U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Cover of Exercise for the Management of Knee Osteoarthritis: A Review of Clinical Effectiveness

Exercise for the Management of Knee Osteoarthritis: A Review of Clinical Effectiveness

CADTH Rapid Response Report: Summary with Critical Appraisal

and .

Ottawa (ON): Canadian Agency for Drugs and Technologies in Health; .

Context and Policy Issues

Knee osteoarthritis (OA) is a chronic progressive condition and a major cause of muscleoskeletal disability in older populations.1,2 Knee OA symptoms include pain, decreased mobility, decreased muscle strength, stiffness, decreased physical function and quality of life.2 Currently there exists no cure for knee OA with management strategies consisting of exercise, weight loss, dietary supplements, biologic agents, orthoses, topical and systemic analgesics, as well as surgical interventions. Lifelong treatment with nonpharmacological therapies aimed at symptom reduction and prevention of further joint damage are generally recommended.3

Pain management with opioids is often prescribed for knee OA patients with severe pain that experience insufficient response to first-line pharmacological and nonpharmacological interventions.4 Functional limitation, female gender, poor self-reported health status, chronic obstructive pulmonary disease, and musculoskeletal disease in addition to knee OA correlate with prescribed opioids.5 Opioids pose a significant risk of side effects and adverse events in all knee OA patients and safer alternative treatments are preferred.3 Management and research for knee OA pain increasingly favours nonpharmacological interventions.4

The purpose of this report is to retrieve and review the existing clinical effectiveness evidence on the treatment of knee OA patients with exercise as compared to no intervention, placebo, or opioids.

Research Question

What is the clinical effectiveness of exercise for the management of knee osteoarthritis?

Key Findings

Evidence from 18 identified systematic reviews suggested the clinical efficacy of exercise compared to no intervention, placebo, or minimal intervention for knee osteoarthritis patients with limited exceptions. Outcomes of pain, physical function, physical performance, and stiffness were frequently associated with statistically significant improvements with exercise, though there were some inconsistencies in the evidence. One identified meta-analysis conducted an indirect comparison of exercise with opioid pain management for knee osteoarthritis patients and the results suggested comparable clinical efficacy. Overall the systematic reviews were conducted with few methodological concerns and the underlying evidence base included high quality randomized controlled trials. The evidence base was associated with potential for measurement and selection bias that limited the confidence in the conclusions of this report. Furthermore the variety of interventions, lengths of follow-up, and frequency or duration of exercise make it difficult to draw conclusions regarding the optimal exercise approach. The systematic reviews also identified under-reporting of adverse event outcomes as a limitation. One identified systematic review identified evidence that the most commonly reported adverse event was falling and that exercise was not associated with increased knee osteoarthritis progression.

Methods

Literature Search Methods

A limited literature search was conducted on key resources including PubMed, The Cochrane Library, University of York Centre for Reviews and Dissemination (CRD), Canadian and major international health technology agencies, as well as a focused Internet search. Methodological filters were applied to limit retrieval to health technology assessments (HTAs), systematic reviews (SRs) and meta-analyses (MAs). Where possible, retrieval was limited to the human population. The search was also limited to English language documents published between January 1, 2012 and July 27, 2017. Following a review of the quantity of available evidence identified in the first level of screening, articles published prior to 2015 were excluded.

Selection Criteria and Methods

One reviewer screened citations and selected studies. In the first level of screening, titles and abstracts were reviewed and potentially relevant articles were retrieved and assessed for inclusion. The final selection of full-text articles was based on the inclusion criteria presented in Table 1.

Table 1. Selection Criteria.

Table 1

Selection Criteria.

Exclusion Criteria

Articles were excluded if they did not meet the selection criteria outlined in Table 1, they were duplicate publications, or were published prior to 2015. Studies with a mixed population (e.g. hip and knee issues) were excluded unless results for knee OA were reported separately.

Critical Appraisal of Individual Studies

The included SRs and MAs were critically appraised using the Assessing the Methodological Quality of Systematic Reviews (AMSTAR) tool.6 Summary scores were not calculated for the included studies; rather, a review of the strengths and limitations of each included study were described narratively.

Summary of Evidence

Quantity of Research Available

A total of 243 citations were identified in the literature search. Following screening of titles and abstracts, 210 citations were excluded and 33 potentially relevant reports from the electronic search were retrieved for full-text review. Three potentially relevant publications were retrieved from the grey literature search. Of these potentially relevant articles, 18 publications were excluded for various reasons, while 18 publications met the inclusion criteria and were included in this report. Appendix 1 describes the PRISMA flowchart of the study selection.

Summary of Study Characteristics

Study Design

All 18 included reports consisted of a systematic review (SR) of the literature,1–5,7–19 and ten of these included reports also provided a meta-analysis of findings from the included studies.1,2,4,7–9,12,14–16 A critical appraisal of the identified evidence was presented in 14 of the included studies.1,2,4,7–9,12,14–20 Brosseau et al., published a three part SR in 2017 and all three parts of the SR are included in this report.3,10,11 The most detailed methodology for the three parts is provided in part one.10 Nine SRs examined randomized controlled trial (RCT) data only,2,8,9,12,14–17,19 the three part SR that only included RCTs with a Physiotherapy Evidence Database (PEDro) evidence score of 6 or more,3,10,11 one SR that included RCTs that had been previously identified and included in Cochrane reviews on the topic,4 one SR that included RCTs and case-control trials,1 one SRs that included RCTs and observational studies while excluding case reports,7 and three SRs that did not place any restrictions on study design in a literature selection criteria.5,13,18

Language

Two MAs included trials published in Chinese languages.8,12 Quicke et al., 2015 also identified and translated one RCT from Czech.18 The remainder of the studies limited the search for evidence to English language databases. None of the studies included in this report searched for evidence from, or excluded evidence from any specified countries.

Patient Population

Nine of the included SRs looked at patients with knee OA.1,3,10,11,13–16,19 Three reports specified unilateral or bilateral knee OA,8,9,12 and one report specified adult patients of 18 years or older.7 A broader patient population was defined in five SRs; one examined RCTs that enrolled adult patients of 18 years or older with hip or knee OA,2 one included RCTs that looked at trials of knee OA as well as trials enrolling a mixed population of hip and knee OA patients,4 one SR reviewed evidence on OA,5 on SR looked at adults with knee conditions,17 and one SR looked at studies that either enrolled patients with a mean age of 45 or more years with knee pain or enrolled adults with knee OA.18 These five SRs all provided analysis of evidence from patients with knee OA specifically.

Interventions and Comparators

The most recent SR from the Agency for Healthcare Research and Quality (AHRQ) was published in May 2017 and had a broad focus on any interventions for knee OA. The exercise related interventions identified in this SR consisted of interventions categorized as agility exercise, strength/resistance training, aerobic exercise, tai chi, and yoga. The comparators in this SR consisted of placebo, usual care (no definition provided), or waiting list.7 Sharma et al. also examined any intervention for OA that included exercise and did not define comparators.5 One SR examined RCTs that looked at muscle strength training and subdivided evidence into muscle strength training interventions that satisfied the American College of Sports Medicine (ACSM) guidelines for training and into interventions that did not satisfy ACSM guidelines. This SR included RCTs that employed comparators of no intervention, waiting list, sham, or placebo.9 One SR examined recreational activities, walking, and conditioning exercise while excluding aquatic activity and therapeutic exercise interventions and included RCTs that used comparators of usual care, minimal interventions such as advice or education, or no intervention.2 The three part SR from Brosseau et al. was divided into three parts defined by the investigated intervention for knee OA. The first part examined mind-body exercise programs such as tai chi and yoga,10 the second investigated the RCT evidence for strengthening exercise programs,11 while the third part examined aerobic exercise programs.3 All three included RCTs that compared these interventions to a non-exercise control.3,10,11 Kan et al. investigated trials that compared yoga to physiotherapy (specific interventions not provided), yoga with physiotherapy, usual care, no specific exercise, or ordinary daily activities.13 Chang et al. examined tai chi Chuan as compared to no intervention, education, or interview.14 Zhang et al. examined traditional Chinese exercise as compared to education or no treatment.8 General descriptions of included exercise interventions were exercise therapy,15,16,19 resistance exercise,12 home exercise programs,1 outpatient exercise and self-care/management.17 A SR focused on safety outcomes defined exercise interventions as a physical activity intervention or exposure to physical activity for a duration of 3 months or longer.18 One SR included RCTs that examined exercise compared to no intervention, sham, or standard care, and also examined RCTs that examined analgesics compared to placebo or standard care. This SR then used this data to indirectly compare the clinical efficacy of exercise and analgesics. A stratified analysis in this SR included a comparison of land and aquatic based exercise to opioids.4 All of the SRs included evidence of an exercise related intervention to a no intervention, sham, placebo or at least a minimal intervention such as education, standard care, or waiting list. The SR that focused on interventions of outpatient exercise and self-care/management identified evidence comparing exercise with self-care/management to no intervention. Although some self-care/management strategies were minimal interventions these comparisons were not considered relevant in this report. Evidence comparing exercise alone to no intervention from this SR is presented in this report.17

Outcomes

The most commonly investigated outcomes of the SRs included in this report are self-reported pain,1,3,4,7–14,17–19 subjective measures of physical function,1,3,7,8,10–14,17–19 and objective measures of physical function.1,3,7,8,10–15,17–19 Measures of quality of life (QoL) were reported in seven SRs,3,7,8,10,11,13,16 while joint stiffness was reported in four SRs.7,8,12,14 Outcomes related to mental health were reported in two SRs,7,8 as were measures of muscle strength.7,19 Adverse events were reported in three of the 15 SRs.7,8,18 The SR by Quicke et al. focused on safety related outcomes of exercise interventions including adverse events, and additionally reported on radiographic/MRI biomarkers of structural OA progression.18 Other unique outcomes reported in the identified SRs included incidence of surgery,7 health-related quality of life (HRQoL) SF-36 components,16 inflammation or effusion,7 walking performance outcomes,15 Survey of Activities and Fear of Falling in the Elderly (SAFE),14 external knee adduction moment,19 Stanford Self Efficacy Scale,17 and one SR did not specify outcomes of interest.5

A tabulated summary of study characteristics is provided in Appendix 2 of this report.

Summary of Critical Appraisal

The majority of the identified SRs were well conducted. The three part SR from Brosseau et al.10,11,18 and the SR from Sharma published in 2016 had the most significant limitations.5 These studies lacked sufficient description of the methodology.5,10,11,18 However, all of the included SRs provided a systematic literature search methodology,1–5,7–19 13 of which provided inclusion and exclusion criteria for literature selection1,2,5,7–9,12,13,15–19 while two SRs only described inclusion criteria.4,14 Other common strengths of the identified evidence included a description of the statistical methodology used1–4,7–12,14–16 a discussion on limitations,1–4,7–16,18,19 and tabulated study characteristics,1,2,4,7–9,12–19 Study selection was done in duplicate in 14 of the included SRs,1,3,7–12,14–19 and of the 12 studies that provided data extraction methodology,1,2,4,7–9,12,13,15,17–19 nine of those data extraction methodologies were conducted in duplicate or double checked by a second reviewer.1,2,7–9,12,13,17,19 A valid critical appraisal system was described and applied in 12 included SRs,4,7–9,12–19 nine of which conducted the critical appraisal in duplicate.7,9,12,13,15–19 Patient populations were well defined in 11 SRs,1,4,7–9,12–16,19 three SRs provided a well-defined intervention of interest,1,9,15 and three SRs examined more unique and well defined outcomes of interest.15,16,19 Ten of the SRs included an MA, and all of these provided an analysis of statistical heterogeneity,1,2,4,7–9,12,14–16 however only four of these also provided an assessment of publication bias.14–16,18 Findings of high statistical heterogeneity, reported as I2 values, were frequently observed for different outcomes and broad definitions of exercise interventions presented concerns of clinical heterogeneity. Two SRs discuss this limitation,15,16 one of which used an I2 value of over 25% as a cutoff to down grade the quality of the evidence of the MAs conducted.16 Two assessments of publication bias were not actually conducted due to an insufficient number of studies.16,18 One SR found evidence of publication bias and used this to downgrade the quality of the evidence associated with the MA.15 While none of the SRs stated a potential conflict of interest (COI), three SRs did not include a COI statement.16,17,19 A literature search of databases of both English and Chinese languages was conducted in two SRs.8,12 Evidence was limited to RCTs in 12 SRs, and while this did not contribute to the methodological quality of these SRs it does indicate that higher quality trial evidence meeting the selection criteria was available.2,3,8–12,14–17,19 Only one SR provided citations of studies that were excluded during literature selection.1 One SR (Sharma, 2016) did not provide a flowchart describing the literature selection process.5 A list of search terms used was not provided by ten SRs.2–4,7,9–13,15 An common important limitation of the identified evidence was a lack of reporting of adverse event data.1,3,4,10–12,15–17,19

A tabulated summary of this critical appraisal by individual SR is provided in Appendix 3 of this report.

