The effectiveness of braces and orthoses for patients with knee osteoarthritis: A systematic review of Japanese-language randomised controlled trials

Abstract

Background:

Knee osteoarthritis is common. The evidence regarding the effectiveness of braces and orthoses for patients with knee osteoarthritis is inconclusive according to English-language literature. English-language reviews to date have not included Japanese-language studies.

Objectives:

This study aimed to collect and synthesise Japanese-language randomised controlled trials investigating the effectiveness of braces and orthoses for patients with knee osteoarthritis.

Study design:

Systematic review.

Methods:

Eight databases were systematically searched from inception to 29 July 2015. Only Japanese-language randomised controlled trials were included. Risk of bias was assessed using Physiotherapy Evidence Database scale. A meta-analysis was not appropriate due to the heterogeneity in the included studies.

Results:

Seven randomised controlled trials with low to high risks of bias were included. Six of seven included studies were conducted by the same author group. Limited evidence supported the positive effects of short-lever elastic knee braces to improve pain and functional disability in specific outcomes. No evidence was found to support the use of foot orthoses, such as laterally wedged insoles, medial arch support and metatarsal arch pad.

Conclusion:

Our systematic review found no conclusive evidence about the effectiveness of any braces and orthoses for patients with medial knee osteoarthritis. Future Japanese-language studies should address methodological flaws exposed in this review and strengthen the international evidence base.

Clinical relevance

This is the first systematic review of Japanese-language randomised controlled trials investigating orthoses for patients with knee osteoarthritis. Clinicians can consider the use of short-lever elastic knee braces to improve specifically pain on squat or walking. Evidence found in this review does not support the use of foot orthoses.

Keywords

Knee osteoarthritis braces and orthoses systematic review randomised controlled trials

Background

Knee osteoarthritis (OA) is an increasingly common condition in the world.¹ The growing prevalence is attributed to ageing societies and rising obesity throughout the world.^2,3 The prevalence of radiographic knee OA (Kellgren and Lawrence⁴ grades 2–4) among the general population aged 45 years or older in the United States was 28.0%, according to a large-scale study in 2007.⁵ The available prevalence of radiographic knee OA in Europe, Australia and Canada is also high, ranging from 13.4% to 67.0%.^6–10 Knee OA typically causes pain, swelling and decreased range of motion, which leads to functional disability and decreased quality of life among patients.¹¹ Due to the increasing prevalence, this disabling condition is anticipated to impose substantial individual suffering and global economic burden to our societies in the future.¹²

Although the pathophysiology of knee OA is not entirely clear, it is generally characterised as a degenerative joint disease which commonly involves the damage to the articular cartilage and subchondral bone,¹³ osteophyte formation, ligament laxity, atrophy of periarticular muscles¹⁴ and synovial inflammation.¹⁵ The medial compartment has been reported to be involved nearly 10 times more often than the lateral compartment.¹⁴ Patients with medial knee OA often have varus alignment, while patients with the involvement of the lateral compartment tend to assume valgus alignment.¹⁶ These malalignments increase the lever arm length between the ground reaction force and the centre of the joint thereby increasing the adduction/abduction moment resulting in increased contact pressure to the medial or lateral compartment.^17,18 In fact, malalignments have been reported to increase the risk of progression of this disease.¹⁹

Conservative management is usually recommended at first for knee OA. Conservative management includes various strategies, such as patient education,²⁰ weight management,²¹ aerobic exercise,²² strengthening exercise,²² manual therapy,²³ ultrasound,²⁴ braces and orthoses,²⁵ walking aids,²⁶ anti-inflammatory medication,²⁷ hyaluronic acid injection²⁸ and corticosteroid injection.²⁹ Braces and orthoses are commonly used to improve pain and functional disability of patients with knee OA.^30,31 Although its mechanism has not been proven yet, mechanical unloading to the affected compartment and neurophysiological pain modulation by proprioceptive inputs have been hypothesised.^30–32

In Japan, knee OA is epidemic due to the infamously ageing society, where the proportion of population aged over 65 years was estimated to be 26.5% in 2015.³³ According to a large cross-sectional study recruiting Japanese adults at various ages in 2012,³⁴ the prevalence of radiological knee OA (Kellgren and Lawrence grades 2–4) was 51.5% (57.9% in women and 38.8% in men). In another study focusing on subjects aged 60 years or older in 2009,³⁵ the figures were even higher (70.2% in women; 47.0% in men).

Historically, scientific studies written in non-English languages have been largely overlooked from international academics due to the dominant use of English.³⁶ However, this linguistic constraint can confound English-language secondary research, undermining the totality of evidence base.³⁷ Considering the high prevalence of this condition in the super-ageing society in Japan, we suspected that there might be some evidence on this topic in Japanese-language literature which had not been introduced into English-language literature. The uncharacteristic inclusion of English-language studies by Japanese authors Toda and colleagues^38–42 in the recent English-language reviews investigating the effectiveness of braces and orthoses for patients with knee OA^25,43–46 also heightened our expectation.

The aim of this systematic review was to collect, assess and synthesise Japanese-language randomised controlled trials (RCTs) about the effects of braces and orthoses on pain and functional disability of patients with knee OA.

Methods

Study protocol

This systematic review was written in line with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement.⁴⁷ The protocol of this review was registered in the international prospective register of systematic reviews PROSPERO network (registration number CRD42015024689) before commencing a systematic search.

Eligibility criteria

The clinical question for the review was developed through a participants, intervention, comparator, outcome measures and study design (PICOS) format.⁴⁸ Specific definitions of the components of the clinical question are provided in Table 1. Studies were included if they reported on patients with knee OA, diagnosed with radiological investigation (Kellgren and Lawrence⁴ grades 1–4) and clinical symptoms. Limitation was not set on age or gender. Interventions of interest included any types of braces or orthoses for knee, foot or ankle. No restriction was placed on comparator interventions. We considered primary outcome measures such as pain and functional disability with any duration of follow-up. Studies examining only kinetic or kinematic parameters were excluded. This systematic review included only RCTs to minimise bias. Only Japanese-language articles published in peer-reviewed journals were included.

Table 1.

PICOS format and search terms.

