Yoga for osteoarthritis of the hip or knee

Objectives

This is a protocol for a Cochrane Review (intervention). The objectives are as follows:

To assess the benefits and harms of yoga for osteoarthritis of the hip or knee, compared to other interventions or no intervention.

Background

Description of the condition

Arthritis is a category of chronic diseases characterised by symptoms of joint pain, stiffness and loss of flexibility, bony tenderness with bony enlargements, and deformity as the disease increases in severity (Senthelal 2020). These symptoms result from inflammation or tissue damage, or both. The pain and potential joint damage associated with arthritis translate into difficulty with a wide range of activities that limit one’s ability to work, perform routine daily activities, live independently, or participate in valued leisure and social activities, thus affecting multiple aspects of health‐related quality of life.

The most common form of arthritis is osteoarthritis (OA), which represents one of the leading causes of disability in older adults (Cross 2014). The prevalence of OA has increased due to the ageing population and increased levels of obesity. OA has been traditionally viewed as resulting from mechanical 'wear and tear' in weight‐bearing joints such as the knees, hips, and spine; but it also occurs in non‐weight bearing joints, including the fingers. The pathobiology of OA has been re‐conceptualised as a biomechanically‐initiated process, in which local and systemic factors propagate joint injury, with cartilage degradation, decreased repair mechanisms, and abnormal new bone formation. Pain, the predominant symptom in osteoarthritis, is multifactorial and reflective of processes at the joint level as well as peripheral nerve sensitisation and changes in brain structure and function, which further propagate and amplify pain perception (Clauw 2017). OA is a degenerative disease and joint pain and stiffness tend to increase over time.

There is no cure for OA, however it can be managed. Management strategies focus on over‐the‐counter pain medications, such as acetaminophen, together with exercise as a first approach to relieving pain and maintaining or improving function (Fernandes 2013; McAlindon 2014; Rausch 2018; Sinusas 2012). Opioid medications are not recommended because of the chronic nature of the disease and the possibility of developing tolerance and addiction. Surgical interventions, which are costly and not without risk, are reserved for circumstances where over‐the‐counter medication, corticosteroid injections, and core non‐pharmacological and educational interventions such as exercise are insufficient. Lifestyle management of OA (e.g. regular exercise, proper diet and sufficient sleep) is needed to reduce pain and improve both physical function and quality of life for those living with OA.

Description of the intervention

Yoga was first mentioned in text of ancient India as a unified state of the body and mind (Feuerstein 2001). Later, the term yoga was used to also include the set of practices that might foster such a state (Satchidananda 2012). While such practices originally consisted of concentration, meditation, breathing and lifestyle, the physical postures (asana) later developed as an aspect of yoga (Svatmarama 1992). As yoga took root in the West, asana played a more central role, often accompanied by intentional breathing and mental practices (Singleton 2010). While the ultimate goal of yoga may have been rooted in spiritual transcendence, there has also been recognition of the direct health benefits that yoga practices may provide, with most Americans practising yoga specifically for wellness (Barnes 2004).

Yoga has become increasingly popular around the world. For example, in the USA the percentage of people reporting using yoga during the past year increased from 5.1% in 2002 (Barnes 2004), to 6.1% in 2007 (Barnes 2008), 9.5% in 2012 (Clarke 2015) and 14.3% in 2017 (Clarke 2018). In a nationally representative Australian sample, 35% of young women and 27% of middle‐aged women reported practising yoga (Sibbritt 2011). Additionally, yoga is increasingly used as a strategy for managing pain and other chronic conditions, with 18% of those using yoga in the USA reporting that they use it to treat a specific clinical condition, the most common being back pain, stress, and arthritis (Stussman 2015). Furthermore, among American adults with arthritis, 7% use yoga (Zhang 2019). Unsurprisingly, rates of yoga use among arthritis patients in India are reported to be higher, at 34% (Jadhav 2011), while in Lebanon 5% of patients with OA reported ever using yoga (Alaaeddine 2012).

The pain, stiffness, loss of flexibility, and potential joint destruction associated with arthritis in general and OA in particular, translate into difficulty with a wide range of activities that limit one’s ability to work, live independently, and participate in leisure activities, resulting in lower levels of physical activity for those with OA compared to the general population (Cook 2007; Gay 2019; Holla 2014). Physical activity, including mind‐body approaches such as yoga, has therefore been recommended as an important component of OA management (Brosseau 2017). The symptoms and limitations of OA can impact stress and mood, with both presenting as common comorbidities among people with arthritis (Sharma 2016). Therefore, an intervention that may provide both physical activity and stress management strategies might provide useful adjunctive self‐care for OA.

