|Home | About | Journals | Submit | Contact Us | Français|
The aim of this article was to systematically review the existing literature on the use of yoga for persons with arthritis. We included peer-reviewed research from clinical trials (published from 1980-2010) that used yoga as an intervention for arthritis patients and reported quantitative findings. Eleven studies were identified, including four RCTs and four NRCTs. All trials were small and control groups varied. No adverse events were reported and attrition was comparable or better than typical for exercise interventions. Evidence was strongest for reduction in disease symptoms (tender/swollen joints, pain) and disability, as well as improved self-efficacy and mental health. Interventions, research methods and disease diagnoses were heterogeneous. Larger, rigorous RCTs are necessary to more effectively quantify the effects of yoga for arthritic populations
Yoga includes a variety of theories and practices that originated in ancient India and have evolved and spread throughout the world. In Sanskrit, yoga means “to yoke” or connect 1. This typically refers to mind-body integration, but over the thousands of years that yoga has evolved, this focus has also been applied to spatial surroundings, nature, other individuals and spiritual interconnectedness 2. The physical practice of yoga, referred to as “hatha,” was originally intended to prepare for meditation, an important spiritual practice in many cultures. In recent decades, hatha yoga has become popular for physical activity and stress management. Other aspects of yoga, including study of ancient texts, dietary practices, acts of service, and moral living may be mentioned, but are not generally a focus of western classes.
After attention to posture, deep breathing and/or chanting, yoga practice often begins with a slow movement sequence to increase blood flow and warm muscles. This is followed by poses that include flexion, extension, adduction, abduction and rotation 1;3. Holding poses builds strength by engaging muscles in isometric contraction 4;5. Moving joints through their full range of motion increases flexibility 6;7, while standing poses promote balance by strengthening stabilizing muscles and improving proprioception to reduce falls 8;9. Thus, yoga incorporates several elements of exercise that may be beneficial for arthritis.
To cope with pain, arthritis patients often reduce activity 10;11. However, inactivity can result in muscle or tendon shortening, articular capsule contraction and weakened ligaments 12. Conversely, regular activity may decrease pain and preserve stability 13;14.
While there was once concern that exercise might increase inflammation and exacerbate pain, regular physical activity is now recommended as part of comprehensive treatment of arthritis 15-18. The American College of Rheumatology (ACR) 19, Osteoarthritis Research Society International (OARSI) 20 and the Ottawa Panel 21 note that stretching, strengthening and conditioning exercises can preserve physical function, increase strength and improve endurance for people with arthritis. All persons with arthritis should consult with their doctor to determine a safe and appropriate approach to increasing physical activity
Unfortunately, long term exercise maintenance is uncommon even for healthy individuals, generally approaching 50% after six months 22. Vigorous exercise is ideal for physical health 23 and may be acceptable for some persons with arthritis 24;25, but could be intolerable and may not be recommended for those with significant joint instability or damage 26;27. Adherence to moderate intensity exercise is more broadly tolerable, but still not attained by the majority of those with arthritis 28. For arthritis patients, emphasis on stretching and strength, posture, balance and the ability to adjust pace and intensity are important components of a safe activity, all of which yoga encompasses. Yoga is multifaceted, including focused breathing, mental engagement, stress management, social connection and/or meditative concentration along with physical activity. Yoga may offer an alternative to traditional exercise and potential psychological benefits or increased enjoyment for enhanced exercise adherence. Yoga could, therefore, provide another way for arthritis patients to be active and engaged in health-promoting behavior. Mind-body interventions such as yoga that teach stress management with physical activity may affect diseases from multiple fronts and may be well-suited for investigation in both OA and inflammatory immune-mediated diseases such as RA.
The goal of this review is to evaluate existing evidence regarding the effects of yoga practice on clinical, functional and psychosocial outcomes for people with arthritis.
Databases including MEDLINE, PsychLIT, PsychINFO and IndMed (an Indian database) were searched from 1980 through May 2010 for research trials that used yoga (including poses, breathing practices, relaxation and/or meditation) as an exercise intervention for arthritis patients. Additional relevant publications found in references from the original search list were also reviewed. Research in progress was searched via abstracts from annual scientific meetings of the American Public Health Association, American College of Rheumatology, Osteoarthritis Research Society International, European League Against Rheumatism and the International Association of Yoga Therapists. The following search terms were used: yoga OR yogic AND arthritis, arthritic, rheumatoid, rheumatic OR osteoarthritis. This review was limited to studies including quantitative statistical analysis and peer-review.
