Thursday, November 3, 2011

Yoga Tops Usual Care for Low Back Pain

Yoga Chronic, nonspecific lower back pain can be a difficult condition to treat and often cannot be managed with analgesics alone. Appropriate exercise has proven beneficial for many patients, and yoga — with an emphasis on both physical and mental aspects of pain relief — may be particularly appealing for some patients. In a recently reported clinical trial, patients with low back pain benefitted from moderately increased functionality as a result of yoga instruction; however, added pain relief was not achieved and there were numerous limitations that weakened this research.

Researchers in the United Kingdom conducted a parallel-group, randomized, controlled trial to compare the efficacy of yoga versus “usual care” for chronic or recurrent low back pain [Tilbrook et al. 2011]. Participants included 313 adults (mean age 46 years, roughly 70% female) experiencing low back pain for roughly 10 years on average. They were randomly assigned either to “usual care” plus yoga (N=156) — a 75-minute, 12-class program on pain-reducing poses and how to improve posture and mobility — or to “usual care” alone (N=157). All participants also received a self-study booklet on back pain titled The Back Book [web version PDF here].

Score on the Roland–Morris Disability Questionnaire (RMDQ) at 3 months was the primary outcome measure and, secondarily, at 6 and 12 months. The low back pain version of the RMDQ contains 24 yes/no questions assessing such factors as disturbances in standing, walking, dressing, sleeping, and everyday activities due to pain. It does not assess the degree or quality of pain itself. Other secondary outcomes measured in this study included pain, pain self-efficacy, and general health measures at 3, 6, and 12 months.

At baseline, about 78% of subjects were currently experiencing back pain and 56% were using unspecified medications for this disorder. Mean prior duration of back pain was 113 months in the “usual care” group and 130 months in the yoga group; however, none of the subjects in either group were suffering severe disability due to their pain (eg, completely unable to perform usual activities or confined to bed).

Writing in the Annals of Internal Medicine, the researchers report that 93 (60%) subjects offered yoga attended at least 3 of the first 6 sessions and at least 3 other sessions. Compared with “usual care,” the yoga group had better back function at 3, 6, and 12 months: adjusted mean RMDQ score difference between groups was 2.17 points (95% Confidence Interval, 1.03 to 3.31 points) lower in the yoga group at 3 months, 1.48 points (CI, 0.33-2.62) lower at 6 months, and 1.57 points (CI, 0.42-2.71) lower at 12 months.

The yoga group had superior pain self-efficacy scores at 3 and 6 months but not at 12 months, indicating improved confidence in performing normal activities despite pain. Meanwhile, however, yoga and “usual care” groups had similar back pain and general health scores at all time points. Eight yoga participants experienced adverse events, largely pain, related to the intervention, while none of the “usual care” subjects reported increased pain due to their treatment.

The researchers conclude that offering a 12-week yoga program to adults with chronic or recurrent low back pain may confer improvements in functionality over “usual care.” However, a closer look at this study reveals that the results are somewhat clouded by unknown factors and the benefits are more modest than might be of consequence to many patients.

COMMENTARY: News media reports of this study praised the benefits of yoga for chronic low back pain without any reservations. However, as is typically the case, journalists most likely just parroted press releases touting the study and either did not read the full report or did not understand its subtleties from evidence-based pain management (EBPM) perspectives.

For starters, there are some unanswered questions and limitations of the study methodology…

  1. A critically undefined factor in this study is what “usual care” for back pain entails in this population. Since the authors say nothing of what “usual” means the reader cannot assess its appropriateness or effectiveness.

  2. Along with that, more than half of subjects were reported to be taking medications for back pain, but no other information is provided regarding who was taking these (which group), the type of medication, or the frequency, quantity, and timing of administration during the study. This factor alone might have confounded all outcomes.

  3. The purpose of also distributing a self-care booklet to all study participants and its effects are unclear. Perhaps, the investigators just wanted “usual care” subjects to have something added to their routine medical regimen to feel that they were part of a research trial; however, this adds another variable to the overall mix. Whether or not subjects actually read and applied the booklet’s recommendations is not reported.

  4. The authors use an Intention-to-Treat analysis of data, which takes into account missing data even for those subjects who were noncompliant with the study protocol. This provides a more “real world” or naturalistic perspective (eg, patients often do not exactly follow prescribed regimens); however, there was considerable missing data in both groups for the primary outcome and more data was missing in the yoga group than the “usual care” group for secondary outcomes. Although the researchers attempted to statistically adjust for this it usually adds bias and produces less confidence in outcome results.