The physical postures in yoga are easily adapted according to individual abilities and limitations using props and other modifications, allowing it to be readily applied to participants with OA (Moonaz 2015). Additionally, the breathing practices and mental strategies involved in yoga have been shown to alter the stress response (Benvenutti 2017; Gothe 2016; Schmalzl 2018) and thereby have application for pain management (Harth 2019; Somers 2009; Sorel 2019). The philosophical concepts of yoga, including present‐moment awareness, non‐harming of self and others, fostering contentment in the face of challenges, and honesty about one’s circumstance may also be useful skills for managing life with a chronic condition (Sullivan 2018).

How the intervention might work

There are several potential mechanisms by which yoga might impact patient‐reported outcomes in OA, such as pain and quality of life. These may include increased exercise, weight loss, reduced psychological stress and anxiety, and improvement in depressive symptoms.

Exercise is recommended for the management of OA to improve both stability and mobility by increasing strength, balance, co‐ordination, and joint mobility and improving flexibility (Beazley 2017; Brenneman 2015; Mazor 2018; McAlindon 2014; Schmid 2014; Yamamoto‐Morimoto 2019; Youkhana 2016). No specific type of exercise has been shown to be best (Juhl 2014). The isometric postures of yoga can help to strengthen major muscle groups necessary for ambulation and activities of daily living (Goncalves 2011). Yoga also emphasises balance challenges through standing postures on one or both feet, which can improve motor control, awareness, and postural stability (Youkhana 2016). As with other forms of physical activity, yoga may reduce falls or fear of falling, or both (Kumar 2016), which might otherwise lead to physical self‐limitations and decreased mobility. The range of motion and moving sequences of yoga may help to maintain joint mobility and reduce joint stiffness (Cheung 2014; Ebnezar 2012).

While the type of yoga recommended for OA is not vigorous and would not result in substantial weight loss due to increased caloric use, previous research has suggested that yoga’s mechanisms for weight loss might be through an increase in mindful behaviours, including mindful eating (Ross 2016; Ruffault 2017). Even small losses of 5% to 10% of body weight have been associated with a significant decrease in OA pain (Atukorala 2016; Christensen 2005; Christensen 2007); therefore, the weight loss resulting from mindful eating may support OA management.

Osteoarthritis is associated with psychological stress and impaired mental health, including depressive symptoms (Rabenda 2007; Sale 2008; Tak 2006; Yohannes 2010), especially in the presence of common co‐morbidities (Wesseling 2013). The practices of yoga have been associated with improved depressive symptoms (Cramer 2013), which are implicated in arthritis disability (Marks 2009). Treatment of depressive symptoms has therefore been shown to impact arthritis symptoms (Lin 2008).

Prolonged activation of the stress response can result in muscle tension, aches, pains, and increased arthritis symptoms (Evers 2014; Hannibal 2014; O'Connor 2000). Chronic pain reduces pain tolerance over time and is associated with the deterioration of grey matter in areas of the brain associated with emotional regulation (Bushnell 2013). The deep abdominal breathing instructed in yoga practice helps to engage the parasympathetic nervous system and thereby counteract the stress response. Additionally, meditation is associated with increased grey matter in the insular cortex (Hernandez 2018; Holzel 2011; Kang 2013), which regulates sensory and emotional experience.

One of the main philosophical tenets of yoga is non‐harming, which is ascribed to both self and others. This may help to facilitate more cautious engagement in physical activity during yoga practice compared to other exercise, and may impact other lifestyle decisions that affect OA progression. The principle of contentment with current circumstance could help to facilitate acceptance of disease states, gratitude with existing function, and a more positive outlook. The philosophy of yoga may therefore help to improve coping with OA regardless of disease severity and symptoms and therefore may be partially responsible for some of the mental health improvements outlined above. 

Why it is important to do this review

For people with arthritis, exercise may reduce disability and preserve autonomy while improving pain and depressive symptoms (Juhl 2014; Kelley 2015; Knapik 2018). Because adoption and maintenance of traditional exercise programmes is low, investigation of new options may offer alternative and potentially more desirable forms of physical activity.

Increasing interest in mind‐body approaches to physical activity, such as yoga, make the scientific study of their potential benefits timely and important. When combined with state‐of‐the‐art medical management, yoga may offer an opportunity for persons with arthritis to safely increase their physical activity and improve physical health, while potentially obtaining simultaneous psychosocial benefits.

There is empirical evidence for the safety and feasibility of yoga in a variety of populations as a part of the comprehensive management of several chronic conditions.. Previous Cochrane Reviews have been conducted to investigate yoga as an intervention for pain conditions. Most recently, an overview of Cochrane Reviews was conducted to assess the effects of exercise (including yoga) on chronic pain in adults; it found generally low‐quality evidence but favourable small‐to‐moderate effects on pain and physical function, and few adverse events (Geneen 2017). 

