The Medical Research Council team said chiropractic, osteopathy or manipulative physiotherapy - plus an exercise programme was most effective.

Writing in the British Medical Journal, they say this three-pronged approach provides significant symptom relief.

They hope their findings will clarify how to treat back pain - one of the most common conditions seen by GPs.

More than 1,300 patients whose back pain had not improved through the "first line" advice of keeping active and not taking bed rest.

They filled in questionnaires on their general health, back pain, beliefs and psychological well-being before being randomly assigned to a treatment. Further questionnaires were completed after one, three and 12 months.

Cost effective

Patients were allocated a class-based physical exercise programme overseen by a physiotherapist, one form of spinal manipulation (such as physio), or a combination of the two.

People in all groups saw some improvement in their back function.

Those assigned to exercise classes in addition to GP care reported a small benefit at three months but not at one year.

Spinal manipulation plus GP care was linked to a small to moderate benefit at three months and a small, on average, benefit at one year.

But the biggest improvement was seen in those patients assigned to combined manipulation and exercise in addition to GP care.

They reported a moderate, on average, improvement at three months and a smaller average improvement at one year.

The researchers also examined the cost of the treatments and found both spinal manipulation and the combined package of spinal manipulation and exercise achieved similar benefits to patients at around the same cost to the NHS.

Martin Underwood, Professor of General Practice at Queen Mary, University of London, who worked on the study, said: "One of the commonest health problems I'm asked to treat as a GP is back pain.

"This trial has shown that physical treatments - a combined package of spinal manipulation and exercise in particular - may offer relief to the large numbers of patients whose back pain persists after they've received best care from their GP."

Professor Ian Russell, Director of the Institute of Medical and Social Care Research at the University of Wales Bangor, another member of the research team, said: "Though the average benefit to patients was only moderate in size, back pain is so widespread and so costly to society that these benefits add up to substantial economic benefits across the UK.

"Our analysis also showed that the cost to the NHS was modest.

"Added to GP 'best care', spinal manipulation gives the best value for money."