptoms, but this evidence is limited by the very-low certainty of its effects, whose clinical relevance was not investigated. Clinicians and researchers should evaluate the use of these interventions in long-term to give a whole picture of their potential effects in knee OA management, since patients desire therapies that provide them longstanding pain relief and disability improvements. Further clinical trials should be committed in following rigorous and high-quality methodologies and to adhere to the Consolidated, Standard of Reporting Trials (CONSORT), as many included studies did not fully comply to the CONSORT guidelines [
]. Clinicians and researchers must clearly provide the stimulation parameters that are critical for the intervention reproducibility (e.g., number of sessions, stimulation time, intensity, frequency) and to reliably compare studies. This fact underlines the need for further homogeneous studies with extractable data and information fully described to achieve more meaningful and relevant conclusions about the effects of the passive mechanical-based therapies on pain, disability and physical performance of patients with knee OA. In the advent of future homogeneous studies that compare these therapies to each other or to homogenous control groups, further analysis can be made using a network meta-analysis framework.
5. Conclusions
Although the statistically significant effects are based on very-low certainty, the potential benefits of using passive mechanical-based therapies, in combination or not with other physical therapies, should not be disregard. Therefore, we may cautiously recommend that clinicians might use them in specific clinical situations to reduce pain and disability in patients with knee OA.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We would like to acknowledge the Fundação para a Ciência e Tecnologia (FCT) for the funding through the references: UIDB/04436/2020, UIDP/04436/2020, Stimcart -PTDC/EME-EME/4520/2021 and the PhD fellowship grant UI/BD/150951/2021.
Authors' contribution
All authors were involved both in the idealization of the systematic review and preparation of the protocol and manuscript. S.O. and R.A. were involved in the databases searches and data extraction. S.O. performed the data synthesis and organization in coordination with R.A., B.B.H., A.L. and O.C. S.O., R.A. and C.V. appraised the risk of bias of the studies included in the systematic review and summarized the certainty of recommendation using GRADE. S.O., R.A. and C.V. performed the meta-analysis. B.B.H., F.S.S. and J.EM. guided and provided advice during all steps of the development of the systematic review. All authors contributed to drafting and approving the final manuscript prior to submission to the peer-reviewed journal.
Systematic review registration
This systematic review with meta-analysis was prospectively registered at Open Science Framework (https://osf.io/qdr6e)
Appendix A. Supplementary data
The following are the Supplementary data to this article:
Effects of therapeutic ultrasound on pain, physical functions and safety outcomes in patients with knee osteoarthritis: A systematic review and meta-analysis.
Implementing the 27 PRISMA 2020 Statement items for systematic reviews in the sport and exercise medicine, musculoskeletal rehabilitation and sports science fields : the PERSiST (implementing Prisma in Exercise, Rehabilitation, Sport medicine and Spor.
Oliveira S, Andrade R, Valente C, Espregueira-mendes J, Silva F, Hinckel BB, et al. Effects of mechanical stimulation on knee cartilage for osteoarthritis treatment : protocol for a systematic review with meta-analysis. 2021.
Schünemann H, Brozek J, Guyatt G, Oxman A, editors. GRADE handbook for grading quality of evidence and strength of recommendations. Updated October 2013. The GRADE Working Group, 2013. Available from guidelinedevelopment.org/handbook.; n.d.
Therapeutic ultrasound versus sham ultrasound for the management of patients with knee osteoarthritis: a randomized double-blind controlled clinical study.
Does adding transcutaneous electrical nerve stimulation to therapeutic ultrasound affect pain or function in people with osteoarthritis of the knee? A randomized controlled trial.
Phonophoresis of Phyllanthus amarus nanoparticle gel improves functional capacity in individuals with knee osteoarthritis: A randomized controlled trial.
Effect of low-intensity long-duration ultrasound on the symptomatic relief of knee osteoarthritis: a randomized, placebo-controlled double-blind study.
Short-term effects of neuromuscular electrical stimulation and ultrasound therapies on muscle architecture and functional capacity in knee osteoarthritis: a randomized study.
Efficacy of focused low-intensity pulsed ultrasound therapy for the management of knee osteoarthritis: a randomized, double blind, placebo-controlled trial.
Phonopheresis associated with nanoparticle gel from phyllanthus amarus relieves pain by reducing oxidative stress and proinflammatory markers in adults with knee osteoarthritis.
The effectiveness of pulsed ultrasound treatment on pain, function, synovial sac thickness and femoral cartilage thickness in patients with knee osteoarthritis: a randomized, double-blind clinical, controlled study.
Loyola-sánchez A, Richardson J, Beattie KA, Otero-fuentes C, A AL, Richardson J, et al. Effect of low-intensity pulsed ultrasound on the cartilage repair in people with mild to moderate knee osteoarthritis: a double-blinded, randomized, placebo-controlled pilot study. Arch Phys Med Rehabil 2012;93:35–42. https://doi.org/10.1016/j.apmr.2011.07.196.
Efficacy and safety of a stimulator using low-intensity pulsed ultrasound combined with transcutaneous electrical nerve stimulation in patients with painful knee osteoarthritis.
Therapeutic effects of low-frequency phonophoresis with a Chinese herbal medicine versus sodium diclofenac for treatment of knee osteoarthritis: a double-blind, randomized, placebo-controlled clinical trial.
00070-9&id=gr1.jpg){kind=link}
00070-9&id=gr2.jpg){kind=link}
00070-9&id=gr3.jpg){kind=link}