The investigation of yoga’s efficacy as part of a comprehensive approach to managing OA has previously been limited to a handful of studies that are small in size and scope. In 2017, a Cochrane Review reporting the results of five studies of exercise (one of which included yoga) for hand osteoarthritis found low‐quality evidence of small effects for exercise in improving hand pain, function and finger joint stiffness (Osteras 2017). Two earlier Cochrane Reviews on exercise for hip or knee osteoarthritis found high‐quality evidence that land‐based therapeutic exercise reduces pain, and high‐ to moderate‐quality evidence that exercise improves physical function; however yoga was not included in these reviews as an intervention (Fransen 2014; Fransen 2015). 

The interest in yoga as an intervention for OA has grown in recent years along with both the prevalence of OA and the practice of yoga, and a systematic review was published in 2019 (Lauche 2019). However, that systematic review only searched for studies through to early 2018, and there is a need for a Cochrane Review focused on yoga to update the evidence now and as it continues to develop in the future, as the evidence is not yet robust and the question continues to be of interest to patients and clinicians. This review will be conducted according to the guidelines recommended by the Cochrane Musculoskeletal Group Editorial Board (Ghogomu 2014).

Objectives

To assess the benefits and harms of yoga for osteoarthritis of the hip or knee, compared to other interventions or no intervention.

Methods

Criteria for considering studies for this review

Types of studies

We will include randomized controlled trials (RCTs). We will include studies reported as full text, those published as abstracts only, and unpublished RCT data. There will be no language restriction.

Types of participants

We will include adults (age 18 years or greater, with no upper age limit) with a diagnosis of osteoarthritis (OA) of the hip or knee. To be inclusive of small‐scale and low‐budget studies in this sparse research area, we will include studies in which participants were diagnosed with OA using any criteria, including but not limited to the American College of Rheumatology (ACR) criteria (Altman 1986; Altman 1991). Participants with either primary or secondary OA, of any degree, will be eligible. We will exclude participants reported to have comorbid conditions that affect the joints, such as rheumatoid arthritis. Studies including a mixture of eligible participants with hip or knee OA, and participants without hip or knee OA, will only be included if it is possible to extract the data separately for the participants with hip or knee OA.

Types of interventions

We will include studies reporting the use of any type of yoga incorporating physical practice (asanas), regardless of the duration, frequency, or intensity of the intervention. We will exclude studies limited to the non‐physical practice elements of yoga (e.g. philosophy). We will require that the study specify that the intervention is ‘yoga’ and will exclude studies of interventions based on yoga but not described as yoga. Yoga that is part of a multimodal programme in which the effect of yoga cannot be isolated (e.g. mindfulness‐based stress reduction (MBSR)) will be excluded.

We will include trials comparing yoga with any non‐yoga intervention, including no intervention or a waiting list or another active intervention. We will carry out the following comparisons.

• Yoga versus no active comparison (e.g. yoga versus no intervention, yoga versus a waiting list).

• Yoga versus active comparison (e.g. yoga versus another exercise programme), for which each type of active intervention is considered separately.

• Yoga plus an active intervention versus the same active intervention alone (e.g. yoga plus another exercise programme versus another exercise programme alone) for which each type of active intervention is considered separately

We will include trials providing cointerventions to participants, provided they are made equally available to both intervention groups and are not part of the randomized treatment.

Types of outcome measures

We will include studies in which yoga is tested as an intervention to improve the physical or mental health and functioning of people with hip or knee osteoarthritis. We will not exclude studies on the basis of reported outcomes. For this review, we will include the Cochrane Musculoskeletal recommended outcomes for OA (see: musculoskeletal.cochrane.org/proposed-outcomes). Because yoga is a mind‐body intervention sometimes used for anxiety and depression, and depression is common in OA (Sharma 2016), we will also include depression/anxiety as an outcome.

We will measure outcomes at short term (closest to four weeks), intermediate term (closest to three months), and long term (six months or longer).

Major outcomes

1. Pain: mean overall pain, measured on a pain rating scale, or pain subscores of composite scales if overall pain scales are not reported.

If data on more than one pain scale are provided for a trial, we will refer to the Cochrane Musculoskeletal hierarchy of recommended outcome measures (see: musculoskeletal.cochrane.org/proposed-outcomes) and extract data on the pain scale that is the highest on the following list.

1. Global pain.

2. Pain on walking.

3. Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain subscale.

4. Pain on activities other than walking.

5. WOMAC global scale.

6. Lequesne osteoarthritis index global score.

7. Other algofunctional scale.

8. Patient’s global assessment.

9. Physician’s global assessment.

10. Other outcome.

11. No continuous outcome reported.

2. Function: mean overall function, measured on a function rating scale, or function subscores of composite scales if overall function scales are not reported.

