Exercise demonstrably works for chronic low back pain, with Pilates, McKenzie therapy and functional rehabilitation as the best-supported forms. The evidence comes from multiple large randomised studies, but the effects are modest and complete recovery is rarely the outcome. In practical terms: start with exercise and cognitive support, and keep medication as a reserve option.
An honest starting point: chronic low back pain (lasting longer than six months) is difficult to cure completely. In most people the exact cause cannot be established, and complete freedom from pain is unlikely. The goal of treatment is therefore more realistic: meaningfully reducing pain and being able to function better in daily life.
Exercise is the best-supported approach. Multiple international guidelines recommend exercise therapy as the first-choice treatment, with an average pain reduction of around 15 to 19 points on a scale of 0 to 100 compared with minimal treatment. A large network meta-analysis of 217 randomised studies involving more than 20,000 participants identified Pilates, McKenzie therapy (a specific exercise method focused on movement patterns) and functional rehabilitation as the most effective forms of exercise. Even so, the rule holds: consistency counts at least as much as the right technique. If someone does not feel drawn to Pilates but enjoys swimming or cycling, that is also an excellent choice.
One specific approach that stands out is motor skill training: exercises aimed at how a person carries out everyday activities such as getting up, bending or walking. In a randomised study of 154 participants, this person-centred method scored nearly eight points better on the disability scale after twelve months than standard strength and flexibility exercises. This effect was sustained at both six and twelve months.
People who do not benefit sufficiently from physical exercise may benefit from cognitive behavioural therapy or multidisciplinary rehabilitation. The American College of Physicians recommends this approach with moderate evidence as a fully valid alternative or supplement, including ahead of medication. Mindfulness-based stress reduction, yoga and tai chi are also evidence-supported options, particularly for people in whom tension or psychological stress plays a role in pain perception.
Acupuncture and spinal manipulation (chiropractic or manual therapy) are mentioned in guidelines, but the evidence for them is of lower quality. A meta-analysis of six studies directly compared exercise therapy with manual therapy and found no clinically meaningful difference. The quality of evidence in that analysis was, however, very low, so it is equally unjustified to dismiss manual therapy as clearly inferior.
Medication is not the first step. NSAIDs such as ibuprofen or naproxen are recommended as the first medication choice, but only when non-pharmacological treatment proves insufficient. Opioids such as strong painkillers are, in guidelines, reserved exclusively as a last resort after everything else has been tried, and only when the benefits outweigh the risks of addiction and side effects. This is a weak recommendation and requires an explicit discussion with a doctor.
Based on multiple international guidelines (including the American College of Physicians, PMID 28192789), a large network meta-analysis (217 RCTs, PMID 34538747), a systematic review with meta-analysis (PMID 40747709), an RCT (PMID 33369625), an overview review (PMID 38693474) and an epidemiological study (PMID 24994051). The total number of studies is difficult to count precisely due to overlap; conservatively estimated at a minimum of 20,000 participants through the network meta-analysis alone.