BACKGROUND: Osteoarthritis (OA) is common and burdensome for patients and health care systems. Our study purpose was to evaluate the long-term efficacy and safety of DMOADs in adults with knee and hip osteoarthritis. METHODS: We searched Cochrane Central Register of Controlled Trials, MEDLINE, EMBASE, and Web of Knowledge without language, publication, or date restrictions from inception through November 2018 for randomized controlled trials assessing 12 classes of DMOADs with at least 12 months of follow-up. Therapeutic effects were evaluated with pairwise and network meta-analysis. Outcomes included pain, function, minimum joint space width or cartilage volume, radiographic progression, and total joint replacement. Analyses were also performed for drug safety. RESULTS: Twenty-eight randomized controlled trials with 11,890 patients were included. Glucosamine and chondroitin minimally improved both structure (minimum joint width or cartilage volume: network results: glucosamine: SMD 0.16; 95% CI [0.04, 0.28], chondroitin: SMD 0.21 [0.10, 0.32]) and symptoms (glucosamine: pain: - 0.15 [- 0.25, - 0.05]; function: - 0.17 [- 0.28, - 0.07], chondroitin: pain: - 0.06 [- 0.15, 0.03], and function: - 0.15 [- 0.26, - 0.03]). Strontium demonstrated improvement in structure (minimum joint width or cartilage volume: 0.20 [0.02, 0.38]), and vitamin D on symptoms (pain: - 0.15 [- 0.27, -0.03]; function: - 0.18 [- 0.31, - 0.06]). Although doxycycline also demonstrated a favorable efficacy ranking, its safety profile was poor (withdrawal: network relative risk 1.69 [1.03, 2.75]). The therapeutic effects of other medications were not ranked as highly. DISCUSSION: Glucosamine and chondroitin yielded statistically significant but clinically questionable long-term benefit on structure and symptoms, though both had favorable safety profiles. Strontium improved structure, and vitamin D improved symptoms. Although doxycycline had a favorable efficacy ranking, its safety profile was poor. None of the 12 classes of drugs appears to have long-term clinically significant benefit.