PICOS format	Definition	Search terms
Participants	Patients with knee osteoarthritis	knee AND (osteoarthritis OR osteoarthrosis OR (degenerative joint disease))
Intervention	Any types of braces or orthoses for knee, foot or ankle	(brace* OR bracing OR supporter OR stabili?er OR orthotic* OR orthosis OR orthoses OR insole* OR wedge) NOT (surg OR (knee arthroplasty) OR (knee replacement))
Comparison	Any interventions	(not set)
Outcome	Pain and functional disability	(not set)
Study design	RCTs	(not set)

PICOS: participants, intervention, comparator, outcome measures and study design; RCT: randomised controlled trial.

Information sources and search strategy

A systematic search was performed using three Japanese-language databases, including CiNii, Ichushi and Medical*Online. An additional search through five English-language databases with Japanese indexes (CINAHL, Embase, MEDLINE, Scopus and Web of Science) was performed to find relevant Japanese-language studies with English-language abstracts. All databases were searched from the earliest available dates to 29 July 2015. The three Japanese-language databases were chosen due to their comprehensiveness. Five English-language databases were selected due to their capacity to limit articles to Japanese language. Search terms were established based on the PICOS format (Table 1). In Japanese-language databases, Japanese equivalents of these words were used. The systematic search was performed collaboratively by two authors (T.N. and K.M.).

After the initial systematic search, duplicates were removed and titles and abstracts of the remaining articles were screened subsequently. In the last process, full texts of potentially relevant articles were assessed for eligibility. Reference lists of included studies were also hand searched.

Risk of bias assessment

Risk of bias assessment was performed using Physiotherapy Evidence Database (PEDro) scale.⁴⁹ The PEDro scale consists of 11 specific criteria. Items 2–9 refer to the internal validity of an article, and items 10 and 11 refer to the statistical analysis, ensuring sufficient data to enable appropriate interpretation of the results.⁴⁹ Item 1 is related to the external validity and therefore not included in the total PEDro score. This critical appraisal tool has been proven to be reliable and valid by previous studies.^50,51 The risk of bias in each study was classified as ‘high’ (4/10 or less), ‘moderate’ (5 or 6/10) or ‘low’ (7/10 or more) according to total PEDro scores.

Two independent Japanese-speaking physiotherapists (T.N. and K.M.) critically appraised all included studies. After completing individual assessments, the scores were compared, and inter-tester reliability of the PEDro scores was determined as Cohen’s kappa statistics.⁵² Kappa statistics were classified as almost perfect agreement (0.81–1.00), substantial agreement (0.61–0.80), moderate agreement (0.41–0.60), fair agreement (0.21–0.40), slight agreement (0.00–0.20) or poor agreement (<0.00).⁵³ When consensus could not be reached between the two assessors, opinions from an English-speaking author (S.M.) were sought to resolve disagreements.

Data extraction and synthesis

Data were extracted on sample characteristics (sample size, age, gender, body mass index (BMI), ethnicity, radiological severity, pain duration, laterality and co-morbidity count), types of interventions, outcome measures, results (pre- and post-intervention data, follow-up data, statistical significance and effect size) and ethical considerations for participants. The results with any duration of follow-ups were extracted. We anticipated that a meta-analysis might be undertaken, should there be homogeneity among included studies and sufficient data available. The corresponding authors of included articles were expected to be contacted for necessary data to complete further analysis.

Results

Study selection

The literature search process using eight electronic databases is illustrated in Figure 1. The initial search identified 437 records. After removing 89 duplicates, titles and abstracts of 348 articles were screened. In the screening, 337 articles were excluded primarily because they were not RCTs. This yielded 11 potentially relevant articles. In a full-text assessment, three studies were excluded because they used pseudo-randomisation.^54–56 Another one study, where interventions were chosen by patients, was also excluded at this stage.⁵⁷ The hand search in reference lists of included studies did not discover any other relevant articles. Accordingly, seven Japanese-language RCTs were included in this review.

Figure 1.

Flowchart of study selection.

Study characteristics

A summary of the most important characteristics of seven included studies is presented in Table 2. The same author group was involved in six of seven studies and all the six trials were conducted in the authors’ private orthopaedic clinic.^58,60–64 Five different articles written in English by this author group had been included in one English-language systematic review.²⁵ The corresponding author was contacted by e-mail and it was confirmed that their studies recruited only patients who had never received treatments using braces or orthoses, and therefore there was no duplicate patient across their studies irrespective of languages. None of the included studies was approved by external ethics committees.

Table 2.

Summary of included studies.

