Reviews of the efficacy of acupuncture as a treatment for acute low back pain have concluded that there is insufficient evidence for its efficacy and that more research is needed to evaluate it. A multicentre randomized controlled trial was conducted at 4 primary-care centres in Spain to evaluate the effects of acupuncture in patients with acute nonspecific low back pain in the context of primary care. A total of 275 patients with nonspecific acute low back pain (diagnosed by their general practitioner) were recruited and assigned randomly to 4 different groups: conventional treatment either alone or complemented by 5 sessions over a 2-week period of true acupuncture, sham acupuncture, or placebo acupuncture per patient. Patients were treated from February 2006 to January 2008. The primary outcome was the reduction in Roland Morris Disability Questionnaire scores of 35% or more after 2weeks' treatment. The patients in the 3 types of acupuncture groups were blinded to the treatments, but those who received conventional treatment alone were not. In the analysis adjusted for the total sample (true acupuncture relative risk 5.04, 95% confidence interval 2.24-11.32; sham acupuncture relative risk 5.02, 95% confidence interval 2.26-11.16; placebo acupuncture relative risk 2.57 95% confidence interval 1.21-5.46), as well as for the subsample of occupationally active patients, all 3 modalities of acupuncture were better than conventional treatment alone, but there was no difference among the 3 acupuncture modalities, which implies that true acupuncture is not better than sham or placebo acupuncture.