Summary of Findings

What is the clinical effectiveness of exercise for the management of knee osteoarthritis?

The most recently published SR that was identified and addressed the research question was from the AHRQ.7 Investigated interventions categorized as aerobic exercise in this SR were identified in five RCTs. Statistically significant benefits in short-term Western Ontario and McMaster Arthritis Index (WOMAC) pain, short-term (4-12 weeks) WOMAC function, and short-term WOMAC overall scores were observed in one RCT that enrolled 27 patients. Mixed results from two small RCTs also lead the authors to conclude that there was insufficient evidence to support clinical efficacy of aerobic exercise on pain, function, and total WOMAC scores in the medium-term (12-26 weeks). Additionally, the authors concluded from two RCTs that reported no significant difference in outcomes of Knee injury and Osteoarthritis Outcome Score (KOOS) pain and WOMAC pain that there was insufficient evidence to draw conclusions regarding the long-term (over 26 weeks) use of aerobic exercise for knee osteoarthritis. The authors found low quality evidence from three RCTs that aerobic exercise demonstrated no significant long-term impact on function.7 This SR also categorized interventions as strength or resistance training. An MA of results from 5 RCTs indicated no significant effect on short-term WOMAC pain, results from 6 RCTs indicated no significant effect on short-term WOMAC function, and results from 3 RCTs indicated no significant effect on medium-term WOMAC function for strength or resistance training interventions. Interventions categorized as agility training demonstrated inconsistent effects. A low strength of evidence supported significant short-term benefit on pain but not function in 3 RCTs, while a low strength of evidence also supported no long-term benefit of agility training on pain or function. The SR from AHRQ also categorized exercise interventions as general exercise therapy. A low strength of evidence supported a beneficial effect on medium-term pain, medium-term function, and long-term pain for patients with knee OA. Three small short-term RCTs were identified in the SR and investigated tai chi with medium-term clinical benefits in WOMAC pain and function observed in two RCTs. This SR also examined evidence for the clinical effectiveness of yoga, however the authors did not draw any conclusions from the one small RCT identified.7

Zhang et al. identified eight RCTs with 325 patients that evaluated the clinical effectiveness of tai chi exercise for knee OA. Pooled results from these eight RCTs found a statistically significant benefit for short-term (8 to 24 weeks) pain and short-term function scores. Seven of the RCTs reported short-term stiffness scores and this pooled data demonstrated a statistically significant improvement compared to controls. Short-term QoL and short-term mental health scores were reported in three of the eight identified RCTs and statistically significant benefits in QoL but not mental health were identified in the MA.8

Chang et al. investigated tai chi Chuan exercise for knee OA and observed statistically significant benefits in an MA of 6 RCTs for WOMAC pain and WOMAC stiffness. Results of physical function outcomes were mixed as WOMAC physical function, and six-minute walk test outcomes were not significantly improved but the stair climb test and Survey of Activities and Fear of Falling in the Elderly (SAFE) were improved in a smaller number of RCTs. The authors also reported one trial that observed improved bone density in patients participating in tai chi Chuan. The authors did not comment on the certainty of conclusions. The average Jadad score of the evidence for each outcome was 4.0 except for the SAFE survey outcome for which the evidence was based on two trials with an average Jadad score of 3.0.14

Bartholdy et al. examined exercise interventions very generally and also subcategorized exercises that were focused on increasing muscle strength by fulfilling ACSM guideline criteria, and those that did not fulfill this criteria.9 The authors examined 22 RCTs that investigated ACSM exercises and 34 RCTs that investigated non-ACSM exercises for the management of knee OA. Knee extensor strength, pain outcomes, and physical function outcomes pooled from these RCTs demonstrated clinical effectiveness of exercise overall, ACSM exercise, and non-ACSM exercise as compared to control. An indirect comparison between ACSM exercises and non-ACSM exercises suggested that ACSM exercises did not provide superior outcomes for these patients.9

Fernandopulle et al. categorized exercise interventions of interest as recreation activity, conditioning exercise, or walking intervention. Pooled results from RCTs that examined knee OA patients without including hip OA patients, identified improved physical performance at six months and 18 months. Walking interventions demonstrated statistically significant improvement in physical function at six months in pooled results from 2 RCTs (n = 75), but not physical function at 12 months in pooled results from three RCTs (n = 156). Walking interventions also had no statistically significant benefit for pain at three months or six months, or physical performance at 12 months. The authors concluded that, based upon a moderate level of evidence, conditioning exercises had clinically significant results for patients with knee OA in the short and longer term, while physical activity interventions, such as tai chi and walking, had very limited evidence to support short-term clinical efficacy for pain or function outcomes. Half of the studies identified in this SR had a low risk of bias however items in the bias assessment were commonly absent in the identified evidence of the SR including blinding, compliance, and intention-to-treat analysis.2

The three part SR published in 2017 from Brosseau et al. examined many exercise interventions for knee OA classified as mind-body exercise programs,10 strengthening exercise programs,11 and aerobic exercises.3 Mixed results were reported for four small RCTs on Hatha yoga, tai chi Qigong, and Sun Style tai chi. The authors concluded that these interventions are to be recommended to achieve the statistically significant beneficial outcomes in the time periods observed in the trials. The authors explain the mixed results as due to the use of different outcome measures for a similar outcomes, and the effect of the follow-up time.10 A summary of identified RCTs on strengthening exercise programs were detailed in the second part of this SR.11 The authors identified 26 high-quality RCTs that found a total of 64 statistically significant benefits in outcomes of pain relief, improved physical function, and QoL of the different strengthening exercise programs examined in these RCTs. The number of outcomes that favoured control intervention or were not statistically significantly different was not reported. The authors conclude that any type of strengthening exercise among the included trials on land-based exercise was effective. Some conclusions from this SR were unclear or not specific.11 The third part of this SR investigated the potential of aerobic exercise in the management of knee OA and identified five high-quality trials. A total of ten statistically significant benefits in pain relief, improved physical function, and QoL were observed for aerobic exercise with or without a muscle strengthening exercise component in these five trials. The authors concluded that aerobic exercises, as a component of an exercise program that also includes muscle strengthening exercise, are recommended to provide clinical benefits in pain relief, improved physical function and QoL. There was insufficient evidence for the authors to strongly recommend aerobic exercises without a muscle strengthening exercise component as the evidence was limited to one trial.3

Li et al. examined the clinical efficacy of resistance exercise for the treatment of knee OA by examining trials that reported outcomes of pain, stiffness, and physical function. Resistance exercise interventions were further classified as high intensity, low intensity, short-term (12 weeks or less), and longer-term (more than 12 weeks). Stiffness outcomes did not demonstrate a statistically significant benefit for high intensity resistance exercise or for longer-term resistance exercise, however resistance exercise was effective for low intensity and short-term stiffness outcomes. Pain relief and physical function were significantly improved in all classifications of resistance exercise interventions. The effect size for pain and physical function benefits were observed to be greater in analyses of trials investigating high intensity as compared to low intensity resistance exercise. The authors identified a lack of blinding and intent-to-treat analysis as methodological concerns that may have impacted the small treatment effects observed in the included trials.12

Kan et al. examined the effects of yoga exercise on knee OA patients in outcomes of pain, physical function, and QoL. In addition to trials that employed an irrelevant comparator, this SR identified two single group before/after trials with Downs and Black’s Quality Index scores of 18/32 and 16/32. Both studies observed statistically significant improvements in outcomes of pain, and one study observed an improved QoL following 12 weeks of yoga. Physical function outcomes were mixed. The authors concluded that good evidence supported a positive clinical benefit of yoga for pain relief of knee OA and that yoga was a safe and tolerable exercise as no adverse events were reported in any of the included trials.13

The SR by Henriksen et al. provided the only identified comparison of exercise to opioids for the treatment of knee OA included in this report. This analysis compared exercise to opioid RCTs that utilized a similar no intervention control and therefore the comparison was indirect. Although trials that examined mixed hip and knee populations were included in the analysis the authors stratified these two populations and found no statistically significant difference between these groups for the primary comparison outcome. An MA of 34 RCTs that enrolled 4179 patients found a statistically significant improvement in pain with exercise as compared to control. A stratified analysis of land exercise and aquatic exercise also found statistically significant improvements in pain for land exercise but not aquatic exercise. An MA of 20 RCTs that enrolled 5627 patients found a statistically significant improvement in pain with analgesics. Further stratified indirect analyses demonstrated no statistically significant difference between land exercise as compared to opioids or aquatic exercise as compared to opioids for pain outcomes. The authors observed that overall the pharmacological trials were conducted with less potential for bias based upon more restrictive selection criteria, more frequent blinding, and more frequent reporting of the handling of missing data which may have influenced the author’s findings. Additionally, the authors point out that the majority of analgesic trials employed a placebo control whereas the exercise trials mainly used no intervention or a usual care control therefore the gross effect size of the analgesic trials may have been underestimated. The authors however concluded that the evidence suggested comparable effects of exercise and analgesics for pain management.4

Three studies were identified in the SR by Sharma in 2016. Short narrative summaries reported that one trial found that neuromuscular and quadriceps strengthening improved pain and function in knee OA patients without improvements in knee adduction moment. Another trial found pressure-pain sensitivity, temporal summation, and pain were reduced with exercise in knee OA patients. The third trial observed that booster sessions with a physical therapist did not improve pain, function, or compliance to home exercise in knee OA patients.5

Anwer et al. examined home exercise programs for knee OA patients in outcomes of pain and function. Statistically significant improvements in pain were observed in an MA of 11 trials and in physical function in an MA of 9 trials. The evidence base for these MAs identified by Anwer et al. was evaluated as having an average PEDro score of 6.78/10. The authors concluded that high methodological quality evidence supported the clinical efficacy of a variety of home exercise programs for pain and physical function outcomes.1

Tanaka et al. examined the clinical efficacy of exercise for knee OA patients in walking physical performance outcomes. Walking distance was examined in 11 RCTs with 1547 participants, graded as very-low quality, and a statistically significant benefit was identified. Walking time was examined in 12 RCTs with 901 participants and a statistically significant benefit was observed (evidence graded as moderate quality). Walking velocity was examined in 6 RCTs with 581 participants and MA identified a statistically significant benefit (evidence graded as low quality). The authors’ assigned grade of the supporting evidence quality was negatively impacted by their assessment of publication bias. The authors concluded that exercise therapy provided benefits in the amount of time spent walking, gait velocity, and possibly the total distance walked for knee OA patients.15

In 2015 Tanaka et al. looked at the impact of exercise on QoL outcomes for knee OA patients using domains of the SF-36 HRQoL questionnaire. MA of moderate and high quality evidence identified statistically significant improvements in the physical component summary, mental component summary, physical functioning, and role-physical domains of the SF-36. Similar analyses found no statistically significant effect of exercise on domains of bodily pain, general health, vitality, social functioning, role-emotional, and mental health in knee OA patients. The evaluated evidence base consisted of high or moderate quality evidence for these analyses and a minimum of four RCTs for each outcome, however the authors report uncertainty about the impact of high statistical heterogeneity and possible publication bias on the calculated effect sizes.16

Button et al. conducted an SR where findings of two previously published relevant SRs10,11 were briefly summarized in a table. At 12 and 18 month follow-ups a statistically significant improvement was demonstrated in physical functioning, pain index, and a standardized physical component for the exercise intervention as compared to a no exercise control. At 12 months follow-up no statistically significant improvement was found in the Stanford Self Efficacy Scale with exercise. The trials enrolled 222 patients and the consensus on the quality of the trials was rated good. Additional findings of this SR, where self-care/management with an exercise component was compared to a control, may be of additional interest. The primary focus of the SR was self-care/management interventions with an exercise component and the authors concluded that the evidence lacked sufficient quality to draw conclusions.17

Quicke et al. focused on the safety of physical activity for patients over 45 years old with knee pain or knee OA. Outcomes of structural OA biomarker imaging were specific to the knee OA patients. Quicke et al. identified five RCTs that measured changes in radiographic OA imaging, none of which observed evidence of significantly greater OA structural progression with exercise. The authors caution that this evidence was limited by quantity and an insufficient follow-up time biased to null safety outcomes. Across all identified studies which included older patients with knee pain and knee OA, the most common adverse event of moderate severity was falling. The authors cite evidence from a prior SR that exercise reduces the number of falls in community dwelling older adults and that the number of observed falls in the trials was comparatively low for the number of participants. Based upon analysis of pain outcomes in a mixed population of older adults with knee pain, the authors suggested that most knee OA patients will experience less pain with long-term exercise.18