Study	PEDro score	Participants	Interventions	Outcome measures	Results
Toda⁵⁸	6/10	N = 41 (4 men and 37 women) Age: 64.8 ± 11.0 years BMI: 24.5 ± 2.5 (LW group) and 25.7 ± 4.4 (LW + AO group) Medial knee OA (KL grades 2–4) Unilateral symptom in 29 patients and bilateral in 12 patients Diagnosis based on American College of Rheumatology criteria⁵⁹ Pain duration: 4.5 ± 5.2 years	LW group (n = 21): laterally wedged insoles for both sides were given to patients. LW + AO group (n = 20): In addition to the same insoles for both sides, ankle orthoses (originally to prevent ankle inversion sprain) for symptomatic sides were also given to patients. Patients were instructed to wear the devices for 5–10 h/day. Anti-inflammatory medication was also prescribed for patients in both groups. Interventions were performed for 4 weeks.	Lequesne index (total score) FTA	There was significant improvement in Lequesne index in LW + AO group compared to LW group (p < 0.05). There was no significant difference in improvement in FTA between two groups.
Toda and Tsukimura ⁶⁰	6/10	N = 35 (11 men and 24 women) Age: 64.7 ± 11.3 years BMI: 24.9 ± 4.2 (LW group) and 24.8 ± 3.8 (LW + MAP group) Medial knee OA (KL grades 2–4) No information about symptom laterality Diagnosis based on American College of Rheumatology criteria⁵⁹ Pain duration: 18.0 ± 3.8 years	LW group (n = 16): laterally wedged insoles were given to patients. LW + MAP group (n = 19): in addition to the same insoles were also given to patients. Patients were instructed to wear the devices for 5–10 h/day. Intra-articular hyaluronic acid injection was used once per week in both groups. Interventions were performed for 4 weeks. Whether interventions were provided to both sides or only to symptomatic sides was not clear.	Lequesne index (total score) Body sway in static standing posture	There was significant improvement in Lequesne index in LW group compared to LW + MAP group (p < 0.05). There was no significant difference in change in body sway in static standing posture after interventions between two groups.
Toda et al.⁶¹	6/10	N = 56 (11 men and 45 women) Age: 61.1 ± 10.3 years BMI: 23.4 ± 3.8 (KB group) and 24.0 ± 2.2 (KB + pad group) Medial knee OA (KL grades 2–4) No information about symptom laterality Diagnosis based on American College of Rheumatology criteria⁵⁹ Pain duration: 2.5 ± 4.9 years	KB group (n = 28): short-lever elastic knee braces were provided. KB + pad group (n = 28): The same knee braces with a pad attached inside to protect skin from creases during deep knee flexion was provided. Patients were instructed to wear the devices for 5–10 h/day. Anti-inflammatory medication (twice per day), intra-articular steroid injection (only in the first visit) and intra-articular hyaluronic acid injection (once per week) were also prescribed for patients in both groups. Interventions were performed for 4 weeks. Whether interventions were provided to both sides or only to symptomatic sides was not clear.	Lequesne index (sub-score)	There was significant improvement in only one subcategory of Lequesne index (pain on squat) in KB + pad group compared to KB group (p < 0.05). There was no significant difference in other subcategories between two groups.
Keyaki et al.⁶²	8/10	N = 62 (21 men and 41 women) Age: 65.6 ± 8.3 years BMI: 24.5 ± 3.5 (KB group) 24.8 ± 3.0 (LKB group) Medial knee OA (KL grade 2–4) No information about symptom laterality Diagnosis based on American College of Rheumatology criteria⁵⁹ Pain duration: 3.5 ± 4.6 years	KB group (n = 31): short-lever elastic knee braces with coils to support MCL were provided. LKB group (n = 31): long-lever braces with hinged metal bars to support MCL and ACL were provided. Patients were instructed to wear the devices for 5–10 h/day. All patients in both groups were treated with intra-articular hyaluronic injection subsequently. Interventions were performed for 2 weeks. Whether interventions were provided to both sides or only to symptomatic sides was not clear.	Lequesne index (sub-score)	There was significant improvement in only one subcategory of Lequesne index (pain during walking on uneven ground) in SKB group compared to LKB group (p < 0.05). There was no significant difference in other subcategories between two groups.
Toda et al.⁶³	6/10	N = 59 (14 men and 45 women) Age: 64.9 ± 9.0 years BMI: 24.6 ± 4.1 (UKB group) and 23.7 ± 3.2 (BKB group) Unilateral medial knee OA (KL grades 2–4) Diagnosis based on American College of Rheumatology criteria⁵⁹ Pain duration: 2.7 ± 3.1 years	UKB group (n = 29): short-lever elastic knee braces were provided for only symptomatic sides. BKB group (n = 30): the same knee braces were provided for both sides. Patients were instructed to wear the devices for 5–10 h/day. All patients in both groups were treated with intra-articular steroid injection at the first visit and hyaluronic injection subsequently. Interventions were performed for 4 weeks.	Lequesne index (sub-score)	There was significant improvement in only one subcategory of Lequesne index (pain on walking) in BKB group compared to UKB group (p < 0.05). There was no significant difference in the other subcategories between two groups.
Toda⁶⁴	5/10	N = 58 (19 men and 39 women) Age: 62.4 ± 10.2 years BMI: 24.5 ± 3.6 (SLW + HKS group), 24.2 ± 3.7 (SLW + KB group) and 25.4 ± 3.9 (SLW + LKB group) Medial knee OA (KL grade 2–4) Unilateral symptom in 50 patients and bilateral in 8 patients Diagnosis based on American College of Rheumatology criteria⁵⁹ Pain duration: not specifically mentioned	SLW + HKS group (n = 20): strapping laterally wedged insoles and heat-retaining knee supporter were provided. SLW + KB group (n = 19): strapping laterally wedged insoles and short-lever elastic knee braces without hinges to support ACL were provided. SLW + LKB group (n = 19): strapping laterally wedged insoles and long-lever knee braces with hinged metal bars to support ACL, MCL and LCL were provided. Patients were instructed to wear the devices for 5–10 h/day. All patients in both groups were treated with intra-articular steroid injection at the first visit and hyaluronic injection subsequently. Interventions were performed for 4 weeks. Whether interventions were provided to both sides or only to symptomatic sides was not clear.	Lequesne index (total score)	There was no difference in improvement in Lequesne index between three groups.
Nakajima and Fukui⁶⁵	4/10	N = 64 (13 men and 51 women) Age: 69.1 ± 9.2 years BMI: not reported. Medial knee OA (stages 2–4 in Hokudai classification)⁶⁶ No information about symptom laterality Diagnosis based on Hokudai classification⁶⁶ Pain duration: not specifically mentioned	LW group (n = 26): laterally wedged insoles were provided. MAS group (n = 22): medial arch support insoles were provided. LW + MAS (n = 16): both laterally wedged insoles and medial arch support insoles were provided. Interventions were performed for 10 weeks. Detailed instructions regarding prescription of orthoses were not available from the article. Whether interventions were provided to both sides or only to symptomatic sides was not clear.	Pain on walking Pain on stairs Both outcomes were subcategories of Japanese Orthopaedic Association score for knee OA	There was no difference in improvement in three groups.

PEDro: Physiotherapy Evidence Database; OA: osteoarthritis; KL: Kellgren and Lawrence; LW: laterally wedged insole; AO: ankle orthosis; FTA: femorotibial angle; BMI: body mass index; MAP: metatarsal arch pad; KB: knee brace; LKB: long-lever knee brace; SKB: short-lever knee brace; MCL: medial collateral ligament; ACL: anterior cruciate ligament; UKB: unilateral knee brace; BKB: bilateral knee brace; SLW: strapping laterally wedged insole; HKS: heat-retaining knee supporter; LCL: lateral collateral ligament; MAS: medial arch support.

Age is described as mean ± standard deviation.

Study designs

All seven studies were RCTs with parallel designs.

