BACKGROUND: Clinical practice guidelines for non-specific low back pain do not recommend the use of non-rigid lumbar supports (NRLSs) despite the publication of several positive randomized controlled studies. OBJECTIVE: We conducted a systematic review with meta-analysis to assess the efficacy of NRLSs in the treatment and prevention of non-specific low back pain. METHODS: We searched for reports of randomized controlled trials in PubMed, Cochrane Library, EMBASE, Science Direct and Pedro databases. Data were analyzed by disease stage (acute, subacute, and chronic) and type of prevention (primary and secondary). The analysis of methodological quality involved the Physiotherapy Evidence Database (PEDro) scale. RESULTS: Of the 1581 records retrieved, only 4 full-text articles were included, with 777 patients: 378 in the NRLS group, and 348 in the control group. NRLSs conferred greater amelioration of disability (effect size -0.54, 95%CI -0.90; -0.17) and pain (-0.29, -0.46; -0.12) than standard management. Insufficient data prevented a comparison of the efficiency for acute, subacute and recurrent low back pain as well as meta-regression of responder phenotypes (sociodemographic and other patient characteristics). CONCLUSION: We demonstrated the overall efficacy of NRLSs for both disability and pain. However, further studies are needed to assess which patients can benefit the most from lumbar supports based on patient phenotype and the characteristics of low back pain. PROSPERO (CRD42018109855)