If data on more than one physical functioning measure are provided for a trial, we will refer to the Cochrane Musculoskeletal hierarchy of recommended outcome measures and extract data on the physical functioning measure that is the highest on the following list.

1. Global disability score.

2. Walking disability.

3. WOMAC disability subscore.

4. Composite disability scores other than WOMAC.

5. Disability other than walking.

6. WOMAC global scale.

7. Lequesne osteoarthritis index global score.

8. Other algofunctional scale.

3. Quality of life: mean quality of life, measured by a validated scale.

We could not find a recommended hierarchy of outcomes for quality‐of‐life measures in OA. If data on more than one quality‐of‐life outcome measure are provided for a trial, we will extract data on the quality of life measure that is the highest on the following list.

1. Osteoarthritis Knee and Hip Quality of Life (OAKHQOL).

2. Osteoarthritis Quality of Life questionnaire (OAQoL).

3. Arthritis Impact Measurement Scales (AIMS).

4. Mental component summary of the Medical Outcomes Study 36‐Item Short Form (SF‐36) or the 12‐item Short Form (SF‐12) Health Survey.

5. EuroQol.

6. General Well‐Being Index (GWBI).

7. Assessment of Quality of Life.

8. Other scale.

4. Depression/anxiety: mean depression or anxiety, as measured by a validated scale.

We could not find a recommended hierarchy of outcomes for depression or anxiety measures in OA, however there are several commonly used measures for depression in arthritis conditions (Smarr 2011). If data on more than one depression or anxiety outcome measure are provided for a trial, we will extract data on the measure that is the highest on the following list.

1. Hospital Anxiety and Depression Scale (HADS).

2. Beck Depression Inventory (BDI).

3. Center for Epidemiological Studies Depression Scale (CES‐D).

4. Other scale.

5. Total adverse events.

6. Withdrawals due to adverse events.

7. Serious adverse events.

Minor outcomes

1. Proportion of participants improved on treatment, as measured by a participant‐reported global impression of clinical change (much or very much improved), or similar measure (e.g. proportion with 30% improvement in pain).

2. Patient satisfaction with the intervention, as measured by the individual trial.

3. Withdrawals due to ‘lack of commitment’ or difficulty in practising yoga.

All outcomes will be reported for short‐term follow‐up (i.e. closest to four weeks after randomization) and, if available, intermediate follow‐up (i.e. closest to three months) and long‐term follow‐up (i.e. six months or more after randomization).

Search methods for identification of studies

Electronic searches

We will search the following databases.

• MEDLINE (PubMed, 1809 to present).

• Embase (Embase.com, 1974 to present).

• CINAHL (EBSCOhost, 1937 to present).

• Cochrane Central Register of Controlled Trials (Wiley).

• IndMED (indmed.nic.in 1985 to 2019). The database is currently discontinued.

• LILACS (Virtual Health Library, 1982 to present).

• PEDro (pedro.org.au, 1929 to present).

We will apply no restrictions on date or language of publication. Because preliminary searches indicate that the total number of non‐duplicate records retrieved across all databases will be fewer than 1000, we will not apply a filter for RCTs to our searches, to ensure maximum search sensitivity. Additionally, we will search ClinicalTrials.gov (www.ClinicalTrials.gov) and the WHO International Clinical Trials Registry Platform (www.who.int/ictrp/en/).

Searches will consist of a combination of text words, keywords and subject headings for two concepts: OA and yoga. Searches will be tailored for each database. See Appendix 1 for the search strategies for all online databases.

Searching other resources

We will check reference lists of all primary studies and review articles for additional references, and contact experts in the field. We will search for errata or retractions from included studies published in full text on PubMed (www.ncbi.nlm.nih.gov/pubmed) and report the date this was done within the review.

We will also check the available proceedings of the following conferences for relevant research.

• Symposium of Yoga Research (SYR).

• International Congress on Complementary Medicine Research (ICCMR) (2010 to 2020).

• European Congress for Integrative Medicine (ECIM) (2009; 2011 to 2020).

Data collection and analysis

Selection of studies

Two review authors (LSW, SM, or EC) will independently screen titles and abstracts of all of the potentially relevant studies we identify as a result of the search, and code them as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve'. We will retrieve the full‐text study reports/publications of eligible or potentially eligible studies, and two review authors (LSW, SM, or EC) will independently screen the full text and identify studies for inclusion. 

