OBJECTIVES: To describe utilization patterns of duloxetine and celecoxib and subsequent opioid use among patients with osteoarthritis (OA) after duloxetine’s approval for OA. METHODS: Pharmacy and medical claims from SDI Health were analyzed for adult osteoarthritis patients (ICD-9-CM: 715.xx) who initiated duloxetine or celecoxib between 11/2010 and 4/2011. Initiation was defined as no pill coverage in 90 days prior. Included patients had continuous enrollment in 6 months before and after the initiation and did not use opioid in 90 days before the initiation. Propensity score matching was used to select patients with similar demographic and clinical characteristics for duloxetine and celecoxib cohorts. Compliance to index medication was assessed via medication possession ratio (MPR), proportion of days covered (PDC), and proportion discontinued (no access for 60 days). Initiating dose and opioid use after index date was assessed. Cox proportional hazard model was estimated to compare time to first opioid use. RESULTS: A total of 1360 patients initiated duloxetine and 1,408 patients initiated celecoxib. After matching, 784 patients were selected for duloxetine (mean age: 66, female: 78%) and celecoxib (mean age: 66; female: 76%) cohorts, respectively. 92.5% of duloxetine cohort started on ≤60mg/day, the recommended dose, and 72.8% of celecoxib cohort started on ≤200mg/day.     Duloxetine cohort had higher MPR and PDC and a lower proportion of discontinuation than celecoxib cohort (MPR: 0.81 vs. 0.70; PDC: 0.51 vs. 0.35; discontinuation:  57% vs. 78%. all p<0.001).  A lower proportion of duloxetine cohort used opioids after index date (48.6% vs. 68.5%, p<0.001), and started on opioid later than celecoxib cohort (mean: 132 vs. 107, p<0.001). After controlling for baseline characteristics, duloxetine cohort initiated opioids later than celecoxib cohort (Hazard ratio: 0.67, 95% confidence interval: 0.58-0.76). CONCLUSIONS: OA patients initiating duloxetine had better compliance and a lower likelihood of opioid utilization than those initiating celecoxib.