A total of 11 articles that described evaluating the effects of a yoga intervention in persons with arthritis were examined. One case series was excluded for lack of quantitative methods 29. Final analysis consisted of ten studies. (Table 1.) Five studies focused on RA, two were for OA only, and two included both RA and OA or arthritis in general. The studies were all published from 1980 to 2010.
Study quality was assessed based on: study design, sample size, intervention protocol and statistical analysis. Studies were classified as low, moderate or high for each. These criteria are based on categories set forth by the US Department of Health and Human Services in their 2002 report, “Systems to Rate the Strength of Scientific Evidence” 30. Funding source was not included as a category, since the majority of studies did not report a funding source, although the available information about funding is described. Since this review includes both randomized and observational trials, categories were adapted for both types. (Table 2 ).
Of the ten studies that were included in this review, four were RCTs 31-34, two compared people with arthritis to healthy controls, matching for age and sex 35;36, two utilized a non-randomized control (NRCT) 37;38, and two were cohort studies 39;40. Three of the RCTs had a waitlist control and the other two were usual care. The NRCTs assigned participants to control if they were unable to attend the first class session. None had an active control group. Six studies were reported as journal articles 31;35;37-40, one was a letter-to-the-editor (LTE) 34 and three were presented as abstracts at annual research meetings 32;33. One was presented as an abstract at an annual meeting, followed by publication in a journal that did not include a process of peer-review 36; therefore, only the abstract was included in this review.
Sample sizes ranged from three 29 to 26 37 intervention completers, with similar comparator group numbers. Only one study had as many as 20 persons per group 37. The necessary sample size to provide sufficient power to detect differences between groups was not generally described. Generally, a 10:1 subject to variable item ratio is recommended in multivariable regression analysis to avoid type I errors 42;43, although this depends on variable distribution 44;45.
Six studies reported on attrition, with rates of 0% 37, 9% 34, 22% 31, 36% 40 and 37% 36;39, with the three most rigorous studies having the lowest rates of attrition. The two cohort trials and one with healthy matched controls had the highest rates of attrition. The greatest retention was from the NRCT and two RCTs. Most studies analyzed data for completers only. Only one study reported the consideration of attrition in final analysis, excluding one dropout prior to baseline 34. Remaining studies did not report attrition 32;33;35;41.
Intervention protocols varied widely. The “dose” of yoga varied substantially between studies, and was often inconsistent within studies. For example, the study with the greatest dose included 120 minutes of practice 5 times per week, followed by once per week for three months with 10-30 minutes of daily home practice. In contrast, the lowest dose included 60 minutes once per week for eight weeks. Yet another study was only 15 days long, but included daily practice in a retreat setting. Some studies required daily home practice, some weekly, and some had no element of home practice. While many protocols were developed and/or taught by licensed or certified yoga professionals (teachers, therapists or scholars) some did not describe the intervention development or delivery. This is further complicated because requirements and regulation of yoga instruction differ by jurisdiction and culture, and credentials of the yoga professionals are not always standardized. Some studies used a style of yoga with a long history and published texts describing teaching methods and practice, while others developed a new protocol for the population under investigation. Some studies failed to describe the protocol in any detail.
Three studies, an RCT for hand OA 31, a cohort study for knee OA 40 and a NRCT of young adults with RA39, used an Iyengar-based yoga (IY) program. This style is known for utilizing props (blocks, straps, bolsters), adjusting to individual anatomy 1. The program for both OA studies was developed by one of the authors who is a senior certified yoga instructor, and the RA protocol was devised by an “experienced IY teacher” 39. The hand OA trial included 10 weeks of “stretching and strengthening exercises emphasizing extension and alignment, group discussion, supportive encouragement and general questions and answers” 31. Poses emphasized respiration and upper body alignment. The protocol is described generally with reference to a previous publication. The knee OA study described a 15 pose series and prop modifications, which could be easily replicated. The RA study by Evans et al listed examples of poses 39. The IY-based programs were 6 40, 8 31 and 10 31 weeks, meeting once or twice weekly for 60-90 minutes.
