Key message

For people with long-lasting low back pain without a known cause (chronic non-specific low back pain), after three months of doing yoga or not doing yoga, yoga is probably better than not doing exercise for improving pain and back-related function, although the improvements are small.

There is probably little or no difference between yoga and other types of back-focused exercise in improving back-related function, but we are uncertain about differences between yoga and other exercise for improving pain.

Back pain was the most common harm reported in yoga trials. Risk of harms was higher with yoga than with no yoga, but similar for yoga and other exercise. There was no suggestion that yoga was associated with a risk of serious harms.

What is non-specific low back pain?

Low back pain is a common health problem. In many cases, there is no known cause for the pain and it is termed 'non-specific' back pain. For some people, the pain may last for three months or more and at this point it is termed 'chronic.' Non-specific low back pain is usually treated with over-the-counter pain medicines and exercise, and does not require surgery or other invasive procedures. Yoga is sometimes used to help treat or manage low back pain.

What did we want to find out?

We wanted to find out if yoga improves function (for example, ability to walk, do jobs around the house, getting dressed), pain and quality of life associated with low back pain.

What did we do?

We searched medical databases for clinical trials comparing yoga practices using physical postures (often called 'hatha yoga') to any other treatment, sham (pretend) yoga, or to no treatment in adults (aged 18 years or older). We also included trials comparing yoga added to other treatments, versus those other treatments alone.

What did we find?

We included 21 trials with 2223 participants. Ten trials were carried out in the USA, five in India, two in the UK, and one each in Croatia, Germany, Sweden, and Turkey. Most participants were women in their 40s or 50s.

Key results

No trials compared yoga to sham yoga.

Ten trials compared yoga to non-exercise, which included usual care, delayed yoga treatment, or education (e.g. booklets and lectures). Six trials compared yoga to back-focused exercise or similar exercise programs. Five trials compared yoga, non-exercise, and another form of exercise.

At three months, there was low- to moderate-quality evidence that yoga was slightly better than no exercise in improving back function and pain, but the differences were not sufficiently important to the person with low back pain. There was low-quality evidence for more clinical improvement with yoga. There was moderate-quality evidence for a slight improvement in both physical (able to be active) and mental (emotional problems) quality of life and low-quality evidence for little to no improvement in depression.

At three months, there was moderate-quality evidence that there was little or no difference between yoga and other types of exercise in improving back function. Evidence was very-low quality for effects on pain at three months and we remain uncertain whether there is any difference between yoga and other exercise for pain. Evidence was also of very-low quality for clinical improvement and changes in physical and mental quality of life.

The most common harms reported in the trials were increased back pain. There was low-quality evidence that the risk of harms was higher with yoga than with non-exercise, and low-quality evidence that the risk of harms was similar between yoga and back-focused exercise. None of the trials reported yoga to be associated with a risk of serious side effects.

What are the limitations of the evidence?

Because we did not find any trials comparing yoga to sham yoga, we cannot say how yoga would affect low back pain if people did not know they were doing yoga. Participants in all the trials were aware of whether they were practicing yoga or not, and this may have influenced their interpretation of whether their back pain had changed. In addition, some trials were very small, there were few trials in some comparisons, and the trials in some comparisons had inconsistent results. Therefore, we downgraded the quality of the evidence to moderate, low, or very low.

How up to date is this evidence?

The evidence is current to August 2021.