Randomised controlled trials (RCTs) that compared the effects of non-force spinal manipulation therapy or mobilisation versus control (no spinal manipulation or mobilisation; comparisons of different forms of these treatments) in adults with acute low back pain (pain lasting <12 weeks) were eligible for inclusion. Eligible trials were required to include at least 20 participants. The primary outcomes of interest were patient-reported pain reduction and functional improvement. Secondary outcomes were global effect, health care utilisation and adverse events. Trials with a follow-up less than one week were excluded.
The included studies were conducted between 1974 and 2006 in Australia, Canada, USA, UK and Italy. Patients were recruited from primary care settings, outpatient settings, physical therapy centres and an engineering company or unclear settings. Participant age ranged from 18 to 70 years. Exclusion criteria varied among trials. Where reported, mean duration of low back pain ranged from less than seven days to six months. Various health care professionals provided treatment, including general practitioners, osteopaths, chiropractors and physical therapists. Where reported, the number of spinal manipulation sessions varied from one to 20 and these were administered over a duration of up to 12 weeks. Some trials also included massage, exercise, education, medication, corset, diathermy and electrical stimulation. Trials included one to four control groups that received various treatments, mostly physical modalities (such as ultrasound, electrical nerve stimulation), exercise, education, medication, mobilisation and sham spinal manipulation therapy. Various outcome measures were used to assess pain.
Two reviewers screened studies for inclusion. Disagreements were resolved by consensus.