An NRCT for RA used a program developed by the authors in consultation with rheumatologists and a certified yoga therapist. This program, conducted by Badsha et al included “stretches, strengthening, meditation and deep breathing” in 6-weeks of biweekly classes 37. A healthy matched-control study by Dash and Telles included poses, breathing practices, meditation, lectures and “joint loosening exercises” in a 14-day yoga training camp 35. The RCT by Haslock and Ellis34 used gentle tailored poses, breath control, meditation, lectures and discussions with intention to soften emotions. For the first 3 weeks, 120-minute sessions were held 5 days per week, followed by weekly 120 minute sessions for 3 months.
Abstracts from an RCT of RA and OA used a gentle yoga program developed by rheumatologists, psychologists and a registered yoga therapist 32;33, incorporating poses, breathing practices, relaxation, meditation, chanting and supplemental reading. An age- and sex-matched control study taught a Social-Cognitive Theory-based Kundalini Yoga intervention to those “having been diagnosed with arthritis” 36. Kundalini Yoga emphasizes the spine, with a focus on raising energy and awareness 46. The study included poses, breathing techniques, meditation and relaxation. In two studies 34;40, the reader is referred elsewhere for description of the practice.
Well-validated instruments were administered by blinded assessors in 6 studies 29;31;34;35;40;41. These included anatomical changes, biomarkers, performance outcomes, and clinical assessment. Only one study used unmasked assessors, noting this as its greatest limitation 37.
All but two studies 34;36 measured baseline variables and outcomes recommended by ACR or OARSI 47;48. None used “sham yoga” to blind participants. Consequently, all self-report data suffers from possible expectation bias. However, the chosen self-report instruments are commonly used for persons with arthritis and known for strong psychometric properties.
One study with healthy controls created a new assessment tool to measure intervention efficacy and participant perceptions 36. The authors had previously used some of the questions in this population and demonstrated strong validity and reliability. Additions to the tool were checked for face and content validity by three academics.
Only two trials (a NRCT and cohort study) reported efforts to ensure that data characteristics supported the methods (such as assuming a normal distribution), and adjusted the statistical plan as necessary 37;40. Eight articles/abstracts described hypotheses up front and linked outcomes to those hypotheses. The other two listed feasibility as their primary outcome 34;36. However, some outcomes were not well-explained in the study’s context. For example, a study comparing RA patients to healthy controls hypothesized that yoga would result in increased strength 35. However, this study measured pre- and post-intervention NSAID dose, without assessing analgesic or other medication use and included no pain measures.
Of 8 possible points, studies ranged from 3-6 in overall study quality (Table 3). Future expansion from pilot studies and abstracts may include greater rigor. While available information is limited, the strongest studies can point toward associations that may be confirmed with additional trials.
Professional organizations have provided evidence-based recommendations for the use of particular outcomes for RA and OA. ACR suggests that trials of RA use the following measurement tools: tender joint count, swollen joint count, patient pain assessment, patient and physician global assessment of disease activity, patient assessment of physical function, laboratory evaluation of one acute phase reactant 47. For OA, OARSI recommends pain as the primary outcome, along with physical function and a patient global assessment 48. These can be measured with any tool that has adequate validity, reliability and responsiveness. They do not rule out the later addition of other outcomes, such as physician global assessment, HRQL, inflammation, stiffness, and time to surgery. While no study included all of the recommended outcomes, most included one or more. See Table 4 for study findings.
The Disease Activity Score (DAS) is an index developed to measure RA disease activity that has been extensively validated for use in clinical trials 49;50. It includes the number of tender and swollen joints along with the erythrocyte sedimentation rate or C-reactive protein and a patient assessment of disease activity. Two RA studies measured DAS-28 (which includes a 28 joint count) and both found statistically significant improvements for patients participating in the yoga intervention as compared to controls 33;37.
Two studies measured ring size as a marker of hand inflammation. Haslock and Ellis reported a trend toward statistical difference in ring size for persons with RA 34, while Garfinkel et al saw no change in ring size in persons with hand OA 31. A difference in anti-inflammatory medications for persons with RA could not be attributed to the intervention since the two groups differed at baseline 35.
