ABSTRACT: Purpose:To assess long-term safety, tolerability, and efficacy of duloxetine in Japanese patients with chronic knee pain due to osteoarthritis. Methods:In this open-label extension study (NCT02335346), Japanese patients with knee osteoarthritis and pain (Brief Pain Inventory [BPI] - Severity average pain score ?4 at start of randomized trial) who had previously received duloxetine 60 mg/day or placebo for 14 weeks in a double-blind randomized trial entered the extension and received duloxetine 60 mg/day for 48 weeks. The primary outcome was safety/tolerability, secondary outcomes were change in BPI-Severity (BPI-S) average pain, BPI-Interference (BPI-I), Patient Global Impression-Improvement (PGI-I), Clinical Global Impression-Improvement (CGI-I), 36-item Short-Form Health Survey (SF36), and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), and exploratory outcomes were knee range of motion (efficacy outcome) and Kellgren-Lawrence grade (safety outcome). Results:Of 323 patients who completed the randomized trial, 93 (50 placebo, 43 duloxetine) entered the extension. Most patients (85, 91.4%) experienced an adverse event, most commonly constipation, nasopharyngitis, somnolence, and dry mouth (?10% of patients). There were eight serious adverse events in seven patients and no deaths. No obvious duloxetine-related changes were observed in laboratory tests, vital signs, or electrocardiograms. The change from baseline in BPI-S average pain score was significant throughout the extension. Significant reductions in BPI-I, PGI-I, CGI-I, WOMAC, and SF36 scores were also maintained through 52 weeks. There were no substantial changes in range of motion or Kellgren-Lawrence grade. Conclusion:In Japanese patients with chronic knee pain due to osteoarthritis, long-term treatment with duloxetine was well tolerated and associated with sustained improvements in pain and health-related quality of life without radiographic deterioration.