Ferreira et al. identified three RCTs that investigated the effect of exercise therapy on knee adduction moment (KAM), WOMAC pain and WOMAC physical function in knee OA patients. No statistically significant effect of exercise on KAM was observed in the three identified RCTs. The first RCT, assessed as a high quality trial, observed a statistically significant benefit of hip abductor/adductor strengthening in WOMAC pain walking, WOMAC pain, and WOMAC physical function. A second RCT, assessed as low quality, did not observe a significant impact of hip abductor/adductor, knee extensor, and ankle plantar flexor strengthening exercises on WOMAC pain or WOMAC physical function. The third RCT was also assessed as low quality and found a significant improvement in WOMAC pain and WOMAC physical function in knee OA patients with a neutrally aligned knee who participated in quadriceps strengthening exercise. Knee OA patients with a malaligned knee did not experience a statistically significant benefit in these outcomes. The authors concluded, based upon a moderate level of evidence, that exercise is effective in reducing pain, improving physical capacity, and enhancing muscle strength, however mechanisms other than the reduction in KAM may be responsible for these benefits.19

The above evidence identified for this report examines exercise interventions for knee OA patients categorized here as aerobic exercise,3,7 strength or resistance training,5,7,9,11,12,19 agility training,7 general exercise therapy,2,7 tai chi1,2,7,8,10,14 yoga,7,11,13 walking interventions,2 land-exercise,4 aquatic exercise,4 any exercise,4,15,17,18 and a home exercise program.1 The two SRs examining aerobic exercise did not identify sufficient evidence to make conclusions regarding the efficacy of aerobic exercise alone in outcomes of physical function, pain, or QoL.3,7 Strength or resistance training did not have statistically significant benefits in two SRs that examined short-term benefits at 4-12 weeks and in less than six months,7,11 however other SRs identified evidence supporting strength or resistance training benefits for knee OA patients in outcomes of pain,5,9,12,19 function,9,12,19 strength,5,9 and stiffness over the short term and over longer strength or resistance training durations.9 One SR classified exercise interventions as agility training. This SR found limited evidence that agility training may have pain benefits within 4-12 weeks for knee OA patients but no significant benefits in pain or function over 26 weeks.7 Interventions classified here as general exercise therapy were examined in two SRs and demonstrated benefits in medium-term (12 to 26 weeks) pain,7 medium-term function,7 and long-term (greater than 26 weeks) pain outcomes,7 based upon low quality evidence, and also demonstrated benefits in physical function outcomes for knee OA patients in the short and in the longer term based upon moderate evidence.2 Exercises categorized as tai chi represent heterogeneous interventions of tai chi however all of the evidence identified in this report supports the clinical efficacy of tai chi for knee OA patients. No high quality supporting evidence was identified by the six SRs that examined tai chi interventions.1,2,7,8,10,14 Evidence from one RCT and two lower quality trials supported clinical benefits of yoga.7,11,13 Evidence was limited to support significant benefits on pain and physical function of walking interventions.2 While one SR found significant pain benefits for exercise categorized as land-based the same statistical significance was not observed for aquatic exercise.4 SRs that did not examine a specific exercise intervention identified significant benefits in pain,4 and physical performance outcomes,15 with a low incidence of adverse events and no evidence of structural progression of OA with exercise.18 These interventions represented a wide range of timing, exposure, and intensity, may have overlap with each other, and broad definitions may have concealed significant clinical effects.

Tabulated summaries of the findings of the included SRs and MAs are provided in Appendix 4 of this report.

Limitations

The findings identified in this report are limited by significant heterogeneity in the intervention of interest, a lack of blinding, a lack of quantitative measures of compliance, and potential publication bias. Categorization of exercise interventions were subjective, undefined, and/or lacked detailed description. Statistical heterogeneity (I2) was high for many analyses and limited the confidence in effect sizes. Another limitation identified by many SRs included blinding as a common methodological limitation in the included trials as patient exercise exposure is inherently difficult to blind. A lack of blinding in combination with the subjective patient reported outcomes increased the potential for measurement bias in exercise trials. Patient adherence to exercise interventions was not well reported in the included SRs although two SRs mentioned it without quantification.12,17 Two publications assessed publication bias and while Chang et al. did not find evidence of publication bias in 11 trials of tai chi,14 Tanaka et al. found evidence publication bias in 28 RCTs that examined broader exercise interventions that included tai chi.15 The effect of publication bias on the confidence of conclusions of this report is unknown.

Conclusions and Implications for Decision or Policy Making

Evidence from 18 SRs, ten of which provided an MA, met the literature selection criteria and supported the clinical efficacy of exercise compared to no intervention, placebo, or minimal intervention for knee OA patients with limited exceptions. Of the identified SRs, 15 had few methodological limitations, and the body of evidence identified in these SRs included high-quality RCTs. Pain,1–5,7–10,12–14,17,19 physical function,1–3,5,7–9,12–14,16,17,19 physical performance,13–15 and stiffness outcomes8,12,4 demonstrated statistically significant improvements with exercise in addition to benefits in less commonly reported outcomes. Evidence of exercise compared to opioids for knee OA pain management was identified in one SR. This indirect comparison MA suggested that exercise and opioid pain management efficacy are of a comparable magnitude.4 In addition to disease related outcomes, the identified evidence also supported QoL benefits of exercise in this population,8,16 including both physical and mental components of the SF-36.16 One SR found evidence that exercise was not associated with OA progression and a low incidence of the most commonly reported adverse event, falling.18 Chang et al. also identified evidence that exercise was associated with a decreased fear of falling in the elderly.14 Exercise represents an inexpensive, nonpharmacological, nonsurgical intervention providing beneficial effects for pain and physical function for knee OA patients. The safety and efficacy of very broadly defined exercise interventions accommodates patient choice in therapeutic exercise selection which may facilitate adherence,18 though optimal form of exercise or frequency and duration remain unclear. Adherence to exercise or physical activity is also an integral long-term management and prevention strategy for comorbidities such as cardiovascular disease, diabetes, and obesity.2

These conclusions are associated with some significant limitations of the evidence base that were reported in the SRs and included some common trial methodological concerns and heterogeneous interventions. The trials included in the SRs commonly lacked blinding, including trials that reported subjective patient reported outcomes.7 This limitation presented the potential for measurement bias. Intent-to-treat analysis was also commonly lacking in included trials, as was quantitative information on compliance.7,12 Since patients who did not experience benefit may have dropped out or have been less adherent to exercise interventions this presented the potential for selection bias.7 Intervention heterogeneity was common in the included SRs and broad categorization in many SRs may have concealed significant treatment effects.7 One SR also reported evidence of potential publication bias in the evidence base for general exercise therapy interventions. The impact of publication bias on the confidence of the conclusions of this report are unknown.15 The evidence is also limited by under-reporting of adverse events associated with exercise, which may be of particular concern in older knee OA patients.4,7

Future research recommendations in the identified studies included examination of more specific subpopulations of knee OA patients. One identified SR found evidence that knee OA patients with a malaligned knee, did not achieve the statistically significant pain or physical function outcome benefits that were observed in knee OA patients with an aligned knee.11 Large RCTs with well-defined interventions (intensity, frequency etc.), information on adherence to exercise programs, and adverse event reporting may also help to address some of the limitations of the evidence.7 Comparative efficacy of different exercise interventions is beyond the scope of this report. A recent SR from 2014, cited by Henriksen et al.,4 concluded that, based on the available evidence at the time, it was not possible to determine the superiority of one exercise intervention over another. An updated review of comparative evidence on exercise interventions may also suggest opportunities to further improve outcomes for knee OA patients.

References

1.
Anwer S, Alghadir A, Brismee JM. Effect of home exercise program in patients with knee osteoarthritis: a systematic review and meta-analysis. J Geriatr Phys Ther. 2016 Jan;39(1):38–48. [PubMed: 25695471]
2.
Fernandopulle S, Perry M, Manlapaz D, Jayakaran P. Effect of land-based generic physical activity interventions on pain, physical function, and physical performance in hip and knee osteoarthritis: a systematic review and meta-analysis. Am J Phys Med Rehabil. 2017 Mar 18. [PubMed: 28323761]
3.
Brosseau L, Taki J, Desjardins B, Thevenot O, Fransen M, Wells GA, et al. The Ottawa panel clinical practice guidelines for the management of knee osteoarthritis. Part three: aerobic exercise programs. Clin Rehabil. 2017 May;31(5):612–24. [PubMed: 28183194]
4.
Henriksen M, Hansen JB, Klokker L, Bliddal H, Christensen R. Comparable effects of exercise and analgesics for pain secondary to knee osteoarthritis: a meta-analysis of trials included in Cochrane systematic reviews. J Comp Eff Res. 2016 Jul;5(4):417–31. [PubMed: 27346368]
5.
Sharma L. Osteoarthritis year in review 2015: clinical. Osteoarthritis Cartilage [Internet]. 2016 Jan [cited 2017 Aug 2];24(1):36–48. Available from: http://www​.ncbi.nlm.nih​.gov/pmc/articles/PMC4693145 [PMC free article: PMC4693145] [PubMed: 26707991]
6.
Shea BJ, Grimshaw JM, Wells GA, Boers M, Andersson N, Hamel C, et al. Development of AMSTAR: a measurement tool to assess the methodological quality of systematic reviews. BMC Med Res Methodol [Internet]. 2007 [cited 2017 Aug 2];7:10. Available from: http://www​.ncbi.nlm.nih​.gov/pmc/articles​/PMC1810543/pdf/1471-2288-7-10.pdf [PMC free article: PMC1810543] [PubMed: 17302989]
7.
Treatment of osteoarthritis of the knee: an update review [Internet]. Rockville (MD): Agency for Healthcare Research and Quality; 2017. [cited 2017 Aug 4]. (Comparative effectiveness review, no. 190). Available from: https://www​.effectivehealthcare​.ahrq.gov​/ehc/products/633/2441​/osteoarthritis-knee-update-report-170505.pdf
8.
Zhang Y, Huang L, Su Y, Zhan Z, Li Y, Lai X. The effects of traditional Chinese exercise in treating knee osteoarthritis: a systematic review and meta-analysis. PLoS One [Internet]. 2017 [cited 2017 Aug 2];12(1):e0170237. Available from: http://www​.ncbi.nlm.nih​.gov/pmc/articles/PMC5266306 [PMC free article: PMC5266306] [PubMed: 28121996]
9.
Bartholdy C, Juhl C, Christensen R, Lund H, Zhang W, Henriksen M. The role of muscle strengthening in exercise therapy for knee osteoarthritis: A systematic review and meta-regression analysis of randomized trials. Semin Arthritis Rheum. 2017 Mar 18. [PubMed: 28438380]
10.
Brosseau L, Taki J, Desjardins B, Thevenot O, Fransen M, Wells GA, et al. The Ottawa panel clinical practice guidelines for the management of knee osteoarthritis. Part one: introduction, and mind-body exercise programs. Clin Rehabil. 2017 May;31(5):582–95. [PubMed: 28183188]
11.
Brosseau L, Taki J, Desjardins B, Thevenot O, Fransen M, Wells GA, et al. The Ottawa panel clinical practice guidelines for the management of knee osteoarthritis. Part two: strengthening exercise programs. Clin Rehabil. 2017 May;31(5):596–611. [PubMed: 28183213]
12.
Li Y, Su Y, Chen S, Zhang Y, Zhang Z, Liu C, et al. The effects of resistance exercise in patients with knee osteoarthritis: a systematic review and meta-analysis. Clin Rehabil. 2016 Oct;30(10):947–59. [PubMed: 26471972]
13.
Kan L, Zhang J, Yang Y, Wang P. The effects of yoga on pain, mobility, and quality of life in patients with knee osteoarthritis: a systematic review. Evid Based Complement Alternat Med [Internet]. 2016 [cited 2017 Aug 2];2016:6016532. Available from: http://www​.ncbi.nlm.nih​.gov/pmc/articles/PMC5061981 [PMC free article: PMC5061981] [PubMed: 27777597]
14.
Chang WD, Chen S, Lee CL, Lin HY, Lai PT. The effects of Tai Chi Chuan on improving mind-body health for knee osteoarthritis patients: a systematic review and meta-analysis. Evid Based Complement Alternat Med [Internet]. 2016 [cited 2017 Aug 2];2016:1813979. Available from: http://www​.ncbi.nlm.nih​.gov/pmc/articles/PMC5011213 [PMC free article: PMC5011213] [PubMed: 27635148]
15.
Tanaka R, Ozawa J, Kito N, Moriyama H. Effects of exercise therapy on walking ability in individuals with knee osteoarthritis: a systematic review and meta-analysis of randomised controlled trials. Clin Rehabil. 2016 Jan;30(1):36–52. [PubMed: 25691583]
16.
Tanaka R, Ozawa J, Kito N, Moriyama H. Does exercise therapy improve the health-related quality of life of people with knee osteoarthritis? A systematic review and meta-analysis of randomized controlled trials. J Phys Ther Sci [Internet]. 2015 Oct [cited 2017 Aug 2];27(10):3309–14. Available from: http://www​.ncbi.nlm.nih​.gov/pmc/articles/PMC4668190 [PMC free article: PMC4668190] [PubMed: 26644699]
17.
Button K, Roos PE, Spasic I, Adamson P, van Deursen RW. The clinical effectiveness of self-care interventions with an exercise component to manage knee conditions: a systematic review. Knee [Internet]. 2015 Oct [cited 2017 Aug 2];22(5):360–71. Available from: http://www​.ncbi.nlm.nih​.gov/pmc/articles/PMC4642743 [PMC free article: PMC4642743] [PubMed: 26056046]
18.
Quicke JG, Foster NE, Thomas MJ, Holden MA. Is long-term physical activity safe for older adults with knee pain?: A systematic review. Osteoarthritis Cartilage [Internet]. 2015 Sep [cited 2017 Aug 2];23(9):1445–56. Available from: http://ac​.els-cdn.com​/S1063458415011590/1-s2​.0-S1063458415011590-main​.pdf?_tid=c29f6568-77ab-11e7-b348-00000aacb360&acdnat​=1501696705​_1023d8e2c03a468aa1e1e5828cc553ee [PubMed: 26003947]
19.
Ferreira GE, Robinson CC, Wiebusch M, Viero CC, da Rosa LH, Silva MF. The effect of exercise therapy on knee adduction moment in individuals with knee osteoarthritis: a systematic review. Clin Biomech (Bristol, Avon ). 2015 Jul;30(6):521–7. [PubMed: 25896448]
20.
Li X, Wang XQ, Chen BL, Huang LY, Liu Y. Whole-body vibration exercise for knee osteoarthritis: a systematic review and meta-analysis. Evid Based Complement Alternat Med [Internet]. 2015 [cited 2017 Aug 2];2015:758147. Available from: http://www​.ncbi.nlm.nih​.gov/pmc/articles/PMC4540999 [PMC free article: PMC4540999] [PubMed: 26347287]