We will identify and record reasons for exclusion of all ineligible studies. We will resolve any disagreement through discussion or, if required, we will consult a third person (CB). We will identify and exclude duplicates and collate multiple reports of the same study so that each study, rather than each report, is the unit of interest in the review. We will record the selection process in sufficient detail to complete a PRISMA flow diagram (PRISMA Group 2009) and 'Characteristics of excluded studies' table.

Data extraction and management

We will use a data collection form for study characteristics and outcome data, which has been piloted on at least one study in the review. One review author (LSW) will extract study characteristics from included studies. A second review author (SM or EC) will spot‐check study characteristics for accuracy against the trial report. We will extract the following study characteristics.

1. Methods: study design, total duration of study, details of any 'run‐in' period, number of study centres and location, study setting, withdrawals, and date of study.

2. Participants: number (N), mean age, age range/variance, sex, race/ethnicity, location of OA (hip/knee), disease duration, severity of condition (as categorised and reported by the trialists), diagnostic criteria, inclusion criteria, and exclusion criteria.

3. Interventions: intervention details including the school of yoga and the major components (physical poses, meditation, breathing, relaxation) of the yoga intervention, comparison details, concomitant medications or other interventions, and excluded medications or other interventions.

4. Outcomes: primary and secondary outcomes specified and collected, and time points reported.

5. Characteristics of the design of the trial as outlined below in Assessment of risk of bias in included studies.

6. Notes: the source of trial funding and any other vested interests of the trialists, together with an assessment of potential influence upon trial design, conduct, or reporting stemming from such interests.

Two review authors (LSW, SM, or EC) will independently extract outcome data from included studies. We will extract the number of events and number of participants per treatment group for dichotomous outcomes, and means and standard deviations (or medians or variances if only these are reported) and number of participants per treatment group for continuous outcomes. We will note in the 'Characteristics of included studies' table if outcome data were not reported in a usable way and when data were transformed or estimated from a graph. We will use Plot Digitizer to extract data from graphs or figures (Jelicic 2016) and these data will also be extracted in duplicate. We will resolve any disagreements by consensus or by involving a third person (EC, SM, or CB). One review author (LSW) will transfer data into the Review Manager 5 (RevMan 2020) file. We will double‐check that data are entered correctly by comparing the data presented in the systematic review with the study reports.

There is no standard guidance on the choice to extract either final outcome data or change‐from‐baseline outcome data, or to extract either unadjusted or adjusted data. In this review, if both final values and change‐from‐baseline values are reported for the same outcome, we will extract final values and conduct sensitivity analyses using change‐from‐baseline values. If both unadjusted and adjusted values are reported for the same outcome within a study, we will extract the unadjusted values and conduct sensitivity analyses using adjusted values. When data are analysed based on an intention‐to‐treat (ITT) sample and another sample (e.g. per‐protocol, as‐treated), we will extract the data based on the ITT sample. We will extract data from the times that are closest to our definition of short term (i.e. four weeks), intermediate term (i.e. three months), and long term (i.e. six months).

Assessment of risk of bias in included studies

Two review authors (LSW, SM, or EC) will independently assess risk of bias from each study for each main review outcome using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020a). We will resolve any disagreements by discussion or by involving another author (CB). We will assess the risk of bias according to the following domains.

1. Bias arising from the randomization process.

2. Bias due to deviations from intended interventions.

3. Bias due to missing outcome data.

4. Bias in measurement of the outcome.

5. Bias in selection of the reported result.

For each domain, we will assign a judgement of 'low risk of bias', 'high risk of bias' or 'some concerns', and provide a quote from the study report together with a justification for our judgement in the 'Risk of bias' table. We will summarise the 'Risk of bias' judgements across different studies for each of the domains listed. We will also consider the impact of missing data by key outcomes.

Where information on risk of bias relates to unpublished data or correspondence with a trialist, we will note this in the 'Risk of bias' table. When considering treatment effects, we will take into account the risk of bias for the effect estimate from the studies that contribute to that result. We will present the figures generated by the 'Risk of bias' tool to provide summary assessments of the risk of bias.

Assessment of bias in conducting the systematic review

We will conduct the review according to this published protocol and report any deviations from it in the 'Differences between protocol and review' section of the systematic review.

Measures of treatment effect

We will analyse dichotomous data as risk ratios (RR) or Peto odds ratios when the outcome is a rare event (approximately less than 10%), and use 95% confidence intervals (CIs). Continuous data will be analysed as mean difference (MD) with 95% CIs when outcomes are measured on a common scale. We will enter data presented as a scale with a consistent direction of effect across studies.

When different scales are used to measure the same conceptual outcome (e.g. pain), a standardized mean difference (SMD) will be calculated instead of the MD, with corresponding 95% CI. SMDs will be back‐translated to a typical scale (e.g. 0 to 10 for pain) by multiplying the SMD by a typical among‐person standard deviation (e.g. the standard deviation of the control group at baseline from the most representative trial) (Schünemann 2020a).

