In recent years, long-term prescribing and use of strong opioids for chronic non-cancer pain (CNCP) has increased in high-income countries. Yet existing uncertainties, controversies and differing recommendations make the rationale for prolonged opioid use in CNCP unclear. This systematic review and meta-analyses (MAs) compared the efficacy, safety and tolerability of strong opioids with placebo/non-opioid therapy in CNCP, with a special focus on chronic low back pain (CLBP). Systematic literature searches were performed in four electronic databases (Medline, Web of Science, Cochrane Library and CINAHL) in July 2019 and updated by regular alerts until December 2020. We included 16 placebo-controlled RCTs for CLBP and five studies (2 RCTs and 3 non-randomized studies) of opioids vs non-opioids for CNCP in the quantitative and qualitative synthesis. Random effects pairwise MAs were performed for efficacy, safety and tolerability outcomes and subgroup analyses for treatment duration, study design, and opioid experience status. Very low- to low-certainty findings suggest that 4-15 weeks (short-/intermediate term) opioid therapy in CLBP (compared to placebo) may cause clinically relevant reductions in pain but also more gastrointestinal and nervous system adverse events (AEs), with likely no effect on disability. In contrast, long-term opioid therapy (≥ 6 months) in CNCP may not be superior to non-opioids in improving pain or disability/pain-related function, but seems to be associated with more AEs, opioid abuse/dependence, and possibly an increase in all-cause mortality. Our findings also underline the importance and need for well-designed trials assessing long-term efficacy and safety of opioids for CNCP and CLBP.