Appendix 1. Selection of Included Studies

Image app1f1

Appendix 2. Characteristics of Included Publications

Table 2Characteristics of Included Publications

Author,
Publication
Date		Study DesignPopulationInterventionComparator(s)Outcomes
AHRQ 20177MA and CA of RCTs and observational studiesAdults (≥18years) with knee OAAny. Exercise related: Agility exercise, strength/resistance training, aerobic exercise, physical therapy/general exercise programs, manual therapy, and whole-body vibrationPlacebo, usual care, or waiting list
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • Joint stiffness
  • Mental health and safety
  • Inflammation or effusion
  • Medication use
  • QoL
  • Surgery
  • Adverse events
Zhang et al., 20178MA and CA of RCTsUnilateral or bilateral knee OA

Exclusions: knee trauma or surgery, history of rheumatoid arthritis, or cognitive impairment		Traditional Chinese exerciseEducation or no treatment
  • Self-reported pain
  • Physical function(subjective and objective scales)
  • Joint stiffness
  • Mental health
  • QoL
  • Adverse events
Bartholdy et al., 20179MA and CA of RCTsUnilateral or bilateral knee OAMuscle strength training as defined by ACSM as well as non-ACSM exercise interventionsNo intervention,waiting list, sham, or placebo
  • Self-reported pain
  • Physical function
  • Knee extensor strength
Fernandopulle et al., 20172MA and CA of RCTs with at least 10 patientsAdults (≥18years) with hip or knee OARecreational activities, walking, and conditioning exercise

Aquatic activity and therapeutic exercise interventions were excluded		Usual care, minimal intervention (advice, education ect.), or no intervention
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • Physical performance
Brosseau et al., 2017(1)10SR of high-quality RCTs (PEDro score≥6)Knee OAMind-body exercise programs (tai chi, yoga)Non-exercise group
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • QoL
Brosseau et al., 2017(2)11SR of high-quality RCTs (PEDro score≥6)Knee OAStrengthening exercise programsNon-exercise group
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • QoL
Brosseau et al., 2017(3)3SR of high-quality RCTs (PEDro score≥6)Knee OAAerobic exercise programsNon-exercise group
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • QoL
Li et al., 201612MA and CA of RCTsUnilateral or bilateral knee OAResistance exercise

Mixed exercise interventions were exclusded		No treatment or educational interventions
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • Joint stiffness
Kan et al., 201613SR with CAKnee OAYogaPhysiotherapy, Yoga with physiotherapy, usual care, no specific exercise, ordinary daily activities
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • QoL
Chang et al., 201614MA and CA of RCTsKnee OAtai chi ChuanNo intervention, education, or interview
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • Joint stiffness
  • SAFE
Henriksen et al., 20164MA and CA of RCTs identified by Cochrane reviewsKnee OA and mixed knee and hip OAExercise or oral analgesicsFor exercise studies: no intervention, sham, or standard care

For oral analgesics: placebo or standard care
  • Self-reported pain
Sharma, 20165SROAAny including exercise and physical therapyAnyAny
Anwer et al., 20161MA and CA of RCTs and case-controlled trialsKnee OAHome exercise programsPhysical therapy or no intervention
  • Self-reported pain
  • Physical function(subjective and objective scales)
Tanaka et al., 201615MA and CA of RCTsKnee OAExercise therapyNo intervention or psychoeducational intervention
  • Physical function (objective scales), specifically: total distance walked, amount of time spent walking, and gait velocity
Tanaka et al., 201516MA and CA of RCTsKnee OAExercise or exercise therapyNo intervention or psychoeducational intervention
  • HRQoL using SF-36
Button et al., 201517SR and CA of RCTsAdults (≥18years) with knee conditionsOutpatient exercise and self-care/managementAny
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • Stanford Self Efficacy Scale
Quicke et al., 201518SR and CA of RCTs, prospective cohort, and case control studiesOlder adults (mean age ≥45years) with knee pain OR adults with knee OAPhysical activity intervention or exposure of ≥3month durationNon-physical activity interventionSafety related outcomes from:
  • Adverse events
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • Radiographic/MRI biomarkers of structural OA progression
Ferreira et al., 201519SR and CA of RCTsKnee OAExercise therapyControl or sham
  • Self-reported pain
  • Physical function (subjective and objective scales)
  • Muscle strength
  • External knee adduction moment

ACSM = American College of Sports Medicine; AHRQ = Agency for Healthcare Research and Quality; CA = critical appraisal; COPD = chronic obstructive pulmonary disease; HRQoL = health-related quality of life; MA = meta-analysis; NR = not reported; OA = osteoarthritis; PEDro = Physiotherapy Evidence Database; QoL = quality of life; SR = systematic review; RCT = randomized controlled trial; WBV = whole-body vibrarion

Appendix 3. Critical Appraisal of Included Publications

Table 3Strengths and Limitations of Systematic Reviews and Meta-Analyses using AMSTAR6

StrengthsLimitations
AHRQ7
  • Provided meta-analysis
  • Systematic literature search methodology described with inclusion/exclusion criteria
  • Flowchart of study selection provided
  • Valid critical appraisal system described and applied
  • Statistical methodology described
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Defined patient population
  • Tabulated study characteristics
  • Study selection and data extraction done in duplicate
  • Reported adverse event data
  • Statement of no COIs (funding AHRQ)
  • Very broad focus on interventions
  • Search terms of literature search not provided
  • No assessment of publication bias
  • No list of excluded studies
Zhang et al., 20178
  • Provided meta-analysis
  • Statistical methodology described
  • Data from RCTs only
  • Systematic literature search methodology described with inclusion/exclusion criteria with search terms
  • English and Chinese language studies included
  • Flowchart of study selection provided
  • Valid critical appraisal system described and applied
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Defined patient population
  • Tabulated study characteristics
  • Reported adverse event data
  • Study selection and data extraction done in duplicate
  • Statement of no COIs (Grants from the State Administration of Traditional Chinese Medicine)
  • No assessment of publication bias
  • No list of excluded studies
Bartholdy et al., 20179
  • Provided meta-analysis
  • Statistical methodology described
  • Data from RCTs only
  • Systematic literature search methodology described with inclusion/exclusion criteria
  • Flowchart of study selection provided
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Valid critical appraisal system described and applied
  • Study selection, data extraction and critical appraisal done in duplicate
  • Tabulated study characteristics
  • Well defined intervention
  • Defined patient population
  • Statement of no COIs (funding from Denmark and The Oak Foundation)
  • Literature search terms provided in supplementary material
  • Risk of bias assessments provided in supplementary material
  • No assessment of publication bias
  • No list of excluded studies
Fernandopulle et al., 20172
  • Provided meta-analysis
  • Statistical methodology described
  • Data from RCTs only
  • Systematic literature search methodology described with inclusion/exclusion criteria
  • Flowchart of study selection provided
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Data extraction methodology provided and was verified by second reviewer
  • Valid critical appraisal system described and applied
  • Study selection and critical appraisal done in duplicate
  • Tabulated study characteristics
  • Statement of no COIs (funding from School of Physiotherapy Research Support Fund)
  • Exact literature search terms not provided
  • Interventions broadly catagorized
  • No assessment of publication bias
  • Includes studies on hip OA
  • No list of excluded studies
Brosseau et al., 2017(1)10
  • Statistical methodology described
  • Data from RCTs with PEDro score ≥ 6 only
  • Systematic literature search methodology described
  • Flowchart of study selection provided
  • Study selection done in duplicate
  • Discussion on study limitations
  • Statement of no COIs (funding from University of Ottawa Research Chair)
  • Literature search terms provided in supplementary material
  • No defined literature inclusion/exclusion criteria
  • No critical appraisal
  • Lacked detailed methodology
  • No tabulated study characteristics
  • No assessment of publication bias
  • No adverse event data
  • No list of excluded studies
Brosseau et al., 2017(2)11
  • Statistical methodology described10
  • Data from RCTs with PEDro score ≥ 6 only
  • Systematic literature search methodology described10
  • Flowchart of study selection provided10
  • Study selection done in duplicate
  • Discussion on study limitations
  • Statement of no COIs (funding from University of Ottawa Research Chair)
  • Literature search terms provided in supplementary material
  • No defined literature inclusion/exclusion criteria
  • No critical appraisal
  • Lacked detailed methodology
  • No tabulated study characteristics
  • No assessment of publication bias
  • No adverse event data
  • No list of excluded studies
Brosseau et al., 2017(3)3
  • Statistical methodology described10
  • Data from RCTs with PEDro score ≥ 6 only
  • Systematic literature search methodology described10
  • Flowchart of study selection provided10
  • Study selection done in duplicate
  • Discussion on study limitations
  • Statement of no COIs (funding from University of Ottawa Research Chair)
  • Literature search terms provided in supplementary material
  • No defined literature inclusion/exclusion criteria
  • No critical appraisal
  • Lacked detailed methodology
  • No tabulated study characteristics
  • No assessment of publication bias
  • No adverse event data
  • No list of excluded studies
Li et al., 201612
  • Provided meta-analysis
  • Statistical methodology described
  • Data from RCTs only
  • Systematic literature search methodology described with inclusion/exclusion criteria
  • English and Chinese language studies included
  • Flowchart of study selection provided
  • Valid critical appraisal system described and applied in duplicate
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Defined patient population
  • Tabulated study characteristics
  • Study selection and data extraction done in duplicate
  • Statement of no COIs (Grants from the State Administration of Traditional Chinese Medicine)
  • Literature search terms provided in supplementary material
  • No adverse event data
  • No assessment of publication bias
  • No list of excluded studies
Kan et al., 201613
  • Valid critical appraisal system described and applied in duplicate
  • Systematic literature search methodology described with inclusion/exclusion criteria
  • Flowchart of study selection provided
  • Discussion on study limitations
  • Defined patient population
  • Tabulated study characteristics
  • Data extraction done in duplicate
  • Statement of no COIs (funding not reported)
  • Reported adverse event data
  • Exact literature search terms not provided
  • No assessment of publication bias
  • No list of excluded studies
Chang et al., 201614
  • Provided meta-analysis
  • Statistical methodology described
  • Data from RCTs only
  • Assessment of publication bias
  • Systematic literature search methodology described with inclusion criteria and search terms
  • Flowchart of study selection provided
  • Valid critical appraisal system described and applied
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Defined patient population
  • Tabulated study characteristics
  • Study selection done in duplicate
  • Statement of no COIs (funding from China Medical University)
  • Reported adverse event data
  • Searched only English language databases
  • Data extraction not done in duplicate
  • No list of excluded studies
Henriksen et al., 20164
  • Provided meta-analysis
  • Statistical methodology described
  • Systematic literature search methodology described with inclusion criteria
  • Flowchart of study selection provided
  • Valid critical appraisal system described and applied
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Tabulated study characteristics
  • Study selection and data extraction done in duplicate
  • Statement of no COIs (funding from The Oak Foundation and undisclosed)
  • Conclusions based upon indirect evidence
  • Restricted to pain outcomes
  • Mixed patient population however provided separate analysis for knee OA
  • Exact literature search terms not provided
  • No assessment of publication bias
  • Limited to RCTs identified by Cochrane Reviews
  • No adverse event data
  • No list of excluded studies
Sharma, 20165
  • Systematic literature search methodology described with inclusion/exclusion criteria and search terms
  • Statement of no COIs (funding NIH/NIAMS)
  • No flowchart of study selection provided
  • No assessment of study quality
  • Lacked detailed methodology
  • No discussion of study limitations
  • Very broad focus on patient population, interventions, comparators, and outcomes
  • No tabulated study characteristics
  • No list of excluded studies
Anwer et al., 20161
  • Provided meta-analysis
  • Statistical methodology described
  • Systematic literature search methodology described with inclusion/exclusion criteria and search terms
  • Flowchart of study selection provided
  • Valid critical appraisal system described and applied in duplicate
  • Citations for excluded studies
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Defined patient population
  • Well defined intervention
  • Tabulated study characteristics
  • Study selection and data extraction done in duplicate
  • Statement of no COIs (funding Deanship of Scientific Research at King Saud University)
  • No assessment of publication bias
  • No reporting of adverse event data
Tanaka et al., 201615
  • Provided meta-analysis
  • Statistical methodology described (some descriptions in separate source)
  • Data from RCTs only
  • Assessment of publication bias (used for critical appraisal of evidence body used in meta-analyses)
  • Systematic literature search methodology described with inclusion/exclusion criteria
  • Flowchart of study selection provided
  • Valid critical appraisal system described and applied in duplicate
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Data extraction methodology provided
  • Defined patient population
  • Well defined intervention and outcomes
  • Tabulated study characteristics
  • Study selection done in duplicate
  • Statement of no COIs (funding Japanese Physical Therapy Association)
  • Exact literature search terms not provided
  • No reporting of adverse event data
  • No list of excluded studies
Tanaka et al., 201516
  • Provided meta-analysis
  • Statistical methodology described
  • Data from RCTs only
  • Assessment of publication bias (too few studies to conduct)
  • Systematic literature search methodology described with inclusion/exclusion criteria and search terms
  • Flowchart of study selection provided
  • Valid critical appraisal system described and applied in duplicate
  • Assessed statistical heterogeneity
  • Discussion on study limitations
  • Defined patient population
  • Well defined outcomes
  • Tabulated study characteristics
  • Study selection done in duplicate
  • No COI statement
  • No reporting of adverse event data
  • No data extraction methodology
  • No list of excluded studies
Button et al., 201517
  • Inclusion of RCTs only
  • Systematic literature search methodology described with inclusion/exclusion criteria and search terms
  • Study selection done in duplicate
  • Flowchart of study selection provided
  • Data extraction methodology provided and was verified by second reviewer
  • Valid critical appraisal system described and applied in duplicate
  • Tabulated study characteristics
  • No explicit discussion on study limitations
  • Broad patient population
  • No reporting of adverse event data
  • No COI statement
  • No list of excluded studies
Quicke et al., 201518
  • Assessment of publication bias (too few studies to conduct)
  • Systematic literature search methodology described with inclusion/exclusion criteria and search terms
  • Flowchart of study selection provided
  • Data extraction methodology described
  • Valid critical appraisal system described and applied in duplicate
  • Discussion on study limitations
  • Study selection done in duplicate
  • Statement of no COI (funding Arthritis Research UK, NIHR)
  • Mixed population for most outcomes
  • No list of excluded studies
Ferreira et al., 201519
  • Data from RCTs only
  • Systematic literature search methodology described with inclusion/exclusion criteria and search terms
  • Data extraction done in duplicate
  • Flowchart of study selection provided
  • Valid critical appraisal system described and applied in duplicate
  • Discussion on study limitations
  • Defined patient population
  • Well defined outcomes
  • Tabulated study characteristics
  • Study selection done in duplicate
  • No reporting of adverse event data
  • No COI statement
  • No list of excluded studies

