OBJECTIVES: Comparative evidence evaluating the efficacy of oral osteoarthritis treatments has frequently included short-term trials. Meta-analyses of longer-term trials are needed. This meta-analysis of trials 12 weeks or longer was conducted to assess the efficacy of duloxetine vs. other oral treatments recommended after the use of acetaminophen for osteoarthritis, including non-steroidal anti-inflammatory drugs (NSAIDs) and opioids. METHODS: Search strategy: A systematic literature review was performed in PUBMED, EMBASE, MedLine In-Process, Cochrane Library, and ClinicalTrials.gov up to September 2011. Randomized controlled trials (RCTs) of duloxetine and all oral NSAIDs and opioids were included if the duration of treatment was twelve weeks or longer, the Western Ontario and McMaster Universities Index (WOMAC) total score was available, and they were published in English. Data collection and analysis: The WOMAC baseline and change from baseline total scores were collected and standardized. Twelve additional study characteristics were collected and study quality was assessed. A frequentist meta-analysis and indirect comparison were performed using the DerSimonian-Laird and Bucher methods. Bayesian analyses with and without study-level covariates were performed using noninformative priors. RESULTS: A search of the literature identified 24 studies which met inclusion criteria. The frequentist analysis and the Bayesian analysis without covariates found no statistically significant difference between the efficacy of duloxetine and the other treatments. Meta-regression suggested baseline scores explain much of the variance in change from baseline scores. The results, however, adjusted for study-level covariates led to the same conclusion. CONCLUSIONS: This analysis suggests that the efficacy of duloxetine in osteoarthritis, as measured by the WOMAC total score at 12 weeks or longer, is similar to competitor drugs.