Only one knee OA study reported on stiffness, but found no improvement 40, although they did report a trend toward improvement for global patient assessment. The hand OA trial saw improved finger tenderness and finger range of motion 51 and a study of general arthritis (diagnostic inclusion criteria unclear) used their own symptom self-report instrument with no improvements demonstrated 36.
Several studies assessed strength, balance, flexibility and/or mobility. Three used hand grip, which has been considered a clinical measure of general strength 52, hand function, pain, disease activity 53 and future disability 54. Improvements were found for two RA studies 34;35, but not in hand OA 31. A NRCT of post-menopausal women with RA showed improved balance 38, while the knee OA pilot found no change in 50 ft walk time 40.
Four of the five studies in RA used the Health Assessment Questionnaire (HAQ), a self-report of disability status, as an outcome measure. Two found significant improvement compared to controls or baseline 37;38 and another showed a trend toward improvement 34. The HAQ also includes a visual analog scale (VAS) of pain, which was used in three of the RA studies. Two demonstrated significant improvement 38;39, while one found no change 37. Although the HAQ is often considered to be a disease-specific instrument intended for use in RA, it has also been used more broadly and was included in a study for persons with hand OA 31. In this study, there was no change in the functional dimension of the HAQ, but pain by VAS did improve significantly; however, the HAQ is not as sensitive to changes in persons with OA 51.
The Arthritis Impact Measurement Scale 2 (AIMS2) and the Medical Outcomes Study Short Form-36 (SF-36) assess HRQL. The SF-36 is a general measure 55 while the AIMS2 is specifically designed for arthritis patients 56. Both contain mental and physical domains. The AIMS2 addresses unique issues of this population, but the SF-36 allows for comparison with non-arthritic participants.
Significant improvement in AIMS2 affect was seen for knee OA, with a trend toward improved symptoms and patient global assessment. Using the SF-36, one study found no improvements 37, while an abstract reported improved emotional roles and energy with a trend toward improved mental health 32.
Improvements in psychosocial health were captured using other instruments, including the Beck Depression Index 38. Two RA studies saw no changes in psychological health, measured by the General Health Questionnaire 34 and SF-36 37. Changes in cortisol levels, a common biomarker of psychological stress, was also measured with significant improvement in daytime measurements, and a trend for improved diurnal and awakening levels 38. Measures of pain included the Western Ontario and McMaster Osteoarthritis Index (WOMAC), a validated index for OA of the knee and hip, for a study of OA 40 and the Pain Disability Index (PDI) for a study of RA 38. Improvements in pain were statistically significant in both studies 39. Persons with arthritis who practiced Kundalini yoga reported increased self-efficacy and frequency of yoga behaviors 36.
Two studies required that no changes were made to treatment during the trial 34;40 and one RA study required stable dose of DMARDs and a limit to glucocorticoid use 38. Badsha et al. reported medication reductions for three persons with RA in the yoga group (3 corticosteroids, 1 etanercept, 2 methotrexate) “as a result of clinical improvement” and none in the control group 37. Dash and Telles noted a statistically significant reduction in NSAID use for the intervention group of persons with RA, though groups differed at baseline 35. Other studies did not report changes in medication or procedures to maintain stable medication use.
Most articles and abstracts did not disclose whether the study was funded. The study by Badsha et al. was funded by the Emirates Arthritis Foundation and by an unrestricted grant from Abbott Pharmaceuticals with no reported conflicts of interest 37. Research by Kolasinki et al was partially supported by American College of Rheumatology Clinical Summer Preceptorship Program 40. The LTE from Haslock and Ellis notes in acknowledgements that “Marks and Spencer contributed to the cost of data processing” 34. No other mention is made of funding support.
The assessment of yoga for arthritis is in its infancy. In general, the studies that are reported in the literature are very small in both size and scope. The use of recommended outcomes and validated measures was typical, but only a few outcomes were included in each study. Therefore, there is too little overlap in disease state and measured variables to pool data or draw preliminary conclusions.
HRQL is an important self-reported outcome that can inform about broad effects of interventions on several life domains. Few studies included HRQL, and none used it as a primary outcome. While instruments like the HAQ and WOMAC measure arthritis disability and its impact on daily activities, they do not assess eight domains of health, ranging from physical limitations to energy and mental health. Additionally, because they are primarily used in arthritis populations, this does not allow for comparison with health adults or other chronic conditions.
