Objective: To evaluate the efficacy and safety of botulinum toxin type A (BoNT-A) injected intra-articularly in 60 subjects with moderate pain and functional impairment secondary to knee osteoarthritis. The study investigators hypothesized that intra-articular BoNT-A would result in statistically significant improvements in pain and function at 8 weeks. Design: Double-blind, randomized, single tertiary care academic medical center trial with 6-month follow-up. Patients: Sixty patients aged 40 years or older with painful osteoarthritis of the knee who had failed physical therapy, medications, and/or injection therapy presenting to the musculoskeletal or orthopedic outpatient clinics at a large tertiary care medical institution. All 60 patients completed 8-week follow-up, but only 32 patients completed the 26-week follow-up. Methods: Subjects were randomized to receive a single injection of corticosteroid, low-dose BoNT-A (100 units), or high-dose BoNT-A (200 units). Outcome measures were compared at baseline, 4, 8, 12, and 26 weeks after injection. Main Outcome Measurements: The primary outcome measure was pain visual analog scale (VAS) at 8 weeks. Secondary outcome measures included Western Ontario McMaster Arthritis Index, Short Form-36 scores, patient global assessment, 40-meter timed walk, and adverse effects. Results: The primary end point was pain VAS score at 8 weeks, which decreased within each group but only reached statistical significance in the low-dose BoNT-A group. In the intra-articular corticosteroid group, VAS decreased from 6.4 ± 1.8 to 5.4 ± 2.3 (P = .15); for low-dose BoNT-A, from 6.6. ± 1.9 to 4.5 ± 2.2 (P = .01); and for high-dose BoNT-A, from 6.6 ± 1.4 to 5.9 ± 2.4 (P = .15). All groups showed statistically significant improvements in Western Ontario McMaster Arthritis Index scores (pain, stiffness, function) at 8 weeks. No serious adverse events were noted in any group. Conclusions: This pilot study supports a possible role for BoNT-A as a treatment option for symptomatic knee osteoarthritis; however, larger double-blind randomized studies are needed to determine whether BoNT-A is more effective than placebo in this patient population. © 2010 American Academy of Physical Medicine and Rehabilitation.