Participants

Overall, 375 patients (93 men and 282 women) participated in the seven studies. Sample sizes ranged from 35 to 64 per study. None of the studies performed sample size calculations. The mean ages of patients were relatively consistent, ranging from 61.1 to 69.1 years. Information regarding participants’ ethnicity was not available. All included studies recruited patients with medial knee OA. In six studies by the same author group,^58,60–64 diagnosis was made based on American College of Rheumatology,⁵⁹ and patients with Kellgren and Lawrence⁴ grades 2–4 were eligible. Only one study used Hokudai classification for diagnosis, and patients with stages 2–4 were included.⁶⁵ The majority of participants in the included studies were categorised into grade 2 or 3 in these classification systems in terms of the radiological severity. The mean BMI ranged from 23.4 to 25.7, except for one trial which did not clarify BMI.⁶⁵ The mean pain durations ranged from 2.5 to 18.0 years in six studies,^58,60–64 whereas two studies did not specifically mentioned pain duration.^64,65 One study recruited only patients with unilateral OA,⁶³ while two studies examined both patients with unilateral and bilateral OA.^58,64 The other four studies did not clarify the laterality of symptoms.^60–62,65

Interventions

Three studies explored the effects of various types of knee braces.^61–63 Although detailed mechanical properties of knee braces were not available in two studies,^61,63 none of the three studies used conventional valgus knee braces according to the visual images in the articles. Four studies investigated different types of foot orthoses, including laterally wedged insoles, medial arch supports and metatarsal arch supports.^58,60,64,65 The maximum thickness of laterally wedged insoles ranged from 7 to 8 mm,^58,60,64,65 and one study did not mention wedge thickness.⁶⁴ Only one study used a combination of knee braces and foot orthoses.⁶⁴ In the six studies by the same author group, patients were instructed to wear devices for 5–10 h/day,^58,60–64 while the other study did not specify the dosage of interventions.⁶⁵ In five studies, whether interventions were provided to both sides or only to symptomatic sides was not clear.^{60–62,64,65} Included studies commonly used co-interventions, such as intra-articular hyaluronic injection,^60–64 intra-articular steroid injection^61,63,64 and anti-inflammatory medication.⁵⁸ One study did not specifically report co-interventions.⁶⁵ Intervention periods were 4 weeks in five studies,^{58,60,61,63,64} 2 weeks in one study⁶² and 10 weeks in the other study.⁶⁵

Outcome measures

The six studies by the same author group consistently used the Lequesne index. The Lequesne et al.⁶⁷ index was originally developed and validated in French as an assessment tool to assess pain, stiffness and functional disability of patients with knee or hip OA. This questionnaire has been translated into several languages.^68–70 To the best of the authors’ knowledge, however, there is no formally translated and validated Japanese version of the Lequesne index. This outcome measure was analysed as a total score in three studies^58,60,64 or as a separate scores in subcategories in the other three studies.^61–63 One study by another author group used two subcategories of the Japanese Orthopaedic Association (JOA) score for knee OA, including pain on walking and stairs.⁶⁵ The JOA score for knee OA in Japanese has been reported to be a reliable and valid assessment tool.⁷¹ All studies assessed outcome measures immediately before and after interventions and did not perform any long-term follow-up assessments.

Risk of bias assessment

The results of the critical appraisal using PEDro scale are presented in Table 3. Included studies had a median score 6 of 10, ranging from 4 to 8. Five studies had moderate risk of bias, while one study had high risk of bias and only one study had low risk. Common methodological flaws included no blinding of participants (100%), therapists (100%) and assessors (85.7%), and no intention-to-treat analysis (85.7%). The two reviewers (T.N. and K.M.) showed an almost perfect level of inter-rater reliability in determining PEDro scores of included articles (overall kappa score of 0.96).

Table 3.

A comparison of between-group effect sizes.

Study	Outcome measure (timeframe)	Intervention	Comparator intervention	Effect size (95% CI)
Toda⁵⁸	Total score in Lequesne index (4w)	LW + AO	LW	0.77 (−0.39 to 1.94)
Toda and Tsukimura⁶⁰	Total score in Lequesne index (4w)	LW + MAP	LW	−0.83 (−2.02 to 0.36)
Toda⁶⁴	Total score in Lequesne index (4w)	SLW + KBH	SLW + HKS	−0.84 (−2.51 to 0.82)
		SLW + KB	SLW + HKS	−0.02 (−1.67 to 1.63)
Toda et al.⁶¹	Nocturnal pain (4w)	KB + pad	KB	−0.35 (−1.10 to 0.36)
	Morning pain (4w)	KB + pad	KB	0.53 (−0.57 to 1.80)
	Pain on standing (4w)	KB + pad	KB	0.33 (−0.62 to 1.35)
	Pain on walking (4w)	KB + pad	KB	−0.27 (−0.98 to 0.42)
	Pain on sitting (4w)	KB + pad	KB	−0.10 (−0.90 to 0.68)
	Maximum distance walked (4w)	KB + pad	KB	−0.12 (−0.71 to 0.46)
	Going up stairs (4w)	KB + pad	KB	0.18 (−0.56 to 0.94)
	Going down stairs (4w)	KB + pad	KB	0.06 (−0.63 to 0.76)
	Squat (4w)	KB + pad	KB	0.83 (0.15 to 1.59)^a
	Walking on uneven ground (4w)	KB + pad	KB	0.20 (−0.53 to 0.97)
Keyaki et al.⁶²	Nocturnal pain (2w)	KB	LKB	0.41 (−0.22 to 1.07)
	Morning pain (2w)	KB	LKB	−0.10 (−1.14 to 0.91)
	Pain on standing (2w)	KB	LKB	0.49 (−0.14 to 1.17)
	Pain on walking (2w)	KB	LKB	0.20 (−0.36 to 0.76)
	Pain on sitting (2w)	KB	LKB	−0.17 (−0.95 to 0.58)
	Maximum distance walked (2w)	KB	LKB	0.46 (−0.11 to 1.07)
	Going up stairs (2w)	KB	LKB	0.53 (−0.18 to 1.32)
	Going down stairs (2w)	KB	LKB	0.09 (−0.52 to 0.71)
	Squat (2w)	KB	LKB	0.24 (−0.33 to 0.83)
	Walking on uneven ground (2w)	KB	LKB	0.92 (0.16 to 1.79)^a
Toda et al.⁶³	Nocturnal pain (4w)	BKB	UKB	0.03 (−0.60 to 0.67)
	Morning pain (4w)	BKB	UKB	0.14 (−0.98 to 1.31)
	Pain on standing (4w)	BKB	UKB	−0.22 (−0.92 to 0.47)
	Pain on walking (4w)	BKB	UKB	0.72 (0.07 to 1.45)^a
	Pain on sitting (4w)	BKB	UKB	−0.71 (−1.58 to 0.07)
	Maximum distance walked (4w)	BKB	UKB	0.42 (−0.20 to 1.07)
	Going up stairs (4w)	BKB	UKB	0.35 (−0.40 to 1.14)
	Going down stairs (4w)	BKB	UKB	0.39 (−0.30 to 1.12)
	Squat (4w)	BKB	UKB	0.39 (−0.23 to 1.05)
	Walking on uneven ground (4w)	BKB	UKB	0.55 (−0.24 to 1.43)
Nakajima and	Pain on walking (10w)	MAS	LW	−0.06 (−1.57 to 1.46)
Fukui⁶⁵	Pain on stairs (10w)	LW + MAS	LW	0.21 (−1.06 to 1.49)
		MAS	LW	0.70 (−0.10 to 1.51)
		LW + MAS	LW	0.89 (−0.06 to 1.84)

w: week; LW: laterally wedged insole; AO: ankle orthosis; MAP: metatarsal arch pad; SLW: strapping laterally wedged insole; KBH: knee brace with hinged metal bars; KB: knee brace without hinged metal bars; HKS: heat-retaining knee supporter; LKB: long-lever knee brace; BKB: bilateral knee brace; UKB: unilateral knee brace; MAS: medial arch support; CI: confidence interval.