In the Results section and the 'Summary of findings' table, we will provide the absolute per cent difference, the relative per cent change from baseline, and the number needed to treat for an additional beneficial outcome (NNTB), or the number needed to treat for an additional harmful outcome (NNTH) (the NNTB or NNTH will be provided only when the outcome shows a clinically significant difference). For dichotomous outcomes, the NNTB or NNTH will be calculated from the control group event rate and the relative risk using the Visual Rx NNT calculator (Cates 2008). The NNTB or NNTH for continuous measures will be calculated using the Wells calculator (available at the Cochrane Musculoskeletal editorial office).

For dichotomous outcomes, the absolute per cent change will be calculated from the difference in the risks between the intervention and control group using GRADEpro GDT (GRADEpro GDT 2015) and expressed as a percentage. The relative per cent change will be calculated as the RR minus 1 and expressed as a percentage. For continuous outcomes, the absolute benefit will be calculated as the improvement in the intervention group minus the improvement in the control group, in the original units, expressed as a percentage. The relative difference in the change from baseline will be calculated as the absolute benefit divided by the baseline mean of the control group, expressed as a percentage.

Unit of analysis issues

We do not expect to identify any cross‐over or cluster‐randomized trials, however if such trials are found we will include them and handle them according to the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020b). Where multiple trial arms are reported in a single trial, we will include only the relevant arms. If two comparisons (e.g. yoga versus waiting list and non‐yoga exercise versus waiting list) are combined in the same meta‐analysis, we will make it clear in the 'Characteristics of included studies' table that more than two intervention groups were present in the study, and halve the control group in our data analysis to avoid double‐counting.

Dealing with missing data

We will contact investigators or study sponsors in order to verify key study characteristics and obtain missing numerical outcome data where possible (e.g. when a study is identified as abstract only, or when data are not available for all participants). Where this is not possible, and the missing data are thought to introduce serious bias, we will explore the impact of including such studies in the overall assessment of results by a sensitivity analysis. Any assumptions and imputations to handle missing data will be clearly described and the effect of imputation will be explored by sensitivity analyses.

For dichotomous outcomes (e.g. number of withdrawals due to adverse events), the withdrawal rate will be calculated using the number of patients randomized in the group as the denominator. For continuous outcomes (e.g. mean change in pain score), we will calculate the MD or SMD based on the number of patients analysed at that time point. If the number of patients analysed is not presented for each time point, the number of randomised patients in each group at baseline will be used.

Where possible, missing standard deviations will be computed from other statistics such as standard errors, CIs or P values, according to the methods recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020c). If standard deviations cannot be calculated, they will be imputed (e.g. from other studies in the meta‐analysis) (Higgins 2020c).

Assessment of heterogeneity

Clinical and methodological diversity will be assessed in terms of participants, interventions, outcomes and study characteristics for the included studies to determine whether a meta‐analysis is appropriate. This will be conducted by observing these data from the data extraction tables. Statistical heterogeneity will be assessed by visual inspection of the forest plot to assess for obvious differences in results between the studies, and using the I² and Chi² statistical tests.

As recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017), the interpretation of an I² value of 0% to 40% might 'not be important'; 30% to 60% may represent 'moderate' heterogeneity; 50% to 90% may represent 'substantial' heterogeneity; and 75% to 100% represents 'considerable' heterogeneity. As noted in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), we will keep in mind that the importance of I² depends on the magnitude and direction of effects and the strength of evidence for heterogeneity.

Regarding the Chi² test, we will deem a P value of 0.10 or less to indicate evidence of statistical heterogeneity. If we identify substantial heterogeneity we will report it and investigate possible causes by following the recommendations in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017). 

Assessment of reporting biases

If we are able to pool more than 10 trials in a meta‐analysis, we will create and examine a funnel plot to explore possible small study biases. In interpreting funnel plots, we will examine the different possible reasons for funnel plot asymmetry as outlined in Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Sterne 2017) and relate this to the results of the review.

To assess outcome reporting bias, we will check trial protocols against published reports. For studies published after 1 July 2005, we will screen the Clinical Trial Register at the International Clinical Trials Registry Platform of the World Health Organization (https://trialsearch.who.int/Default.aspx) for the a priori trial protocol. We will evaluate whether selective reporting of outcomes is present.