  5. The study involved 12 yoga instructors delivering classes in 5 geographical areas. There was no analysis to determine whether particular instructors and/or locales made a difference in outcomes; however, these factors could account for variability in class attendance and response to the yoga intervention.

A closer inspection of some of the data evidence from this study reveals more of interest…

  • The figure below displays primary outcome RMDQ-score point-estimates with 95% Confidence Intervals (CIs) for each group throughout the study period (from Tilbrook et al.).
    Yoga Graph 
  • In the “usual care” group (top line of graph) it is apparent that during the full year of the study there were no significant improvements in RMDQ scores (ie, the CIs cross 0, except for a minor reduction at 6 months). Neither “usual care” — whatever that may have consisted of — nor the self-care back booklet appeared to be of any consequence for these patients in terms of improved functionality.

    In the yoga group (bottom line), the most gains were made during the first 3 months while classes were in session, and these persisted at the 6- and 12-month followup periods. All were significantly different from the “usual care” group (ie, CIs do not or only slightly overlap). Since yoga participants also received the booklet and “usual care” during this time, and such interventions produced no improvements in the other group, effects are most likely due to the yoga component itself.

  • Looking at change data for functionality, Tilbrook et al. claim that a difference from one time period to the next of 1.1 to 2.5 on the RMDQ is often considered as being clinically significant and important. In their study, compared with the “usual care” group, yoga participants benefitted from, on average, a 2.17 reduction (ie, fewer activity limitations) at 3 months, 1.48 reduction at 6 months, and 1.57 reduction at 12 months. All were statistically significant (P <0.011 to 0.001).

    However, RMDQ scores were not particularly high at the outset (mean=7.84 out of 24), and the maximum 27% decrease at the time classes concluded at 3 months may not have been of great consequence to many patients. Roughly, yoga participants had 2 fewer functional restrictions  on average than they did at baseline; however, of most importance might have been which specific activities improved. For example, being able to sleep more comfortably might be of more consequence to patients than being able to walk a bit faster, but the authors do not assess these qualities.

  • Furthermore, it should be noted that some have questioned the RMDQ as a good measure of clinically relevant changes in functionality. Julia Hush and colleagues, in Australia, conducted a series of focus groups among patients with back pain to assess the external validity of the RMDQ [Hush et al. 2010]. Participants concurred that this questionnaire did not capture all important functional domains, their complex personal experiences of pain, or relevant changes in their condition over time. The authors concluded that the RMDQ may contribute to misleading conclusions about back pain treatment efficacy and patient recovery.

  • It also might be of some concern in the Tilbrook et al. study that yoga did not produce improvements in back pain or general health scores at any time point. And, the improvements in self-efficacy scores during only the first 6 months suggests that yoga helped participants to better cope with or overcome their pain; possibly due to the breath control, relaxation techniques and mental focus afforded by yoga instruction. However, little detail on this self-efficacy measure is provided in the article and the authors note that there was a considerable amount of missing data for calculating the secondary outcome variables.

Although this study has considerable limitations and challenges to internal and external validity, it should not detract from the potential for exercise to be of some benefit for persons with chronic pain conditions. In a recent UPDATE [here] we discussed a study finding that the prevalence of chronic pain in patients exercising 1 to 3 times a week for at least 30 minutes was 10% to 38% less than in those not exercising.

Also recently, we reported on a study comparing yoga, conventional stretching exercises, and self-care (using a back pain booklet) [here]. Compared with the self-care group, both yoga and stretching groups showed significant and roughly equivalent improvements in functionality at 3 months, which suggested that the stretching component of yoga was of greatest benefit. Interestingly, the magnitude of functional improvements — approximately 1.7 to 2.5 points on a version of the Roland-Morris Disability Questionnaire — were similar to those in the present study by Tilbrook et al. Also similar, yoga produced a significant reduction in pain bothersomeness scores but not in pain itself.

Combined, the two studies of yoga for chronic back pain weakly suggest that this sort of exercise program in appropriately selected patients may provide both short- and longer-term benefits. However, yoga is not completely without risks of adverse effects, such as temporary increases in pain, and patients might expect only modest gains in functionality without associated reductions in pain. So, its best application might be as an adjunct to comprehensive back care that could include other non-drug therapies plus analgesics as necessary.

> Hush JM, Refshauge KM, Sullivan G, et al. Do Numerical Rating Scales and the Roland-Morris Disability Questionnaire capture changes that are meaningful to patients with persistent back pain? Clin Rehabil. 2010(Jul);24(7):648-657 [
abstract here].
> Tilbrook HE, Cox H, Hewitt CE, et al. Yoga for Chronic Low Back Pain: A Randomized Trial. Annals Intern Med. 2011(Nov);155(9):569-578 [
abstract here].