95% CI of effect size excluded 0, indicating statistical significance.

Data extraction and synthesis

Due to the heterogeneous interventions and comparator interventions across the included studies, it was thought to be inappropriate to undertake a meta-analysis. A qualitative synthesis was undertaken instead. None of the included studies reported effect sizes. Cohen’s d and 95% confidence interval (CI) were calculated for between-group effect sizes based on available data (Table 3).^72,73 Cohen’s d was calculated by dividing mean change difference between intervention and comparator groups by pooled standard deviation.⁷⁴

Effectiveness of knee braces

Three studies by the same author group investigated the effectiveness of different types of knee braces.^61–63 One study by Toda et al.⁶¹ reported significantly greater effects of knee braces with pads to improve pain on squat compared to knee braces without pads. Knee braces used in this study were short-lever, elastic, equipped with coils but not hinges. The location of coils was not clearly mentioned in the article. The reported findings corresponded with our further analysis (Cohen’s d = 0.83, 95% CI = 0.15–1.59). The authors hypothesised that the pads inside the knee braces might have protected skin from creases of braces during squat, resulting in reduction in pain. It should be noted that there was no significant difference between the two interventions in the other nine outcomes in the Lequesne index.

One study by Keyaki et al.⁶² found that elastic short-lever knee braces were significantly more effective to improve pain on walking on uneven ground compared to long-lever braces (Cohen’s d = 0.92, 95% CI = 0.16–1.79). According to the study, short-lever knee braces were equipped with coils to support the medial collateral ligament (MCL), while long-lever braces were equipped with metal hinges and bars on both medial and lateral sides to support MCL and anterior cruciate ligament (ACL). The location of coils in the short-lever brace was not specified in the article. This study reported no significant difference between the two interventions in the other nine outcomes in Lequesne index.

Another study by the same author group reported a significantly greater effect on pain on walking when knee braces were applied to both knees compared to when the same type of braces were applied to only symptomatic sides (Cohen’s d = 0.72, 95% CI = 0.07–1.45). Knee braces used in this study were elastic and not equipped with hinges. Detailed mechanical properties of the knee braces were not available in the article. No significant difference was reported between the two interventions in the other nine outcomes in the Lequesne index.

Effectiveness of foot orthoses

Three studies examined the effectiveness of various types of foot orthoses.^58,60,65 One study by Toda⁵⁸ found that the combined use of laterally wedged insoles (8-mm maximum thickness) and figure-eight ankle orthoses was more effective than laterally wedged insoles alone to improve the total score of Lequesne index. However, our analysis using effect size revealed no statistical significance (Cohen’s d = 0.77, 95% CI = −0.39 to 1.94).

Another study by the same author group reported significantly better effect of laterally wedged insoles (8-mm maximum thickness) compared to the combined use of the same insoles and metatarsal arch pads to improve Lequesne index. According to our further analysis, however, its effect size was not statistically significant (Cohen’s d = 0.83, 95% CI = −0.36 to 2.02).

The last study by Nakajima and Fukui⁶⁵ compared the effectiveness of laterally wedged orthoses (7-mm maximum thickness), medial arch support and a combination of both. They found no difference among three interventions, which corresponded to our analysis using effect sizes (see Table 3).

Effectiveness of combined interventions

One study investigated the effectiveness of combinations of foot orthoses and knee braces: laterally wedged insoles with subtalar strapping and a heat-retaining knee supporter, the same insoles and short-lever knee braces without hinges and the same insoles and long-lever knee braces with hinged metal bars. The thickness of the laterally wedged insoles was not stated. The article described that short-lever knee braces were elastic and designed to support the ACL, while long-lever knee braces were equipped with hinged metal bars on both medial and lateral sides to support MCL, lateral collateral ligament (LCL) and ACL. This study reported no significant difference between three groups, which corresponded to our additional analysis of effect sizes (see Table 3).

Discussion

Summary of main results

The purpose of this review was to collect and synthesise all available Japanese-language RCTs investigating the effectiveness of braces and orthoses for patients with knee OA. The systematic search identified seven eligible RCTs with low to high risk of bias. Meta-analysis was thought to be inappropriate due to the absence of common interventions and comparator terms. The reported positive effects of knee braces were limited to specific symptoms and might not be related to clinically relevant overall effects on pain and functional disability. Although two studies reported significant findings on the effects of foot orthoses on the Lequesne index, our further analysis revealed that effect sizes were not statistically significant.

Source of bias and limitations of studies

All seven included studies failed to blind patients and clinicians.^58,60–65 Given that blinding patients to interventions is important particularly when the outcome measures are subjective, this methodological flaw might have affected the findings of these studies which consistently used subjective questionnaires, including the Lequesne index and JOA. Six of seven studies had or potentially had drop-outs and they did not perform intention-to-treat analysis. The lack of intention-to-treat analysis might have biased the estimates of treatment effects in these studies.⁷⁵

Reporting styles were generally poor and information regarding results was insufficient across all the included studies. No information was available with regard to laterality of OA in four studies.^60–62,65 Although co-morbidity count and ethnicity have been reported to be strong prognostic factors for symptomatic progression of knee OA,⁷⁶ none of the included studies clarified these factors. The mechanical properties of knee braces (e.g. supporting LCL, MCL and/or ACL) investigated in two studies were not reported in detail in two studies.^61,63 It was not clearly stated whether interventions were provided to both sides or only to symptomatic sides in five studies.^{60–62,64,65} Potential heterogeneity in participants and inconsistencies in intervention methodologies within these studies might have confounded the results.

