Synovial fluid in patients may differ in molecular weight depending on the presence and degree of osteoarthritis. Treatment is not directed at this relationship. Patients with osteoarthritis of the knee with resting visual analogue scale (VAS) pain of >45 mm were included in a randomized, prospective, double-blind cohort followed for 16 weeks. Patients were randomized at baseline to receive a three intra-articular injection series with one of: dual molecular weight (DMW; 580-780 kDa + 1.2 to 2.0 million Da); low molecular weight (LMW; 500-730 kDa); high molecular weight (HMW; 6 million Da); or saline placebo over 3 weeks. Patients completed baseline assessment of rest and walking VAS pain (primary efficacy variable), collection of a 5-point categorical global satisfaction score, and record of adverse events. Two-hundred and twenty-five patients (age 68 +/- 8 y) were screened and 200 were randomized to one of the four groups. There were no differences at baseline between groups. At 4, 12 and 16 weeks, respectively, walking VAS pain was significantly improved in all treatment groups vs. placebo: DMW (79.6%, p < 0.001; 85.6, p < 0.001; 89.3%, p < 0.001); LMW (73.6%, p < 0.001; 76.4, p < 0.001; 81.3%, p < 0.001) and HMW (69.1%, p < 0.001; 81.0, p < 0.001; 79.1%, p < 0.001). Patients in the DMW group had significantly greater improvement (p < 0.007) in VAS walking pain by 3 weeks (following the second injection) compared to all groups. This difference was persistent at 16 weeks. Greater improvement in patients who received the DMW product was achieved by the second injection persistent at 16 weeks.