Purpose: To evaluate whether platelet-rich plasma or intra-articular plasma are superior to placebo in improving pain and function in patients with knee osteoarthritis (OA) in a 6-month follow-up period. Methods: Randomized, double-blind, clinical trial with 3 groups. Compared intra-articular treatment of knee OA of two experimental groups, one with PRP (N = 20) and other with plasma (N = 20), with a control group receiving saline solution (N = 21). Two ultrasound-guided joint injections were performed at intervals of 2 weeks. PRP injected was fresh, obtained by double centrifugation, leucocyte-poor, nonactivated and contained on average 1 million platelets/mm3. The primary endpoint was general pain in the injected knee measured by the visual analogue scale (VAS) 0-10 cm evaluated at week 24, with intermediate evaluations at weeks 6 and 12. The secondary outcomes were: WOMAC (Western Ontario McMaster University Osteoarthritis Index), KOOS (Knee Injury and Osteoarthritis Outcome Score), OMERAC-OARSI criteria and TUGT (Timed Up and Go Test). Results: The general pain (VAS) assessment in week 24 showed significant improvement regarding the 3 groups (p = 0.000 for all), but there was no statistical difference between them (p = 0.891). The PRP group presented reduction of the mean of pain from 6.1 (SD 1.6) to 3.2 (SD 2.5), which represented a decrease of 2.9 cm or 47.5%. The plasma group presented reduction of the mean of pain from 5.9 (SD 1.4) to 3.5 (SD 2.4), which represented a decrease of 2.4 cm or 40.7%. The placebo group showed a decrease in mean of pain from 6.6 (SD 1.4) to 3.1 (SD 2.6), which represented a decrease of 3.5 cm or 53.0%. In week 6, mean of pain was 3.6 (SD 2.4), 3.7 (SD 2.5) and 3.5 (SD 2.1) in the PRP, plasma and placebo groups, respectively, and in week 12 it was 2.0 (SD 1.4), 3.2 (SD 2.9) and 3.9 (SD 2.7), respectively. In weeks 6 and 12 there was also no statistical difference between the groups, despite the PRP group demonstrated a trend of improvement of pain in week 12 (see attached graphic). In WOMAC and KOOS there was significant improvement in all domains of the 3 groups, with no statistical difference between them in weeks 6, 12 and 24. The same occurred with TUGT. The percentage of patients who achieved OMERACT-OARSI improvement criteria in week 24 for the PRP, Placebo and Plasma groups were, respectively, 80%, 86% and 80%, and there was no statistical difference between the groups. Similar values were found in weeks 6 and 12. Adverse effects were more common within the PRP group, with the most frequent of them being mild to moderate knee pain lasting 2 days on average. Conclusions: PRP, plasma and saline improved knee OA pain and function outcomes in weeks 6, 12 and 24 with statistical and clinical significance. There was no difference between the groups, despite a favourable trend for PRP superiority in week 12.