Sample sizes were consistently small (35–64 per study) and none of the studies performed sample size calculation. Potentially insufficient sample sizes might have compromised statistical precision, leading to wide 95% CI and statistical insignificance.⁷⁷ Although effect sizes were large in some comparisons, statistical significance was identified only in two comparisons out of 38 (see Table 3).⁷⁸ Finally, the use of invalidated Japanese version of the Lequesne index evaluated in six studies might have introduced another bias.^58,60–64

Limitations of this review

This review was potentially limited by its search strategy, as we aimed to include only published literature. This strategy may be subject to publication bias because trials with positive findings tend to be published more often and more quickly compared with trials with negative findings.⁷⁹ We found that six of seven included studies were investigated by the same author group.^58,60–64 Thus, the findings in our review might be affected by potential selection bias in these studies. Since all included studies evaluated only short-term outcomes, long-term effects of braces and orthoses were unclear. This review also suffered from the lack of homogeneity in the included studies, preventing the quantitative synthesis of effect sizes with enhanced statistical precision. Although language bias was an inevitable consequence of this systematic review, we hope that the findings in this study will be integrated into broader framework in further research and complement the international evidence base.

Comparison with other reviews

Although one recent English-language systematic review concluded that valgus knee braces might be more effective to decrease external knee adduction moment during walking compared to no intervention, clinical benefits were uncertain.⁸⁰ Another recent review reported low-quality evidence that patients with medial knee OA might benefit more from knee braces than from no treatment and moderate-quality evidence that laterally wedged insoles are as effective as neutral insoles.²⁵ This review also found the lack of evidence to support superior effects of laterally wedged insoles compared to no treatment. In another systematic review, a significant difference in effect sizes of laterally wedged insoles was found between the two types of control interventions (either neutral insoles or no treatment) with smaller effect sizes seen in studies where control interventions were neutral insoles.⁴⁶ Based on this finding, we speculate that various comparator interventions used in the included studies in our review might have led to statistically insignificant effect sizes for foot orthoses (see Table 3).

Implications for future research

Due to the paucity of high-quality RCTs, evidence regarding positive effects of braces and orthoses for patients with knee OA is inconclusive. Future Japanese-language RCTs should reflect on and improve the methodological flaws exposed in this review to contribute to a stronger evidence base. Particularly, future studies should recruit sufficient numbers of patients based on sample size calculations to ensure higher statistical precision and improve reporting styles, disclosing sufficient information to enable further analysis in secondary studies. Future studies should seek ethics approval from relevant human ethics committees.

Implications for clinical practice

Our analysis of effect sizes revealed statistically significant differences favouring knee braces with pads for improvement in pain on squat (Cohen’s d = 0.83, 95% CI = 0.15–1.59), short-lever elastic braces for improvement in pain on walking on uneven grounds (Cohen’s d = 0.92, 95% CI = 0.16–1.79) and the use of short-lever elastic knee braces for both sides to improve pain on walking (Cohen’s d = 0.72, 95% CI = 0.07–1.45) (see Table 3). These were the only three comparisons where statistical significance was identified. Therefore, the use of these knee braces can be discussed between clinicians and patients with the consideration of patient preferences and costs, when pain on squat or pain during walking on uneven or flat surface is the main symptom of patients. Given that evidence showed that braces and orthoses might not be effective to improve pain and functional disability in general, these interventions might not be justifiable when patients also suffer from multiple complaints listed in the Lequesne index, such as nocturnal pain or difficulties of using stairs.

Conclusion

Our systematic review found no conclusive evidence about the effectiveness of any braces and orthoses for patients with medial knee OA. The use of short-lever elastic knee braces without hinges for symptomatic side or both sides might have potential clinical utility in specific cases. Future Japanese-language studies should address methodological flaws exposed in this review and strengthen the international evidence base.

Footnotes

Author contribution

All authors contributed equally in the preparation of this manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

References

Cross

Smith

Hoy

. The global burden of hip and knee osteoarthritis: estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis. Epub ahead of print 19 February 2014. DOI: 10.1136/annrheumdis-2013-204763.

Zhang

Jordan

. Epidemiology of osteoarthritis. Clin Geriatr Med 2010; 26(3): 355–369.

Bliddal

Christensen

. The treatment and prevention of knee osteoarthritis: a tool for clinical decision-making. Expert Opin Pharmacother 2009; 10(11): 1793–1804.

Kellgren

Lawrence

. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957; 16(4): 494–502.

Jordan

Helmick

Renner

. Prevalence of knee symptoms and radiographic and symptomatic knee osteoarthritis in African Americans and Caucasians: the Johnston County Osteoarthritis Project. J Rheumatol 2007; 34(1): 172–180.

Ding

Cicuttini

Boon

. Knee and hip radiographic osteoarthritis predict total hip bone loss in older adults: a prospective study. J Bone Miner Res 2010; 25(4): 858–865.

Tamm

Lintrop

Veske . Prevalence of patello- and tibiofemoral osteoarthritis in Elva, Southern Estonia. J Rheumatol 2008; 35(3): 543–544.

Janssen

Mark

. Separate and combined influence of body mass index and waist circumference on arthritis and knee osteoarthritis. Int J Obes 2006; 30(8): 1223–1228.

Szoeke

CEI

Cicuttini

Guthrie

. Factors affecting the prevalence of osteoarthritis in healthy middle-aged women: data from the longitudinal Melbourne Women’s Midlife Health Project. Bone 2006; 39(5): 1149–1155.

10.

Odding

Valkenburg

Algra

. Associations of radiological osteoarthritis of the hip and knee with locomotor disability in the Rotterdam Study. Ann Rheum Dis 1998; 57(4): 203–208.

11.

Alkan

Fidan

Tosun

. Quality of life and self-reported disability in patients with knee osteoarthritis. Mod Rheumatol 2014; 24(1): 166–171.

12.

Altman

. Early management of osteoarthritis. Am J Manag Care 2010; 16: S41–S47.

13.

Huber

Trattnig

Lintner

. Anatomy, biochemistry, and physiology of articular cartilage. Invest Radiol 2000; 35(10): 573–580.

14.

Garstang

Stitik

. Osteoarthritis: epidemiology, risk factors, and pathophysiology. Am J Phys Med Rehabil 2006; 85(11 Suppl.): S2–S11; quiz S12–S14.

15.

Scanzello

Goldring

. The role of synovitis in osteoarthritis pathogenesis. Bone 2012; 51(2): 249–257.

