Background: There has been much debate regarding the use of intra-articular injections of platelet-rich plasma (PRP) as symptomatic treatment for knee osteoarthritis. The heterogeneity of the preparation and injection protocols limits the extrapolation of data from randomized controlled trials and meta-analyses.

Objectives: The objective of this expert consensus was to develop the first clinical practice recommendations for PRP injections in knee osteoarthritis.

Methods: Fifteen physicians (10 rheumatologists, 4 specialists in rehabilitation and sport medicine and 1 interventional radiologist) from different countries were selected given to their expertise in the fields of PRP and osteoarthritis. Twenty-five recommendations were finally retained after several meetings using the modified Delphi method to establish clinical consensus. All experts voted their agreement or not for each recommendation using a score between 1 (totally inappropriate) and 9 (totally appropriate). Depending on the median value and extreme scores, recommendations were judged as appropriated or unappropriated with a strong or relative agreement but could also be judged as uncertain due to indecision or absence of consensus.

Results: The main recommendations are listed below:

• - Intra-articular injections of PRP constitute an efficient treatment of early or moderate symptomatic knee osteoarthritis. Median = 8 [6-9] – Appropriate. Relative agreement.

• - Intra-articular injections of PRP may be useful in severe knee osteoarthritis (Kellgren-Lawrence grade IV). Median = 7 [6-7] – Appropriate. Relative agreement.

• - Intra-articular injections of PRP in knee osteoarthritis should be proposed as second-line therapy, after failure of non-pharmacological and pharmacological (oral and topic) symptomatic treatment. Median = 9 [5-9] – Appropriate. Relative agreement.

• - Intra-articular injections of PRP should not be performed in osteoarthritis flare-up with significant effusion. Median = 7 [5-9] – Appropriate. Relative agreement.

• - Intra-articular PRP treatment may include 1 to 3 consecutive injections. Median = 9 [7-9] – Appropriate. Strong agreement.

• - Leukocyte-poor PRP should be preferred for knee OA treatment. Median = 8 [5-9] – Appropriate. Relative agreement.

• - PRP injections should be performed under ultrasound or fluoroscopic guidance. Median = 8 [3-9] – Uncertain. No consensus.

• - PRP should not be mixed with injectable anesthetic or corticosteroid. Median = 9 [6-9] – Appropriate. Relative agreement.

Conclusion: Twenty-five recommendations were discussed by an international multidisciplinary task force group in order to provide a basis for standardization of clinical practices and future research protocols.

Disclosure of Interests: Florent Eymard Consultant of: Regenlab, Paul Ornetti: None declared, Jérémy Maillet Consultant of: Regenlab, Eric Noel Consultant of: Regenlab, Philippe Adam Consultant of: Regenlab, Virginie Legré Boyer Consultant of: Regenlab, Thierry Boyer Consultant of: Regenlab, Fadoua Allali: None declared, Vincent Grémeaux Bader: None declared, Jean-François Kaux: None declared, Karine Louati: None declared, Martin Lamontagne Consultant of: Pendopharm, Fabrice Michel: None declared, Pascal Richette: None declared, Hervé Bard Consultant of: Regenlab