AHRQ = Agency for Healthcare Research and Quality; COI = conflict of interest; NIH = National Institutes of Health; OA = osteroarthritis; PEDro = Physiotherapy Evidence Database; RCT = randomized controlled trial;

Appendix 4. Main Study Findings and Author’s Conclusions

Table 4Summary of Relevant Findings of Included Studies

Main Study FindingsAuthor’s Conclusion
AHRQ7
Aerobic Exercise (MD < 1 favours exercise)Aerobic Exercise (pg 24)
WOMAC pain scores at 6 weeks (1 RCT, n = 27) (P < 0.05)
MD (95% CI): -4.02 (-6.01, -2.03)
WOMAC function scores at 6 weeks (1 RCT, n = 27) (P < 0.05)
MD (95% CI): -15.35 (-24.02, -6.68)
WOMAC total scores at 6 weeks (1 RCT, n = 27) (P < 0.05)
MD (95% CI): -18.58 (-29.65, -7.51)
WOMAC pain scores at 12 weeks (1 RCT, n = 37) (P < 0.05)
MD (95% CI): -14.9 (-27.0, -2.6)
WOMAC function scores at 12 weeks (1 RCT, n = 37) (P > 0.05)
MD (95% CI): NS
WOMAC function scores at 12 weeks (1 RCT, n = 26) (P > 0.05)
MD (95% CI): NS
WOMAC stiffness scores at 12 weeks (1 RCT, n = 37) (P < 0.05)
MD (95% CI): -10.8 (-21.3, -0.7)
WOMAC total or KOOS ADL scores at 12 weeks (1 RCT, n = 37) (P > 0.05)
MD (95% CI): NS
WOMAC pain scores at 18 months (1 RCT, n = 222) (P > 0.05)
MD (95% CI): NS
KOOs pain scores at 12 months (1 RCT, n = 80) (P > 0.05)
MD (95% CI): NS
Functional outcomes at 12 months (2 RCTs) (P > 0.05)
NS
Functional outcomes at 18 months (1 RCT) (P > 0.05)
NS
  • “Evidence was insufficient to draw conclusions about short-term effects of aerobic exercise on pain, function, and total WOMAC scores (one RCT).”
  • “Evidence was insufficient to draw conclusions about medium-term effects of aerobic exercise on pain, function, and total WOMAC scores (two RCTs).”
  • “Evidence was insufficient to draw conclusions on effects of long-term aerobic exercise on pain (2 RCTs).”
  • “Aerobic exercise showed no significant long-term effects on function, based on three RCTs (low evidence).”
Strength or Resistance Training (SMD < 1 favours exercise)Strength or Resistance Training (pg 26)
WOMAC pain scores short-term (5 RCTs, n = 215, I2 = 72.3%) (P = NS)
SMD (95% CI): -0.55 (-1.46, 0.37)
WOMAC function scores short-term (6 RCTs, n = 245, I2 = 68.4%) (P = NS)
SMD (95% CI): -0.60 (-1.38, 0.17)
WOMAC function scores medium-term (3 RCTs, n = 187, I2 = 69.3%) (P = NS)
SMD (95% CI): -0.43 (-2.16, 1.30)
  • “It is unclear whether strength and resistance training have a beneficial effect on patients with OA of the knee. Pooled analyses support a nonstatistically significant benefit, and individual study findings suggest possible benefit on pain and function and significant benefit on total WOMAC scores.”
  • “Strength and resistance training had no statistically significant beneficial effect on short-term pain or function based on pooled analyses of 5 RCTs but a significant short-term beneficial effect on the composite WOMAC total score based on 3 RCTs (low strength of evidence).”
  • “Strength and resistance training showed a nonsignificant medium-term beneficial effect on function in a pooled analysis of 3 RCTs (low strength of evidence).”
  • “Evidence was insufficient to assess long-term effects of strength and resistance training.”
  • “No studies assessed the effects of any factors such as sex, obesity, or disease severity on outcomes of strength and resistance training.”
Agility Training (MD < 1 favours exercise)Agility Training (pg 31)
WOMAC pain scores at 8 weeks (1 RCT, n = 44) (P < 0.05)
MD (95% CI): -3.13 (-5.86, -0.40)
NRS-measured pain at 6 weeks (1 RCT, n = 159) (P > 0.05)
MD (95% CI): NS
WOMAC pain scores at 8 weeks (1 RCT, n = 14) (P < 0.05)
MD (1-10 scale) (95% CI): -4.00 (-5.32, -2.68)
  • “It is unclear whether agility training alone has any benefit for patients with knee OA. Identified studies showed inconsistent effects across time points and outcomes.”
  • “Agility training showed significant short-term beneficial effects on pain but not on function in 3 RCTs (low strength of evidence).
  • “Agility training showed no consistent beneficial effects on medium-term pain or function.”
  • “Agility training showed no long-term beneficial effect on pain (3 RCTs) or function (2 RCTs) (low strength of evidence).”
General Exercise Therapy (MD < 1 favours exercise)General Exercise Therapy (pg. 34)
KOOS pain scores at 3 months (1 RCT, n = 180) (P < 0.05)
MD (0-100) (95% CI): -10.00 (-15.28, -4.72)
WOMAC pain scores at 20 weeks (1 RCT, n = 126) (P < 0.05)
MD (0-17) (95% CI): -1.90 (-3.28, -0.52)
KOOS function scores at 3 months (1 RCT, n = 180) (P < 0.05)
MD (0-100) (95% CI): -9.00 (-14.28, -3.72)
WOMAC function scores at 20 weeks (1 RCT, n = 126) (P < 0.05)
MD (0-68) (95% CI): -5.10 (-9.81, -0.39)
VAS pain scores at 1 year (1 RCT, n = 75) (P < 0.05)
MD (0-10) (95% CI): -2.00 (--3.84, -0.16)
VAS pain scores at 1 year (1 RCT, n = 300) (P < 0.05)
MD (0-10) (95% CI): -0.60 (-.78, -0.42)
VAS pain scores at 1 year (1 RCT, n = 192) (P > 0.05)
MD (0-10) (95% CI): NS
WOMAC function scores at 32 weeks (1 RCT, n = 126) (P < 0.05)
MD (0-17) (95% CI): -2.00 (--3.37, -0.63)		“General exercise programs appear to have beneficial medium-term effects on pain and function and long-term effects on pain for patients with knee OA, based on a relatively small number of heterogeneous RCTs.”
  • “Evidence was insufficient to assess the effects of general exercise therapy programs on short-term pain or function.
  • “General exercise therapy programs had a beneficial effect on medium term pain and function, based on two RCTs (low strength of evidence).”
  • “General exercise therapy programs showed beneficial long-term effects on pain, based on 4 RCTs (low strength of evidence), but evidence was insufficient to assess long-term effects on function or quality of life.”
tai chi (MD < 1 favours exercise)tai chi (pg. 36)
WOMAC function scores at 9 weeks (1 RCT, n = 20) (P < 0.05)
MD (95% CI): -5.54 (-9.72, -1.36)
WOMAC pain scores at 21 weeks (1 RCT, n = 20) (P < 0.05)
MD (95% CI): -1.58 (-2.76, -0.40)
WOMAC function scores at 21 weeks (1 RCT, n = 20) (P < 0.05)
MD (95% CI): -5.52 (-9.70, -1.34)
WOMAC function scores at 24 weeks (1 RCT, n = 204) (P < 0.05)
MD (1-1700) (95% CI): -131.10 (-251.36, -10.85)		“Tai chi appears to have some short- and medium-term benefit for patients with OA of the knee, based on three small, short-term RCTs and one larger, 18-week RCT (total n=290).”
  • “Tai chi showed significant beneficial short-term effects on pain, comparable with those of conventional physical therapy, in one large RCT, but no significant effects in two small, brief RCTs (low strength of evidence).”
  • “Tai chi showed significant benefit for medium-term pain and function in 2 RCTs (low strength of evidence).”
  • “Evidence was insufficient to assess long-term effects of Tai chi on pain, function, and other outcomes.”
Yoga(pg.38)
  • It is unclear whether yoga has any benefit for patients with OA of the knee, as we identified only one small RCT (n=36).
Zhang et al., 20178
tai chi (MD < 1 favours exercise)
Pain scores short-term (8 RCTs, n = 325, I2 = 54%) (P < 0.0001)
SMD (95% CI): -0.77 (-1.13, -0.41)
Function scores short-term (8 RCTs, n = 325, I2 = 37%) (P < 0.0001)
SMD (95% CI): -0.75 (-0.98, -0.52)
Stiffness scores short-term (7 RCTs, n = 228, I2 = 51%) (P = 0.006)
SMD (95% CI): -0.56 (-0.96, -0.16)
QoL scores short-term (3 RCTs, n = 95, I2 = 0%) (P = 0.005)
SMD (95% CI): 0.57 (0.17, 0.97)
Mental health scores short-term (3 RCTs, n = 95, I2 = 52%) (P = 0.08)
SMD (95% CI): 4.12 (-0.50, 8.73)