Data synthesis

We will undertake meta‐analyses only where this is meaningful, i.e. if the participants, interventions, and outcomes are similar enough for pooling to be clinically appropriate. Our primary analysis will combine participants with hip and knee OA, and we will analyse hip and knee OA separately in subgroup analyses if data permit (Subgroup analysis and investigation of heterogeneity). We will use a random‐effects model because we expect the components and delivery characteristics of the yoga interventions to vary such that we will be estimating intervention effects that follow a distribution across studies (Deeks 2017).

The primary analyses will be restricted to trials for which there is either an overall low risk of bias for the outcome analysed, or some concern about bias for the outcome based solely on potential bias in the measurement of the outcome. The rationale for including trials for which there is some concern based on measurement of the outcome is that the outcomes in this review are assessed and reported by participants (e.g. outcomes such as pain, function, or health‐related quality of life); and because of the nature of the intervention we do not expect participants to be blinded to interventions. Thus, all self‐reported outcomes will likely be at some risk of bias in the measurement of the outcome.

Subgroup analysis and investigation of heterogeneity

We plan to carry out the following subgroup analyses, if data permit.

1. Yoga in hip OA versus yoga in knee OA, to investigate whether effects vary in osteoarthritis of different joints.

2. Yoga with only a physical practice versus yoga interventions with meditation, relaxation, or breathwork, to investigate whether effects vary when non‐physical practice components of yoga are provided to participants.

3. Yoga in older age groups (mean age of 65 years or more) versus younger age groups, to investigate whether the intervention is associated with different effects in older versus younger participants.

4. Yoga of different styles (Iyengar, etc.), to investigate whether effects vary depending on the style of yoga.

We will use the following outcomes in subgroup analyses.

1. Pain.

2. Physical function.

We will use the formal test for subgroup interactions in Review Manager 5 (RevMan 2020) and will use caution in the interpretation of subgroup analyses, as advised in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017). The magnitude of the effects will be compared between the subgroups by means of assessing the overlap of the CIs of the summary estimate. Non‐overlap of the CIs indicates statistical significance.

Sensitivity analysis

We plan to carry out the following sensitivity analyses to investigate the robustness of the treatment effect for the primary outcomes of pain and physical function.

1. All trials combined, regardless of their overall risk of bias.

2. Outcomes measured with change‐from‐baseline data.

3. Outcomes measured with adjusted values.

4. Trials in which participants fulfilled ACR diagnostic criteria.

Interpreting results and reaching conclusions

We will follow the guidelines for interpreting results provided in Chapter 15 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2020a), and will be aware of distinguishing a lack of evidence of effect from a lack of effect. We will base our conclusions only on findings from the quantitative or narrative synthesis of included studies for this review. We will avoid making recommendations for practice, and our implications for research will suggest priorities for future research and outline what the remaining uncertainties are in the area.

Summary of findings and assessment of the certainty of the evidence

We will create a 'Summary of findings' table using the following outcomes.

• Pain.

• Function.

• Quality of life.

• Depression/anxiety.

• Total adverse events.

• Withdrawals due to adverse events.

• Serious adverse events.

The comparison in the first table will be yoga versus no active intervention at short term. We will also create 'Summary of findings' tables for yoga versus active interventions at short term, and yoga added to an active intervention versus the active intervention alone at short term. We will create separate 'Summary of findings' tables for different active intervention comparators.

Two people (LSW, SM) will independently assess the certainty of the evidence. We will use the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of a body of evidence as it relates to the studies which contribute data to the meta‐analyses for the prespecified outcomes, and report the certainty of evidence as high, moderate, low, or very low. We will consider the following criteria for upgrading the certainty of evidence, if appropriate: large effect, dose‐response gradient, and plausible confounding effect. We will use methods and recommendations described in Chapters 8, 11, and 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020a; Schünemann 2020b; Schünemann 2020a). We will use GRADEpro GDT software to prepare the'Summary of findings' tables (GRADEpro GDT 2015). We will use version 3 of the GRADEPro view to display our 'Summary of findings' tables. We will justify all decisions to downgrade or upgrade our assessment of the quality of evidence using footnotes, and we will make comments to aid the reader's understanding of the review where necessary. We will provide the NNTB or NNTH, and the absolute and relative per cent change in 'Summary of findings' table, as described in Measures of treatment effect.

Notes

This protocol is based on a template developed by the Cochrane Musculoskeletal Group editorial base.

