Yoga for chronic non-specific low back pain
L Susan Wieland,Nicole Skoetz,Karen Pilkington,Shireen Harbin,Ramaprabhu Vempati,Brian M Berman
DOI: https://doi.org/10.1002/14651858.CD010671.pub3
2022-11-18
Abstract:Background: Non-specific low back pain is a common, potentially disabling condition usually treated with self-care and non-prescription medication. For chronic low back pain, current guidelines recommend exercise therapy. Yoga is a mind-body exercise sometimes used for non-specific low back pain. Objectives: To evaluate the benefits and harms of yoga for treating chronic non-specific low back pain in adults compared to sham yoga, no specific treatment, a minimal intervention (e.g. education), or another active treatment, focusing on pain, function, quality of life, and adverse events. Search methods: We used standard, extensive Cochrane search methods. The latest search date was 31 August 2021 without language or publication status restrictions. Selection criteria: We included randomized controlled trials of yoga compared to sham yoga, no intervention, any other intervention and yoga added to other therapies. Data collection and analysis: We followed standard Cochrane methods. Our major outcomes were 1. back-specific function, 2. pain, 3. clinical improvement, 4. mental and physical quality of life, 5. depression, and 6. Adverse events: Our minor outcome was 1. work disability. We used GRADE to assess certainty of evidence for the major outcomes. Main results: We included 21 trials (2223 participants) from the USA, India, the UK, Croatia, Germany, Sweden, and Turkey. Participants were recruited from both clinical and community settings. Most were women in their 40s or 50s. Most trials used iyengar, hatha, or viniyoga yoga. Trials compared yoga to a non-exercise control including waiting list, usual care, or education (10 trials); back-focused exercise such as physical therapy (five trials); both exercise and non-exercise controls (four trials); both non-exercise and another mind-body exercise (qigong) (one trial); and yoga plus exercise to exercise alone (one trial). One trial comparing yoga to exercise was an intensive residential one-week program, and we analyzed this trial separately. All trials were at high risk of performance and detection bias because participants and providers were not blinded to treatment, and outcomes were self-assessed. We found no trials comparing yoga to sham yoga. Low-certainty evidence from 11 trials showed that there may be a small clinically unimportant improvement in back-specific function with yoga (mean difference [MD] -1.69, 95% confidence interval [CI] -2.73 to -0.65 on the 0- to 24-point Roland-Morris Disability Questionnaire [RMDQ], lower = better, minimal clinically important difference [MCID] 5 points; 1155 participants) and moderate-certainty evidence from nine trials showed a clinically unimportant improvement in pain (MD -4.53, 95% CI -6.61 to -2.46 on a 0 to 100 scale, 0 no pain, MCID 15 points; 946 participants) compared to no exercise at three months. Low-certainty evidence from four trials showed that there may be a clinical improvement with yoga (risk ratio [RR] 2.33, 95% CI 1.46 to 3.71; assessed as participant rating that back pain was improved or resolved; 353 participants). Moderate-certainty evidence from six trials showed that there is probably a small improvement in physical and mental quality of life (physical: MD 1.80, 95% CI 0.27 to 3.33 on the 36-item Short Form [SF-36] physical health scale, higher = better; mental: MD 2.38, 95% CI 0.60 to 4.17 on the SF-36 mental health scale, higher = better; both 686 participants). Low-certainty evidence from three trials showed little to no improvement in depression (MD -1.25, 95% CI -2.90 to 0.46 on the Beck Depression Inventory, lower = better; 241 participants). There was low-certainty evidence from eight trials that yoga increased the risk of adverse events, primarily increased back pain, at six to 12 months (RR 4.76, 95% CI 2.08 to 10.89; 43/1000 with yoga and 9/1000 with no exercise; 1037 participants). For yoga compared to back-focused exercise controls (8 trials, 912 participants) at three months, we found moderate-certainty evidence from four trials for little or no difference in back-specific function (MD -0.38, 95% CI -1.33 to 0.62 on the RMDQ, lower = better; 575 participants) and very low-certainty evidence from two trials for little or no difference in pain (MD 2.68, 95% CI -2.01 to 7.36 on a 0 to 100 scale, lower = better; 326 participants). We found very low-certainty evidence from three trials for no difference in clinical improvement assessed as participant rating that back pain was improved or resolved (RR 0.97, 95% CI 0.72 to 1.31; 433 participants) and very low-certainty evidence from one trial for little or no difference in physical and mental quality of life (physical: MD 1.30, 95% CI -0.95 to 3.55 on the SF-36 physical health scale, higher = better; mental: MD 1.90, 95% CI -1.17 to 4.97 on the SF-36 mental health scale, higher = better; both 237 participants). No studies reported depression. Low-certainty evidence from five trials showed that there was little or no difference between yoga and exercise in the risk of adverse events at six to 12 months (RR 0.93, 95% CI 0.56 to 1.53; 84/1000 with yoga and 91/1000 with non-yoga exercise; 640 participants). Authors' conclusions: There is low- to moderate-certainty evidence that yoga compared to no exercise results in small and clinically unimportant improvements in back-related function and pain. There is probably little or no difference between yoga and other back-related exercise for back-related function at three months, although it remains uncertain whether there is any difference between yoga and other exercise for pain and quality of life. Yoga is associated with more adverse events than no exercise, but may have the same risk of adverse events as other exercise. In light of these results, decisions to use yoga instead of no exercise or another exercise may depend on availability, cost, and participant or provider preference. Since all studies were unblinded and at high risk of performance and detection bias, it is unlikely that blinded comparisons would find a clinically important benefit.