The meta-analysis was done using two studies that presented a high risk of bias and six studies that presented a low risk according to the Cochrane Collaboration.		“Our systematic review revealed that short-term TCE was potentially beneficial in terms of reducing pain, improving physical function and alleviating stiffness. These results may suggest that TCE could prove useful as an adjuvant treatment for patients with knee OA.” (pp. 2/17)

“However, in terms of the small number of studies, and the significance of the two outcomes [QoL and mental health] may need a long time to observe due to the TCEs may play an indirect role in improving the quality of life and mental health by regulating the whole body function, We considered there was little significance on the results of these two outcomes.” (pp. 9/17)
Bartholdy et al., 20179
Overall Exercise Effect (SMD > 1 favours exercise)
Knee extensor strength (56 RCTs, I2 = 80.9%)
SMD (95% CI): 0.59 (0.39, 0.75)
Pain (56 RCTs, I2 = 80.9%)
SMD (95% CI): 0.57 (0.42, 0.73)
Function (50 RCTs, I2 = 83.2%)
SMD (95% CI): 0.56 (0.39, 0.73)
ACSM guideline Exercise Effect (SMD > 1 favours exercise)
Knee extensor strength (22 RCTs, I2 = 86.9%)
SMD (95% CI): 0.83 (0.49, 1.17)
Pain (22 RCTs, I2 = 80.9%)
SMD (95% CI): 0.62 (0.32, 0.93)
Function (19 RCTs, I2 = 86.9%)
SMD (95% CI): 0.64 (0.28, 1.00)
Non-ACSM guideline Exercise Effect (SMD > 1 favours exercise)
Knee extensor strength (34 RCTs, I2 = 42.2%)
SMD (95% CI): 0.38 (0.27, 0.50)
Pain (34 RCTs, I2 = 73.2%)
SMD (95% CI): 0.52 (0.35, 0.68)
Function (31 RCTs, I2 = 71.8%)
SMD (95% CI): 0.49 (0.33, 0.65)		“This review supports the current recommendations on exercise as effective means for improving clinical outcomes among patients with knee OA.” (pp. 19)
“...exercise interventions with a focused aim at muscle strength [ACSM guideline exercises] do seemingly not provide superior clinical outcomes when compared to other types of exercise.” (pp. 19)
Fernandopulle et al., 20172
Recreational activity (SMD < 1 favours exercise)
WOMAC function score at 3 months (3 RCTs, n = 173, I2 = 0%) (P < 0.0001)
MD (95% CI): -9.56 (-13.95, -5.17)
*This analysis includes some hip OA patients
Conditioning exercise (SMD > 1 favours exercise)
Physical performance at 6 months (6MWT) (3 RCTs, n = 484, I2 = 56%) (P < 0.0001)
MD (95% CI): 42.72 (27.78, 57.66)
Conditioning exercise (SMD < 1 favours exercise)
WOMAC function score at 6 months (3 RCTs, n = 542, I2 = 83%) (P = 0.0002)
MD (95% CI): -3.74 (-5.70, -1.78)
*This analysis includes some hip OA patients
Physical performance at 6 months (timed stairs) (2 RCTs, n = 212, I2 = 99%) (P = 0.06)
MD (95% CI): -2.29 (-4.65, 0.06)
Physical performance at 18 months (timed stairs) (2 RCTs, n = 289, I2 = 0%) (P = 0.0003)
MD (95% CI): -0.49 (-0.75, -0.23)
Walking intervention (SMD < 1 favours exercise)
Physical function at 6 months (2 RCTs, n = 75, I2 = 0%) (P < 0.00001)
MD (95% CI): -10.38 (-12.27, -8.49)
Physical function at 12 months (3 RCTs, n = 156, I2 = 0%) (P = 0.83)
MD (95% CI): -0.03 (-0.35, 0.28)
Pain at 3 months (2 RCTs, n = 64, I2 = 0%) (P = 0.46)
MD (95% CI): 0.19 (-0.31, 0.68)
Pain at 6 months (2 RCTs, n = 88, I2 = 94%) (P = 0.14)
MD (95% CI): -1.55 (-3.62, 0.52)
Walking intervention (SMD > 1 favours exercise)
Physical performance at 12 months (6MWT) (2 RCTs, n = 105, I2 = 0%) (P = 0.93)
MD (95% CI): -1.88 (-43.46, 39.71)

50% of the included studies had a low risk of bias in most items except for blinding, compliance, and intention-to-treat analysis		“Conditioning exercises have moderate level of evidence for effectiveness on physical function in individuals with knee OA, in the short- and in the longer-term.” (pp. 19)

“PA [physical activity] interventions such as recreational activity (tai chi, Baduajin) and walking have very limited evidence to suggest a positive effect on pain and function in individuals with knee OA in the short-term.” (pp. 18)
Brosseau et al., 2017(1)10
Hatha Yoga exercise program versus Control (waitlist) EL I RCT (n = 36, [PEDro score 8/10])pp. 58
Hatha Yoga exercise program
  • Clinically important benefit demonstrated for WOMAC pain at 8 weeks
  • No statistically significant benefit for WOMAC physical function at 8 weeks
  • No benefit detected for QoL (SF-12) at 8 weeks
Recommendations: The eight-week Hatha Yoga program (60 minute classes once per week, plus 30 minute home program four times per week) for older women with knee osteoarthritis for management for pain relief (WOMAC subscale) at the eight weeks end of treatment measure is recommended.
Participation in the program is also suggested for improved physical function (WOMAC subscale) at end of treatment of eight weeks. There is a neutral improvement for quality of life (SF-12 subscale) at end of treatment of eight weeks.
Tai Chi Qigong exercise program versus Control (waitlist), EL I RCT (n = 44, [PEDro score 8/10])Tai Chi Qigong exercise program
  • Clinically important benefit demonstrated for QoL (SF-36) at 8 weeks
  • No statistically significant benefit for WOMAC pain at 8 weeks
Recommendations: The eight-week tai chi Qigong program (60 minute classes twice per week) for the management of knee osteoarthritis for improved quality of life (SF-36 subscale) at end of treatment eight weeks is recommended. The use of the program is also suggested for pain relief (WOMAC subscale) and improved physical function (WOMAC subscale) at end of treatment of eight weeks.
Sun style tai chi exercise program versus Control (waitlist), EL I RCT (n = 97, [PEDro score 8/10])Sun style tai chi exercise program
  • Clinically important benefit demonstrated for WOMAC physical function at 12 weeks
  • No benefit detected for WOMAC pain at 12 weeks
  • No benefit detected for QoL (SF-12) at 12 weeks
Recommendation: The 12-week Sun style tai chi exercise program (60 minute classes once per week) for management of knee osteoarthritis for improved physical function (WOMAC subscale) at the end of treatment of 12 weeks is recommended. There is a neutral improvement for pain relief (WOMAC subscale) and for quality of life (SF-12 subscale) at the end of treatment of 12 weeks.
Sun style tai chi exercise program versus Control (health education), EL I RCT (n = 55, [PEDro score 7/10])
  • Clinically important benefit demonstrated for WOMAC pain at 20 weeks
Recommendation: The 20-week Sun style tai chi exercise program (20 to 40-minute classes three times per week) for the management of knee osteoarthritis for pain relief (WOMAC subscale) and improved physical function (WOMAC subscale) at the end of treatment (20 weeks) is strongly recommended.
Brosseau et al., 2017(2)11
Home-based progressive strengthening exercise programpp. 607
  • Improved WOMAC pain and physical function at 4 months
  • No improvement in SF-36 QoL at 4 months
  • Strengthening exercises (with/without other types of therapeutic exercises), involving characteristics (type of resistance; type of contractions, modes of supervision, exercise program intensity and duration) can greatly reduce pain, improve physical function and quality of life for knee osteoarthritis patients.
  • Contribution of adjunctive therapies (e.g. patellar taping, manual-therapy, etc.) combined with strengthening exercise need to be studied.
  • There is a need to develop combined behavioral and muscle strengthening exercise strategies to improve long-term maintenance of strengthening exercise.

“Any type of strengthening exercise among the included trials on land-based exercise was identified as effective in this systematic review.” (pp. 607)
Progressive hip muscle strengthening home-based exercise program
  • Improved WOMAC pain and physical function at 1 week

Group education program followed by an unsupervised home-based exercise program
  • Improved WOMAC pain and physical function at 8 weeks but not 4 weeks

Isokinetic strengthening exercise program
  • Improved VAS pain and LI physical function at 8 weeks end of treatment and 6 month FU

Physiotherapy intervention program
  • Improved VAS pain at 3 months treatment end and 6 month FU
  • No improvement in KOOS pain or QoL at 3 months end of treatment and 6 month FU

Group-based supervised progressive strengthening and coordination exercise program
  • Improved KOOS QoL at 12 weeks treatment end
  • No improvement in KOOS pain or physical function at 12 weeks end of treatment

Osteoarthritis education and supervised strengthening exercise program with home exercises
  • Improved VAS pain at 6 months
  • No improvement in VAS pain, VAS QoL, and QoL (QoL scale) at 6 week end of treatment
  • No improvement in VAS QoL, or QoL (QoL scale) at 6 months

Supervised isokinetic, isotonic and isometric muscle strengthening exercise programs
  • Improved VAS pain and LI physical function at 8 weeks end of treatment and at 1 year FU

Supervised isokinetic muscle strengthening exercise program and hot packs application
  • Improved VAS pain and LI physical function at 1 year FU
  • No improvement in VAS pain, or LI physical function at 8 week end of treatment

Progressive exercise program, education and usual care
  • Improved WOMAC physical function at 6 weeks end of treatment
  • No improvement in WOMAC pain or physical function at 6 month FU

High and low-resistance strengthening exercise programs
  • Improved WOMAC pain and physical function at 8 weeks end of treatment

Non-weight-bearing and weight-bearing exercise programs
  • Improved WOMAC physical function at 8 weeks end of treatment

Quadriceps strengthening exercise program
  • Improved WOMAC pain at 24 months end of treatment

Progressive resistance exercise program of knees and hip muscles
  • Improved VAS pain, WOMAC pain, WOMAC physical function, SF-36 physical function and SF-36 QoL at 90 days end of treatment
  • No improvement in 6MWT physical performance at 90 days end of treatment

Strengthening and balance exercise program
  • Improved physical function (usual walkin speed) and VAS pain at 3 month end of treatment

Home-based physiotherapist prescribed supervised quadriceps strengthening exercise program
  • Improved WOMAC pain at 12 weeks end of treatment
  • No improvement in WOMAC physical function at 12 weeks end of treatment

Concentric-eccentric quadriceps strengthening exercise program
  • Improved WOMAC pain and WOMAC physical function at 8 weeks end of treatment

Physiotherapy exercise program
  • Improved VAS pain at 8 weeks end of treatment and 3 month FU
  • No improvement in KOOS QoL or KOOS pain at 8 weeks end of treatment and 3 month FU

Lower extremity strengthening exercise program
  • Improved HADS QoL, WOMAC pain, and WOMAC physical function at 6 months end of treatment

Physiotherapy exercise interventions
  • Improved VAS pain at night, VAS pain weight-bearing, and VAS pain at rest at 9 months FU
  • No improvement in VAS pain at night or AFI physical function at 3 month end of treatment
  • No improvement in AFI physical function at 9 month FU
  • No improvement in physical function (usual walking speed) at 3 month end of treatment or at 9 month FU

Mechanical diagnosis and therapy exercise program
  • Improved pain (P4 subscale), KOOS pain, and KOOS physical function at 2 weeks end of treatment
  • No improvement in pain (P4 subscale), KOOS pain, or KOOS physical function at 10 weeks FU

Strengthening exercise program with patient education
  • Improved LI pain and LI physical function at 8 weeks end of treatment
  • No improvement in 6MWT physical function or SF-36 QoL at 8 weeks end of treatment