Study designs varied and each has drawbacks, including lack of masking, lack of control groups, group crossover, and biased group assignment. In these cases, limitations were often noted, but efforts made to reduce bias were not always explained. No study administered comparison treatment arms. This would strengthen findings, but requires a larger sample size and greater resources, which is a challenge in time-intensive, behavioral research trials 31.
‘Yoga’ describes a range of practices. While most studies described a comprehensive intervention (poses, breathing, relaxation and/or meditation), the styles, doses and format varied. Researchers must be clear about the delivered intervention, and that it is population appropriate. Determining which aspects of the practice are safe and beneficial can only follow when it is known what has been tested. Especially with patients who have considerable musculoskeletal limitations and symptoms, what is taught and how it is modified should be detailed in future research and practice. Beyond this, researchers should provide protocol transparency so that larger, more rigorous trials can replicate the interventions utilizing the same methods to confirm or dispute findings. Furthermore, when design methods and statistical analyses are not well-described, research cannot be properly evaluated and readers are unable to determine whether methodological flaws may be responsible for error in findings.
For classroom-based interventions, it may be challenging to recruit participants willing to travel and dedicate several hours per week for months, especially with unpredictable, painful diseases. Understudied interventions are often limited to the safest and healthiest (by age and/or disease status) to first ensure no harm before expanding to vulnerable populations. This reduces qualifying participants, particularly for a rarer disease (such as RA). However, this can bias the sample and limit generalizability to all persons with the disease.
Arthritis encompasses many diagnoses. The two most common, OA and RA, have significant pathophysiologic differences, and effects of an intervention such as yoga may not be the same for each. Moreover, the effects of yoga on isolated hand versus knee OA may also have significantly different outcomes. Combining persons with different diseases in the same intervention and analyzing the data collectively could mask an effect that is strong for just a portion of participants, or could suggest a universal effect that only applies to a subgroup with one particular form of arthritis. The use of biomarkers as treatment outcomes will improve current understanding of how additional biochemical and pathophysiological parameters of diseases respond to interventions.
The research in this review was conducted in diverse populations across the globe, which suggests broad interest, as well as cross-cultural acceptance. However, perceptions of yoga, teaching methods and differences in arthritis treatment are likely to influence intervention effects and, possibly, results in different findings.
Overall, the most consistent findings were for tender orswollen joints in persons with RA, which improved for all three studies that used this outcome. Another common outcome was pain, which improved in six out of eight studies, measured by various instruments. Disability improved in three out of four studies. Self-efficacy improved in both studies for which it was measured. Mental health and energy improved for two out of three studies. For grip strength, improvements were seen in both studies of RA, but not the study of OA that included it. Results for global health and physical functioning were inconsistent. Measures of disease symptoms and physical functioning were more commonly used than markers of physical fitness or psychosocial functioning. Because different instruments were often used to assess the same outcome, interpretation of results across studies is compromised.
A goal of future studies should be to create standardized protocols that are optimized to enhance safety, enjoyment and long-term adherence (including specific poses and modifications). Studies have varied considerably with regard to the frequency and duration of yoga practice, as well as the style and specific class content. Authors need to thoroughly describing the practice studied, including specifying the yoga lineage (Iyengar, Kundalini, etc.) in the methods section or separately publishing intervention details.
Interdisciplinary collaboration in the design of yoga interventions is appropriate for this population. Yoga experts, rheumatology clinicians or clinical researchers alone are not equipped to create an authentic and appropriate yoga program without consultation with each other. Such a program requires careful attention to the stress on joints and connective tissue, as well as the consideration of joint range of motion and adaptation during potential disease flares. An arthritis-appropriate program that has been created in careful collaboration and well-tested through rigorous research methods is required as a next step in the evolution of this research.
It would be of significant interest to study the mechanisms by which yoga affects arthritis symptoms. The use of standardized outcome measures and appropriate statistical methods are essential for confirming findings. The field awaits large, comprehensive trials that may validate improvements indicated by this collection of small pilot studies.
This work was supported by Grant No. F31 AT003362-01A1 from the National Institutes of Health and a Doctoral Dissertation Award from the Arthritis Foundation.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The authors have nothing to disclose.