Appendices

Appendix 1. Search strategies

PubMed (pubmed.gov)

(yoga[tiab] OR yogic[tiab] OR meditat*[tiab] OR qigong[tiab] OR qi gong[tiab] OR tai ji[tiab] OR tai chi[tiab] OR asana*[tiab] OR yoga[mesh] OR meditation[mesh] OR qigong[mesh] OR “tai ji”[mesh])

AND

(osteoarthrit*[tiab] OR osteoarthro*[tiab] OR gonarthrit*[tiab] OR gonarthros*[tiab] OR coxarthrit*[tiab] OR coxarthros*[tiab] OR degenerative arthrit*[tiab] OR osteoarthritis[mesh:noexp] OR “osteoarthritis, hip”[mesh] OR “osteoarthritis, knee”[mesh])

Embase (embase.com)

(yoga:ab,ti OR yogic:ab,ti OR meditat*:ab,ti OR qigong:ab,ti OR ‘qi gong’:ab,ti OR ‘tai ji’:ab,ti OR ‘tai chi’:ab,ti OR asana*:ab,ti OR yoga/de OR meditation/de OR qigong/de OR ‘tai chi’/de)

AND

(osteoarthrit*:ab,ti OR osteoarthro*:ab,ti OR gonarthrit*:ab,ti OR gonarthros*:ab,ti OR coxarthrit*:ab,ti OR coxarthros*:ab,ti OR degenerative arthrit*:ab,ti OR osteoarthritis/de OR ‘hip osteoarthritis’/de OR ‘knee osteoarthritis’/de)

CINAHL (EBSCOhost)

(TI (yoga or yogic or meditat* or qigong or qi gong or tai ji or tai chi or asana*) OR AB (yoga or yogic or meditat* or qigong or qi gong or tai ji or tai chi or asana*) OR MH “yoga pose” OR MH “yoga” OR MH “meditation” OR MH “qigong” OR MH “tai chi”)

AND

(TI (osteoarthrit* or osteoarthros* or gonarthrit* or gonarthros* or coxarthrit* or coxarthros* or degenerative arthrit*) OR AB (osteoarthrit* or osteoarthros* or gonarthrit* or gonarthros* or coxarthrit* or coxarthros* or degenerative arthrit*) OR MH “osteoarthritis” OR MH “osteoarthritis, hip” OR MH “osteoarthritis, knee”)

CENTRAL (cochranelibrary.com)

(yoga or yogic or meditat* or qigong or qi gong or tai ji or tai chi or asana*):ti,ab,kw

AND

(osteoarthrit* or osteoarthros* or gonarthrit* or gonarthros* or coxarthrit* or coxarthros* or degenerative arthrit*):ti,ab,kw

IndMed (web.archive.org/web/20190220063501/http:/indmed.nic.in:80/)

(yoga or yogic or meditation or qigong or qi gong or tai ji or tai chi or asana or asanas)

AND

(osteoarthritis or osteoarthrosis or gonarthritis or gonarthrosis or coxarthritis or coxarthrosis or degenerative arthritis)

LILACS (Virtual Health Library)

yoga or yogic or meditation or qigong or qi gong or tai ji or tai chi or asana [Words]

AND

osteoarthritis or osteoarthrosis or gonarthritis or gonarthrosis or coxarthritis or coxarthrosis or degenerative arthritis [Words]

PEDro (pedro.org.au)

yoga AND osteoarthritis

Clinicaltrials.gov (www.ClinicalTrials.gov)

yoga OR yogic OR meditation OR qigong OR qi gong OR tai ji OR tai chi OR asana | osteoarthritis OR osteoarthrosis OR gonarthritis OR gonarthrosis OR coxarthritis OR coxarthrosis OR degenerative arthritis

WHO International Clinical Trials Registry Platform (www.who.int/ictrp/en/)

(yoga OR yogic OR meditation OR qigong OR qi gong OR tai ji OR tai chi OR asana) AND (osteoarthritis OR osteoarthrosis OR gonarthritis OR gonarthrosis OR coxarthritis OR coxarthrosis OR degenerative arthritis)

 

Contributions of authors

L Susan Wieland, Steffany Moonaz, Andrea G Shipper, Elise Cogo and Clifton Bingham III designed and wrote the protocol. All authors approved the final protocol.

Sources of support

Internal sources

• No sources of support provided

External sources

• National Institutes of Health (NIH), USA

  This project was supported by the National Center for Complementary and Integrative Health (NCCIH) grant R24 AT001293. The views and opinions expressed herein are those of the protocol authors and do not necessarily reflect those of the NIH or the NCCIH.

Declarations of interest

L Susan Wieland has no known conflicts of interest.
Steffany Moonaz is a yoga therapist who has received compensation for developing educational materials including online videos and books, as well as conference travel and research consulting related to the use of yoga for people with arthritis. She has no direct financial conflicts of interest and has authored a book in 2018 entitled Yoga Therapy for Arthritis: A Whole Person Approach to Movement and Lifestyle, and provides arthritis‐related continuing education to yoga professionals. 
Andrea G Shipper has no known conflicts of interest.
Elise Cogo has no known conflicts of interest.
Clifton Bingham III has no known conflicts of interest.

New