Progressive supervised squat exercise program
  • Improved WOMAC pain and WOMAC physical function at 12 weeks end of treatment
  • No improvement in 6MWT physical function at 12 weeks end of treatment
Brosseau et al., 2017(3)3
Leg functional aerobic and strengthening exercise program versus Control (placebo ultrasound), EL I RCT (n = 83, [PEDro score 7/10])pp. 621
  • Improved WOMAC physical function at 4 week end of treatment
  • No statistically significant benefit for 6MWT physical function at 4 week end of treatment
  • A short-term aerobic exercise program with or without strengthening exercises is promising for reducing pain, as well as improving physical function and quality of life for individuals with knee osteoarthritis.
  • No strong conclusions can be drawn at the present time about the specific and potential beneficial effects of aerobic exercise programs alone in the management of knee osteoarthritis.
Aerobic exercise programme versus Control (OA health education), EL I RCT (n = 293, [PEDro score 6/10])
  • No statistically significant benefit for self-reported physical function at 18 month end of treatment

Individual and group supervised aerobic and strengthening exercise programs versus Control (waitlist), EL I RCT (n = 126, [PEDro score 7/10])
  • Improved WOMAC pain at 8 week end of treatment
  • No statistically significant benefit for SF-36 QoL or WOMAC physical function at 8 week end of treatment

Community physiotherapy exercise interventions, EL I RCT (n = 217, [PEDro score 8/10])
  • Improved WOMAC pain and WOMAC physical function at 3 month FU
  • No statistically significant benefit for WOMAC pain or WOMAC physical function at 3 month or 9 month FU

Aerobic and strengthening/resistance exercise programme versus Control (health education on OA and exercises), EL I RCT (n = 131, [PEDro score 8/10])
  • No statistically significant benefit for 6MWT physical function or WOMAC pain at 18 month end of treatment

Multi-component exercise programme versus Control (no intervention), EL I RCT (n = 56, [PEDro score 7/10])
  • No statistically significant benefit for KOOS pain, 6MWT physical function or KOOS QoL at 12 week end of treatment

Cycling exercise program
  • Improved WOMAC pain, WOMAC physical function, and KOOS QoL at 12 week end of treatment

Aerobic, strengthening exercise program and osteoarthritis health education
  • Improved AIMS2 pain and AIMS2 physical function at 3 month end of treatment
Li et al., 201612
Resistance exercise vs control (SMD < 1 favours exercise)pp. 958
Pain (20 RCTs, n = 2030, I2 = 37%) (P < 0.00001)
SMD (95% CI): -0.43 (-0.57, -0.29)
Stiffness (7 RCTs, n = 254, I2 = 0%) (P = 0.02)
SMD (95% CI): -0.31 (-0.56, -0.05)
Physical function (19 RCTs, n = 2077, I2 = 58%) (P < 0.00001)
SMD (95% CI): -0.53 (-0.70, -0.37)
  • Resistance exercise is effective on pain relief, stiffness alleviation, and physical function improvement in knee osteoarthritis.
  • High intensity resistance exercise had a larger effect size for pain and physical function than low intensity.
  • Within 12 weeks of resistance exercise was effective for stiffness.

“The treatment effect for many of the studies was only small and may have been influenced by the low methodological rigor of the studies included in the meta-analyses.” (pp. 958)
High intensity resistance exercise vs control (SMD < 1 favours exercise)
Pain (5 RCTs, n = 233, I2 = 44%) (P = 0.001)
SMD (95% CI): -2.16 (-3.45, -0.86)

Stiffness (3 RCTs, n = 112, I2 = 0%) (P = 0.15)
SMD (95% CI): -0.58 (-1.37, -0.21)
Physical function (6 RCTs, n = 303, I2 = 52%) (P < 0.0001)
SMD (95% CI): -0.42 (-0.62, -0.22)
Low intensity resistance exercise vs control (SMD < 1 favours exercise)
Pain (4 RCTs, n = 407, I2 = 0%) (P < 0.00001)
SMD (95% CI): -0.46 (-0.66, -0.26)
Stiffness (1 RCTs, n = 19) (P = 0.05)
SMD (95% CI): -34.00 (-68.56, 0.56)
Physical function (4 RCTs, n = 407, I2 = 0%) (P < 0.00001)
SMD (95% CI): -0.53 (-0.70, -0.37)
12 weeks or less -resistance exercise vs control (SMD < 1 favours exercise)
Pain (12 RCTs, n = 543, I2 = 36%) (P = 0.00001)
SMD (95% CI): -0.58 (-0.76, -0.41)
Stiffness (3 RCTs, n = 69, I2 = 0%) (P = 0.03)
SMD (95% CI): -0.54 (-1.04, -0.05)
Physical function (11 RCTs, n = 590, I2 = 64%) (P < 0.00001)
SMD (95% CI): -0.67 (-0.84, -0.50)
12 weeks or more -resistance exercise vs control (SMD < 1 favours exercise)
Pain (8 RCTs, n = 1487, I2 = 0%) (P < 0.00001)
SMD (95% CI): -0.27 (-0.38, -0.17)
Stiffness (4 RCTs, n = 185, I2 = 0%) (P = 0.09)
SMD (95% CI): -0.51 (-1.10, 0.08)
Physical function (8 RCTs, n = 1487, I2 = 0%) (P < 0.0001)
SMD (95% CI): -0.30 (-0.40, -0.20)

Eight of the included studies presented low risk in blinding assessment, while four studies presented high risk in intent-to-treat analysis. Another limitation of the evidence that was reported was an insufficient description of compliance in many studies.
Kan et al., 201613
Two studies identified with a relevant control (single group pre-post studies)
One study (n=45, Quality score = 18/32) found Yoga (60min/session, 3sessions/week, 12week treatment) statistically significant improvements in:
VAS pain
6MWT
30 second chair stand test
KOOS QoL

NS differences in:
Stair climbing
One study (n=45, Quality score = 16/32) found Modified Iyengar yoga (90min/session, 1session/week, 8week treatment) statistically significant improvements in:
WOMAC pain

NS differences in:
50 minute walk time		“Yoga may be a safe and tolerable exercise for patients with KOA since no studies reported adverse event both during and after yoga intervention.” (pp. 9)

“This systematic review showed that yoga has positive effect on pain relief on people with KOA with good evidence. A relative long period (12 weeks) of yoga intervention may help to improve the short-distance mobility in patients with KOA. More RCTs with high quality and larger sample size are needed.” (pp. 9)
Chang et al., 201614
tai chi Chuan vs control (SMD < 1 favours exercise)
WOMAC pain (6 RCTs, n = 250, I2 = 80%)
SMD (95% CI): -0.41 (-0.67, -0.14)
Average Jadad score: 4.0
WOMAC stiffness (6 RCTs, n = 250, I2 = 59%)
SMD (95% CI): -0.20 (-0.45, 0.05)
Average Jadad score: 4.0
WOMAC physical function (5 RCTs, n = 207, I2 = 41%)
SMD (95% CI): -0.16 (-0.44, 0.11)
Average Jadad score: 4.0
6MWT (3 RCTs, n = 97, I2 = 82%)
SMD (95% CI): -0.16 (-1.23, 0.90)
Average Jadad score: 4.0
SAFE (2 RCTs, n = 134, I2 = 99%)
SMD (95% CI): -0.63 (-0.98, -0.27)
Average Jadad score: 3.0
Stair climb test (2 RCTs, n = 53, I2 = 74%)
(95% CI): -0.74 (-1.34, -0.15)
Average Jadad score: 4.0		“In summary, tai chi Chuan had beneficial outcomes for patients with knee osteoarthritis, that is, improving knee extensor endurance, aerobic capacity, and body balance and coordination and reducing the body weight and bone density loss. Positive effects can be observed in the physical component in body functions and structures as well as activities and participation domains. There was insufficient evidence to support that tai chi Chuan had beneficial mental effect on patients with knee osteoarthritis, because of insufficient data in the recruited articles.” (pp. 9)
Henriksen et al., 20164
This analysis includes studies with mixed knee and hip OA with the authors intentions of making indirect analyses regarding exercise vs pharmaceutical interventions for knee OA specifically.“This meta-epidemiological study provides indirect evidence of comparable effects of exercise and oral analgesics for treating pain secondary to knee OA. These results can inform and support clinical management of patients that for some reason are unable to exercise or who consider exercise unviable.” (pp. 427)
Exercise vs control (SMD , > 1 favours exercise)
Pain (34 RCTs, n = 4179, I2 = 48%) (P < 0.001)
SMD (95% CI): 0.46 (0.34, 0.59)
Analgesics vs control (SMD > 1 favours analgesics)
Pain (20 RCTs, n = 5627, I2 = 63%) (P < 0.001)
SMD (95% CI): 0.41 (0.23, 0.59)
Indirect analysis
Exercise vs Analgesics (SMD > 1 favours exercise)
Pain (54 RCTs, n = 9806) (P = 0.61)
SMD (95% CI): 0.06 (-0.16, 0.28)
Exercise aquatic vs control (SMD > 1 favours exercise)
Pain (4 RCTs, I2 = 0%) (P = 0.259)
SMD (95% CI): 0.22 (-0.16, 0.28)
Exercise land vs control (SMD > 1 favours exercise)
Pain (30 RCTs, I2 = 51%) (P = 0.000)
SMD (95% CI): 0.49 (0.36, 0.63)
Indirect analysis
Exercise aquatic vs opioids (SMD > 1 favours exercise)
Pain (P = 0.34)
SMD (95% CI): -0.23 (-0.69, 0.24)
Indirect analysis
Exercise land vs opioids (SMD > 1 favours exercise)
Pain (P = 0.75)
SMD (95% CI): 0.05 (-0.25, 0.35)

Blinding (86% vs 6%) and handling of missing data (95% vs 57%) was conducted with less bias in trials of analgesics than in trials of exercise.		“This meta-epidemiological study based on trials included in Cochrane reviews, included 54 RCTs with 9806 patients, suggests comparable effects of exercise and orally administered analgesics for the conservative management of pain secondary to knee OA. Almost two-thirds of the reviewed trials pertained to exercise, which points out a trend toward testing exercise interventions for knee OA pain. This highlights the changed landscape of management and research in knee OA pain, which increasingly favors nonpharmacological interventions over drug interventions. Our results indicate that exercise and oral analgesics are comparable in terms of their pain relieving benefits, which supports the recommendation that exercise should be considered as the first choice, possibly accompanied by analgesics for control of severe pain or pain exacerbations.” (pp. 427)
Sharma, 20165
Three identified relevant studies on Exercise and Physical exercise interventions for Therapy were briefly described narratively in this SR.
  • Neuromuscular and quadriceps strengthening similarly improved pain and function but did not change the external knee adduction moment in the setting of moderate-severe medial knee OA with varus alignment in Australia. (pp. 11)
  • Pressure-pain sensitivity, temporal summation, and pain were reduced with exercise in patients with knee OA. (pp. 11)
  • Booster sessions with a physical therapist did not influence pain, function, or home exercise adherence in patients with knee OA. (pp. 11)
No specific relevant conclusions for exercise interventions for knee OA patients.
Anwer et al., 20161
Home exercise vs no intervention (SMD > 1 favours exercise)
Pain (11 trials, I2 = 75.26%) (P = 0.000)
SMD (95% CI): 0.464 (0.244, 0.685)

Group and Home exercise vs no intervention (SMD > 1 favours exercise)
Pain (2 trials, I2 = 0%) (P = 0.000)
SMD (95% CI): 0.804 (0.383, 1.224)
Home exercise vs no intervention (SMD > 1 favours exercise)
Function (9 trials, I2 = 68.85%) (P = 0.001)
SMD (95% CI): 0.354 (0.152, 0.555)		“Based on the high methodological quality of studies included in this systematic literature review and meta-analysis, it can be concluded that home exercise programs reduced knee pain and improved function in individuals with knee OA. The large evidence base supports the effectiveness of a variety of home programs including open and closed kinematic chain exercises.
In addition, small but growing evidence supports the effectiveness of other types of exercise such as tai chi, balance, and proprioceptive training for individuals with knee OA.” (pp. 47)
Tanaka et al., 201615
Exercise vs control (SMD > 1 favours exercise)pp. 49
Walking distance (11 RCTs, n = 1547, I2 = 54%) (P = 0.007)
SMD (95% CI): 0.44 (0.27, 0.60)
Quality of supporting evidence (GRADE): Very-low