16.

Tetsworth

Paley

. Malalignment and degenerative arthropathy. Orthop Clin North Am 1994; 25(3): 367–377.

17.

Bruns

Volkmer

Luessenhop

. Pressure distribution at the knee joint. Influence of varus and valgus deviation without and with ligament dissection. Arch Orthop Trauma Surg 1993; 113(1): 12–19.

18.

Egloff

Hugle

Valderrabano

. Biomechanics and pathomechanisms of osteoarthritis. Swiss Med Wkly 2012; 142: w13583.

19.

Tanamas

Hanna

Cicuttini

. Does knee malalignment increase the risk of development and progression of knee osteoarthritis? A systematic review. Arthritis Rheum 2009; 61(4): 459–467.

20.

Hansson

Jonsson-Lundgren

Ronnheden

. Effect of an education programme for patients with osteoarthritis in primary care – a randomized controlled trial. BMC Musculoskelet Disord 2010; 11: 244.

21.

Gudbergsen

Boesen

Lohmander

. Weight loss is effective for symptomatic relief in obese subjects with knee osteoarthritis independently of joint damage severity assessed by high-field MRI and radiography. Osteoarthritis Cartilage 2012; 20(6): 495–502.

22.

Roddy

Zhang

Doherty

. Aerobic walking or strengthening exercise for osteoarthritis of the knee? A systematic review. Ann Rheum Dis 2005; 64(4): 544–548.

23.

Jansen

Viechtbauer

Lenssen

. Strength training alone, exercise therapy alone, and exercise therapy with passive manual mobilisation each reduce pain and disability in people with knee osteoarthritis: a systematic review. J Physiother 2011; 57(1): 11–20.

24.

Ulus

Tander

Akyol

. Therapeutic ultrasound versus sham ultrasound for the management of patients with knee osteoarthritis: a randomized double-blind controlled clinical study. Int J Rheum Dis 2012; 15(2): 197–206.

25.

Duivenvoorden

Brouwer

van Raaij

. Braces and orthoses for treating osteoarthritis of the knee. Cochrane Database Syst Rev, 2015, http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004020.pub3/abstract

26.

Carbone

Satterfield

Liu

. Assistive walking device use and knee osteoarthritis: results From the Health, Aging and Body Composition Study (Health ABC Study). Arch Phys Med Rehabil 2013; 94(2): 332–339.

27.

Gallelli

Galasso

Falcone

. The effects of nonsteroidal anti-inflammatory drugs on clinical outcomes, synovial fluid cytokine concentration and signal transduction pathways in knee osteoarthritis. A randomized open label trial. Osteoarthritis Cartilage 2013; 21(9): 1400–1408.

28.

Rutjes

AWS

Jüni

da Costa

. Viscosupplementation for osteoarthritis of the knee: a systematic review and meta-analysis. Ann Intern Med 2012; 157(3): 180–191.

29.

Law

Nguyen

Frank

. Current concepts on the use of corticosteroid injections for knee osteoarthritis. Phys Sportsmed 2015; 43(3): 269–273.

30.

Brouwer

van Raaij

Verhaar

. Brace treatment for osteoarthritis of the knee: a prospective randomized multi-centre trial. Osteoarthritis Cartilage 2006; 14(8): 777–783.

31.

Butler

Barrios

Royer

. Effect of laterally wedged foot orthoses on rearfoot and hip mechanics in patients with medial knee osteoarthritis. Prosthet Orthot Int 2009; 33(2): 107–116.

32.

Birmingham

Kramer

Kirkley

. Knee bracing for medial compartment osteoarthritis: effects on proprioception and postural control. Rheumatology 2001; 40(3): 285–289.

33.

Bureau

. Monthly report, http://www.stat.go.jp/english/data/jinsui/tsuki/index.htm (2015, accessed 22 July 2015).

34.

Muraki

Akune

Oka

. Incidence and risk factors for radiographic knee osteoarthritis and knee pain in Japanese men and women: a longitudinal population-based cohort study. Arthritis Rheum 2012; 64(5): 1447–1456.

35.

Muraki

Oka

Akune

. Prevalence of radiographic knee osteoarthritis and its association with knee pain in the elderly of Japanese population-based cohorts: the ROAD study. Osteoarthritis Cartilage 2009; 17(9): 1137–1143.

36.

Tonkin

. Language and the ingenuity gap in science. Crit Inq Lang Stud 2011; 8(1): 105–116.

37.

Schlosser

. Appraising the quality of systematic reviews (Technical Brief no. 17), http://www.ncddr.org/kt/products/focus/focus17/ (2007, accessed 22 July 2015).

38.

Toda

Segal

Kato

. Effect of a novel insole on the subtalar joint of patients with medial compartment osteoarthritis of the knee. J Rheumatol 2001; 28(12): 2705–2710.

39.

Toda

Tsukimura

. A six-month followup of a randomized trial comparing the efficacy of a lateral-wedge insole with subtalar strapping and an in-shoe lateral-wedge insole in patients with varus deformity osteoarthritis of the knee. Arthritis Rheum 2004; 50(10): 3129–3136.

40.

Toda

Tsukimura

. A 2-year follow-up of a study to compare the efficacy of lateral wedged insoles with subtalar strapping and in-shoe lateral wedged insoles in patients with varus deformity osteoarthritis of the knee. Osteoarthritis Cartilage 2006; 14(3): 231–237.

41.

Toda

Segal

. Usefulness of an insole with subtalar strapping for analgesia in patients with medial compartment osteoarthritis of the knee. Arthritis Rheum 2002; 47(5): 468–473.

42.

Toda

Tsukimura

. Influence of concomitant heeled footwear when wearing a lateral wedged insole for medial compartment osteoarthritis of the knee. Osteoarthritis Cartilage 2008; 16(2): 244–253.

43.

Raja

Dewan

. Efficacy of knee braces and foot orthoses in conservative management of knee osteoarthritis: a systematic review. Am J Phys Med Rehabil 2011; 90(3): 247–262.

44.

Malvankar

Khan

Mahapatra

. How effective are lateral wedge orthotics in treating medial compartment osteoarthritis of the knee? A systematic review of the recent literature. Open Orthop J 2012; 6: 544–547.

45.

Penny

Geere

Smith

. A systematic review investigating the efficacy of laterally wedged insoles for medial knee osteoarthritis. Rheumatol Int 2013; 33(10): 2529–2538.

46.

