Yoga is safe for most people, and it helps back pain a littleStrong evidence
In a Cochrane review of 21 trials (2,223 people with back pain lasting 3 months or more), yoga improved back pain and everyday function a little more than not exercising, and about as much as other back exercise. In 1,702 German practitioners, injuries ran at about 0.6 per 1,000 hours of practice, which the authors judged as safe as, or safer than, other exercise. The US National Center for Complementary and Integrative Health (NCCIH) calls yoga generally safe for healthy people when done properly with a qualified instructor, and says serious injuries are rare. Fewer than 1 in 100 US adults who had ever done yoga stopped because of a side effect, most often back pain.
Sources 1, 2, 3, 4
A short rise in back pain is the commonest side effectEmerging evidence
In the back pain trials, about 43 in 1,000 people had a side effect with yoga over 6 to 12 months, mainly more back pain, against 9 in 1,000 who didn't exercise (low certainty), and about as many as with other back exercise (84 against 91 in 1,000). In one trial, side effects were mostly mild joint and back pain that settled on its own, in 9 of 127 people doing yoga and 14 of 129 having physical therapy; in another, 13 of 87 yoga and 13 of 75 stretching class attendees had a mild or moderate flare, mostly more back pain, and one yoga attendee had a slipped disc; in the UK trial, 12 of 156 people doing yoga reported side effects, mostly more pain, against 2 of 157 with usual care. The earlier version of the Cochrane review found yoga wasn't linked with serious side effects.
These were beginners' classes for people who already had back pain, so they show how often a flare happens when people start, not how often yoga causes new back pain. In vigorous Ashtanga practitioners, the low back was one of the three commonest injury sites, with the hamstring and the knee.
Sources 1, 5, 6, 7, 8, 9
Which poses, and what they ask of the backEmerging evidence
At a US clinic, of 89 people whose injuries were put down to yoga, the poses most often blamed involved bending the spine forward or arching it back to its extremes. Most had overuse muscle pain (74.2%); others had a flare of pain from existing joint wear (51.7%) or facet joint arthritis (38.2%), and 14.6% had a spinal compression fracture, which the authors link with thin bones. It was a clinic's series with no count of how many people did yoga safely. In the German survey, forward and backward bends were linked with 23.8% of acute problems; in an Australian survey, forward bends with 4.8% of injuries and backbends with 3.1%; and teachers and clinicians in a survey of 1,336 linked lower back injuries with forward bends, twists and backbends.
No lab study has measured the load on the lower back in yoga poses: a study of older adults holding yoga poses measured the ankle, knee and hip only. In our reading, the deep, held, repeated or forced versions are where backs complain: wheel, camel and bow; seated and standing folds with straight legs, especially when tight hamstrings leave the lower back to do the bending; and twists levered deeper with the arms. Very rarely, a very hard class strains the back muscles badly: one case report describes swollen back muscles and muscle breakdown after power yoga in a 36-year-old man, who recovered fully.
Sources 10, 2, 11, 12, 13, 14
Telling the causes apartEmerging evidence
A GP or physiotherapist asks which poses and movements bring it on, where it hurts and whether anything spreads down a leg, and checks how you bend in each direction. The NHS says it's not always obvious what causes back pain, and it often gets better on its own. Many backs have a sore direction and an easy one: pain on arching that eases bending forward, or the reverse. Leg symptoms below the knee point to a nerve. Pain at the sit bone in a forward fold is often the hamstring tendon, not the back. One-sided pain by the dimple of the pelvis may be the SI joint, a source in 15% to 30% of long-lasting back pain without leg symptoms in one review, though no single test can pin it down.
In a teenager, pain on arching is typical of a stress fracture of the spine: of 733 children and teenagers with back pain who had a magnetic resonance imaging (MRI) scan, 94.2% of those with a fresh stress fracture had pain arching back. But no examination test can confirm it: two reviews found the one-legged backbend test isn't reliable enough to rule it in or out.
Sources 15, 16, 17, 18, 19
When a scan or test mattersStrong evidence
For most back pain, the National Institute for Health and Care Excellence (NICE) says not to offer imaging routinely outside specialist settings. A young person with lasting pain on arching is the exception: a stress fracture can't be diagnosed by examination alone and needs a scan. A review of imaging in children takes two plain X-rays as the best first test and uses MRI for early diagnosis, and MRI picks up early stress injuries without radiation (moderate evidence). In an older adult with sudden back pain and thin bones, an X-ray looks for a compression fracture. Leg weakness, numbness or bladder or bowel changes need urgent assessment, as the warning signs say.
Sources 20, 21, 22, 23
How it usually goesEmerging evidence
The NHS says back pain usually improves within a few weeks, though it can last longer or come back. The flares in yoga trials were mostly mild and settled on their own, and people in those trials still ended up with better backs on average. In the German survey, 76.9% of acute yoga problems recovered fully.
A stress fracture of the spine is much commoner in young athletes who arch and twist a lot (found in 41.7% of athletes with back pain across 9 studies, against about 6% of adults), and it usually heals well without surgery: athletes returned to sport after about 3 to 4 months on average across one review, and about 9 in 10 got back to sport in a meta-analysis (another group: athletes in arching sports; we found no published yoga cases). No one can promise a date.
Sources 15, 5, 7, 2, 24, 21, 25, 26