Exercise vs control (SMD < 1 favours exercise)
Walking time (12 RCTs, n = 901, I2 = 49%) (P < 0.00001
SMD (95% CI): -0.50 (-0.70, -0.30)
Quality of supporting evidence (GRADE): Moderate
Exercise vs control (SMD > 1 favours exercise)
Walking velocity (6 RCTs, n = 581, I2 = 94%) (P < 0.00001)
SMD (95% CI): 1.78 (0.98, 2.58)
Quality of supporting evidence (GRADE): Low
  • Exercise therapy is effective for improving the amount of time spent walking, gait velocity, and maybe the total distance walked in people with knee osteoarthritis. The effect sizes for exercise therapy’s ability to improve walking ability remained unclear due to the heterogeneity of the studies.
  • Exercise intervention to increase strength, flexibility, and aerobic capacity may both improve walking ability and provide pain relief to people with symptomatic knee osteoarthritis.
Tanaka et al., 201516
Exercise vs control (SMD > 1 favours exercise)
SF-36 Physical component summary (7 RCTs, n = 771, I2 = 76%)
SMD (95% CI): 0.52 (0.21, 0.83)
Quality of supporting evidence (GRADE): Moderate
Exercise vs control (SMD > 1 favours exercise)
SF-36 Mental component summary (7 RCTs, n = 771, I2 = 77%)
SMD (95% CI): 0.44 (0.12, 0.75)
Quality of supporting evidence (GRADE): Moderate
Exercise vs control (SMD > 1 favours exercise)
SF-36 Physical functioning (4 RCTs, n = 587, I2 = 0%)
SMD (95% CI): 0.28 (0.12, 0.45)
Quality of supporting evidence (GRADE): High
Exercise vs control (SMD > 1 favours exercise)
SF-36 Role-physical (4 RCTs, n = 587, I2 = 19%)
SMD (95% CI): 0.26 (0.10, 0.43)
Quality of supporting evidence (GRADE): High
Exercise vs control (SMD > 1 favours exercise)
SF-36 Bodily pain (4 RCTs, n = 587, I2 = 55%)
SMD (95% CI): 0.22 (-0.04, 0.47)
Quality of supporting evidence (GRADE): Moderate
Exercise vs control (SMD > 1 favours exercise)
SF-36 General health (5 RCTs, n = 608, I2 = 6%)
SMD (95% CI): 0.15 (-0.01, 0.31)
Quality of supporting evidence (GRADE): High
Exercise vs control (SMD > 1 favours exercise)
SF-36 Vitality (4 RCTs, n = 587, I2 = 0%)
SMD (95% CI): 0.09 (-0.07, 0.26)
Quality of supporting evidence (GRADE): High
Exercise vs control (SMD > 1 favours exercise)
SF-36 Social functioning (5 RCTs, n = 608, I2 = 54%)
SMD (95% CI): 0.17 (-0.08, 0.43)
Quality of supporting evidence (GRADE): Moderate
Exercise vs control (SMD > 1 favours exercise)
SF-36 Role emotional (4 RCTs, n = 587, I2 = 0%)
SMD (95% CI): 0.07 (-0.10, 0.23)
Quality of supporting evidence (GRADE): High
Exercise vs control (SMD > 1 favours exercise)
SF-36 Mental Health (5 RCTs, n = 608, I2 = 41%)
SMD (95% CI): 0.15 (-0.08, 0.37)
Quality of supporting evidence (GRADE): Moderate		“In conclusion, this systematic review demonstrated that exercise therapy can improve HRQOL, as assessed by the SF-36, of knee OA sufferers. However, this review failed to clarify whether the effect sizes of exercise therapies on HRQOL, particularly the PCS and MCS, are influenced by dissimilar protocol designs among trials or by publication bias.” (pp. 3313)
Button et al., 201517
Relevant exercise interventions compared to no intervention control as reported in another SR were reported
12 and 18 month FU (n = 222)
Statistical improvement with exercise in:
Physical functioning (P = 0.02)
Pain Index (P = 0.000)
Stardardized physical component (P = 0.002)
Overall consensus on trial quality: Good
12 month FU (n = 222)
No statistical improvement with exercise in:
Stanford Self Efficacy Scale (P > 0.05)
Overall consensus on trial quality: Good

The focus of this SR was on self-management with an exercise component. Many additional studies included in this SR reported findings of potential interest as the self-management intervention was often education, goal setting, peer support, and other interventions considered as no intervention in other studies included in this report.		“The studies included in this review demonstrated an ‘unclear’ risk of bias and conflicting evidence regarding the long-term effect of self-care and exercise interventions. Nine of the included studies failed to have a long-term follow-up, which threatens the external validity of their findings. The four studies that did demonstrate long-term clinical effectiveness all used an OA population and had a strong focus on information provision, goal setting, and developing self-management skills. The exercise component of these interventions was poorly developed and could be strengthened by improving the exercise content, prescription, and progression. The evidence on exercise prescription needs to have a higher priority alongside self-care interventions.” (pp. 370)
Quicke et al., 201518
Most analyses were described narratively in this safety focused SR and the authors did not distinguish if the data was from RCTs that enrolled older patients with knee pain or older patients with knee OA.

Six studies were identified however that examined structural OA biomarker imaging responses to exercise:
“Of the five RCTs that measured changes in radiographic OA using imaging, none provided any evidence of significantly greater structural progression of OA between those in physical activity vs non-physical activity groups or those within physical activity group over time. A single small RCT found trends for improvements in the majority of OA parameters measured using MRI over time within the physical activity group whilst a single RCT found trends towards joint space narrowing within physical activity groups.” (pp. 1447)		“Falling was the most common moderate severity adverse event (n = 5).” (pp. 1450)

“Patients can be reassured that mild or temporary increases in pain with therapeutic exercise occur in a minority of individuals but pain does not equal harm or mean structural progression of knee OA and most will experience less pain if they persist with long-term exercise.” (pp. 1451)

“... although there was no evidence of increased frequency of TKR [total knee replacement surgery] or increased OA structural progression with physical activity, these results should also be interpreted with caution. This is because relatively few studies (five and six for each respective safety domain) contributed extractable data whilst the responsiveness of radiographs to detect OA structural change over periods less than 2 years is suboptimal which would tend to bias these safety outcomes towards the null.” (pp. 1451)
Ferreira et al., 201519
Three RCTs with data on KAM (3-D inverse dynamics)
1. Hip abductor/adductor strengthening vs no intervention (n = 76) (MD < 1 favours exercise)
WOMAC pain walking
MD (0-10) (95% CI): -1.37 (-2.16, -0.59)
WOMAC pain
MD (0-20) (95% CI): -2.40 (-3.25, -1.54)
WOMAC physical function
MD (0-68) (95% CI): -2.40 (-3.25, -1.54)
KAM
MD (95% CI): 0.13 (-0.12, 0.38)
Cochrane Collaborations’ Tool assessment of bias: High quality trial

This RCT also observed significantly increased hip joint torques, and knee extension torque (muscle strength) in the intervention group as compared to controls.
2. Hip abductor/adductor; knee extensors; ankle plantar flexors strengthening vs sham (n = 37) (MD < 1 favours exercise)
WOMAC pain
MD (0-20) (95% CI): -0.67 (-2.03, 0.69)
WOMAC physical function
MD (0-68) (95% CI): -2.99 (-7.77, 1.79)
KAM
MD (95% CI): 0.12 (-0.36, 0.80)
Cochrane Collaborations’ Tool assessment of bias: Low quality trial
3. Quadriceps strengthening vs no intervention (n = 92) (MD < 1 favours exercise)
WOMAC pain malaligned knee patients
MD (0-20) (95% CI): -1.6 (-7.06, 3.86)
WOMAC physical function malaligned knee patients
MD (0-68) (95% CI): -4.10 (-9.94, 1.74)
KAM malaligned knee patients
MD (95% CI): 0.18 (-0.06, 0.42)
WOMAC pain aligned knee patients
MD (0-20) (95% CI): -13.9 (-19.24, -8.55)
WOMAC physical function aligned knee patients
MD (0-68) (95% CI): -5.40 (-10.90, 0.10)
KAM aligned knee patients
MD (95% CI): -0.02 (-0.38, 0.34)
Cochrane Collaborations’ Tool assessment of bias: Low quality trial

The one high quality and two low quality trials constitute a moderate level of evidence to support the authors conclusions.		“The present systematic review showed that ET [exercise therapy] is effective in reducing pain, improving physical capacity, and enhancing muscle strength but has no effect on the KAM [knee adduction moment]. Considering such results, clinical efficacy of different protocols of ET was not followed by any alteration in the KAM in individuals with KOA. Several robust systematic reviews and practice guidelines endorse the positive clinical effects of ET in this population, but this is the first systematic review demonstrating that the dynamic KAM was not reduced by ET, even when clinical benefits were evident. Furthermore, a tendency favoring KAM increase was shown.” (pp. 525)

“The lack of knee adduction moment reduction indicates that exercise therapy may not be protective in knee osteoarthritis from a joint loading point of view. Alterations in neuromuscular control, not captured by the knee adduction moment measurement, may contribute to alter dynamic joint loading following exercise therapy. To conclude, mechanisms other than the reduction in knee adduction moment might explain the clinical benefits of exercise therapy on knee osteoarthritis.” (pp. 521)

6MWT = six-minute walk test; AHRQ = Agency for Healthcare Research and Quality; AIMS2 = Abnormal Involuntary Movement Scale 2; CI = confidence interval; EL = evidence level; FU = follow-up; HADS = Hospital Anxiety and Depression Scale; KAM = knee adduction moment; KOOS = Knee injury and Osteoarthritis Outcome Score; LI = Lequesne Index; MD = mean difference; NS = No significant difference; QoL = quality of life; RCT = randomized controlled trial; SAFE = Survey of Activities and Fear of Falling in the Elderly; SMD = standardized mean difference; TKR = total knee replacement; WOMAC = Western Ontario and McMaster Arthritis Index;

About the Series

CADTH Rapid Response Report: Summary with Critical Appraisal
ISSN: 1922-8147
CADTH receives funding from Canada’s federal, provincial, and territorial governments, with the exception of Quebec.

Suggested citation:

Exercise for the management of knee osteoarthritis: a review of clinical effectiveness. Ottawa: CADTH; 2017 Aug. (CADTH rapid response report: summary with critical appraisal).

Disclaimer: The information in this document is intended to help Canadian health care decision-makers, health care professionals, health systems leaders, and policy-makers make well-informed decisions and thereby improve the quality of health care services. While patients and others may access this document, the document is made available for informational purposes only and no representations or warranties are made with respect to its fitness for any particular purpose. The information in this document should not be used as a substitute for professional medical advice or as a substitute for the application of clinical judgment in respect of the care of a particular patient or other professional judgment in any decision-making process. The Canadian Agency for Drugs and Technologies in Health (CADTH) does not endorse any information, drugs, therapies, treatments, products, processes, or services.

While care has been taken to ensure that the information prepared by CADTH in this document is accurate, complete, and up-to-date as at the applicable date the material was first published by CADTH, CADTH does not make any guarantees to that effect. CADTH does not guarantee and is not responsible for the quality, currency, propriety, accuracy, or reasonableness of any statements, information, or conclusions contained in any third-party materials used in preparing this document. The views and opinions of third parties published in this document do not necessarily state or reflect those of CADTH.

CADTH is not responsible for any errors, omissions, injury, loss, or damage arising from or relating to the use (or misuse) of any information, statements, or conclusions contained in or implied by the contents of this document or any of the source materials.

This document may contain links to third-party websites. CADTH does not have control over the content of such sites. Use of third-party sites is governed by the third-party website owners’ own terms and conditions set out for such sites. CADTH does not make any guarantee with respect to any information contained on such third-party sites and CADTH is not responsible for any injury, loss, or damage suffered as a result of using such third-party sites. CADTH has no responsibility for the collection, use, and disclosure of personal information by third-party sites.

Subject to the aforementioned limitations, the views expressed herein are those of CADTH and do not necessarily represent the views of Canada’s federal, provincial, or territorial governments or any third party supplier of information.

This document is prepared and intended for use in the context of the Canadian health care system. The use of this document outside of Canada is done so at the user’s own risk.

This disclaimer and any questions or matters of any nature arising from or relating to the content or use (or misuse) of this document will be governed by and interpreted in accordance with the laws of the Province of Ontario and the laws of Canada applicable therein, and all proceedings shall be subject to the exclusive jurisdiction of the courts of the Province of Ontario, Canada.

Copyright © 2017 Canadian Agency for Drugs and Technologies in Health.

The copyright and other intellectual property rights in this document are owned by CADTH and its licensors. These rights are protected by the Canadian Copyright Act and other national and international laws and agreements. Users are permitted to make copies of this document for non-commercial purposes only, provided it is not modified when reproduced and appropriate credit is given to CADTH and its licensors.

Except where otherwise noted, this work is distributed under the terms of a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International licence (CC BY-NC-ND), a copy of which is available at http://creativecommons.org/licenses/by-nc-nd/4.0/

Bookshelf ID: NBK525040PMID: 30199179

Views

In this Page

Other titles in this collection

Related information

Similar articles in PubMed

See reviews...See all...

Recent Activity

ClearTurn OffTurn On

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...