Parkes

Maricar

Lunt

. Lateral wedge insoles as a conservative treatment for pain in patients with medial knee osteoarthritis: a meta-analysis. JAMA 2013; 310(7): 722–730.

47.

Liberati

Altman

Tetzlaff

. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009; 339: b2700.

48.

Moher

Liberati

Tetzlaff

. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg 2010; 8(5): 336–341.

49.

Moseley

Herbert

Sherrington

. Evidence for physiotherapy practice: a survey of the Physiotherapy Evidence Database (PEDro). Aust J Physiother 2002; 48(1): 43–49.

50.

Maher

Sherrington

Herbert

. Reliability of the PEDro scale for rating quality of randomized controlled trials. Phys Ther 2003; 83(8): 713–721.

51.

De Morton

. The PEDro scale is a valid measure of the methodological quality of clinical trials: a demographic study. Aust J Physiother 2009; 55(2): 129–133.

52.

Shrout

. Measurement reliability and agreement in psychiatry. Stat Methods Med Res 1998; 7(3): 301–317.

53.

Landis

Koch

. The measurement of observer agreement for categorical data. Biometrics 1977; 33(1): 159–174.

54.

Toda

. Effect of elastically fixed lateral wedged insole on the subtalar joint for lower extremity alignment of patients with medial compartment osteoarthritis of the knee. Jpn J Rheum Joint Surg 2001; 20(2): 149–155.

55.

Toda

Kin

Tsukimura

. The effects of insoles with subtalar stabilizer for patients with medial knee osteoarthritis. Jpn J Phys Ther 2001; 35(11): 829–833.

56.

Toda

Tsukimura

. Effect of the insole with subtalar strapping for patients with varus deformity osteoarthritis of the knee: comparison with the placebo group. Bull Jpn Soc Prosthet Orthotic Educ Res Dev 2004; 20(1): 42–47.

57.

Deie

Ochi

Shibuya

. New knees functional brace and new type insole for medial osteoarthritis knee. J Jpn Orthop Surg Soc 2011; 85(10): 745–749.

58.

Toda

. The clinical efficacy of full-length shoe inserted lateral wedged insole combined with ankle sprain supporter for osteoarthritis of the knee. Jpn J Joint Dis 2014; 33(1): 7–12.

59.

Hochberg

Altman

Brandt

. Guidelines for the medical management of osteoarthritis. Part II. Osteoarthritis of the knee. American College of Rheumatology. Arthritis Rheum 1995; 38(11): 1541–1546.

60.

Toda

Tsukimura

. Is a shoe inserted lateral wedged insole combined with transverse arch pad in feet effective for osteoarthritis of the Knee? Orthop Surg Traumatol 2013; 56(6): 777–782.

61.

Toda

Tsukimura

Koji

. Efficacy of knee sleeves with medial compartment pad for patients with medial compartment osteoarthritis of the knee. Rinshouseikeigeka [Clin Orthop] 2012; 47(6): 597–602.

62.

Keyaki

Toda

Noriko

. The comparison of efficacy on ADL between two types of knee sleeves for the patients with osteoarthritis of the knee. Jpn J Phys Ther 2010; 44(11): 1027–1031.

63.

Toda

Tsukimura

Koji

. Comparison of clinical effects in patients with unilateral symptomatic osteoarthritis of the knee who wore bilateral soft orthotic devices or a device on the symptomatic knee only. Jpn J Joint Dis 2010; 29(2): 223–229.

64.

Toda

. An orthotic device for patients with osteoarthritis of the knee: the efficacy of a soft orthotic device for the knee with subtalar strapped insole. Jpn J Joint Dis 2009; 28(1): 93–99.

65.

Nakajima

Fukui

. Foot orthoses for knee osteoarthritis. Rheumatology 2003; 30(2): 114–119.

66.

Sasaki

Yasuda

. Clinical evaluation of the treatment of osteoarthritic knees using a newly designed wedged insole. Clin Orthop Relat Res 1987; 221: 181–187.

67.

Lequesne

Mery

Samson

. Indexes of severity for osteoarthritis of the hip and knee: validation – value in comparison with other assessment tests. Scand J Rheumatol Suppl 1987; 16(Suppl. 65): 85–89.

68.

Basaran

Guzel

Seydaoglu

. Validity, reliability, and comparison of the WOMAC osteoarthritis index and Lequesne algofunctional index in Turkish patients with hip or knee osteoarthritis. Clin Rheumatol 2010; 29(7): 749–756.

69.

Nadrian

Moghimi

Nadrian

. Validity and reliability of the Persian versions of WOMAC Osteoarthritis Index and Lequesne Algofunctional Index. Clin Rheumatol 2012; 31(7): 1097–1102.

70.

Faucher

Poiraudeau

Lefevre-Colau

. Algo-functional assessment of knee osteoarthritis: comparison of the test-retest reliability and construct validity of the WOMAC and Lequesne indexes. Osteoarthritis Cartilage 2002; 10(8): 602–610.

71.

Okuda

Omokawa

Okahashi

. Validity and reliability of the Japanese Orthopaedic Association score for osteoarthritic knees. J Orthop Sci 2012; 17(6): 750–756.

72.

Sullivan

Feinn

. Using effect size – or why the P value is not enough. J Grad Med Educ 2012; 4(3): 279–282.

73.

Thalheimer

Cook

. How to calculate effect sizes from published research: a simplified methodology. Work Learn Res 2002; 1–9.

74.

Rosnow

Rosenthal

. Computing contrasts, effect sizes, and counternulls on other people’s published data: general procedures for research consumers. Psychol Meth 1996; 1(4): 331–340.

75.

Gupta

. Intention-to-treat concept: a review. Perspect Clin Res 2011; 2(3): 109–112.

76.

Bastick

Runhaar

Belo

. Prognostic factors for progression of clinical osteoarthritis of the knee: a systematic review of observational studies. Arthritis Res Ther 2015; 17(1): 152.

77.

Gardner

Altman

. Confidence intervals rather than P values: estimation rather than hypothesis testing. Br Med J 1986; 292(6522): 746–750.

78.

Cohen

. A power primer. Psychol Bull 1992; 112(1): 155–159.

79.

Thornton

Lee

. Publication bias in meta-analysis: its causes and consequences. J Clin Epidemiol 2000; 53(2): 207–216.

80.

Moyer

Birmingham

Bryant

. Biomechanical effects of valgus knee bracing: a systematic review and meta-analysis. Osteoarthritis Cartilage 2015; 23(2): 178–188.