Skip to content

Bennell KL, et al. Platelet-rich plasma versus placebo injection for knee OA: RESTORE. JAMA. 2021;326:2021–2030. PMID 34812863

One-paragraph summary

RESTORE randomized 288 adults aged ≥50 years with symptomatic medial knee OA (Kellgren–Lawrence 2–3) to three weekly leukocyte-poor PRP or saline injections with participant, injector and assessor blinding; 269 (93%) completed 12 months. Pain improved in both groups, but the between-group pain difference was -0.4/10 (95% CI -0.9 to 0.2), below the prespecified 1.8-point clinically important difference. Medial tibial cartilage volume changed -1.4% versus -1.2%, a difference of -0.2% (95% CI -1.9% to 1.5%).

Key findings

  • 97% of each group received all three injections.
  • Twenty-nine of 31 prespecified secondary outcomes showed no significant between-group difference.
  • The trial tested a characterized leukocyte-poor product; it did not test every possible PRP formulation.
  • Large within-group improvement illustrates why uncontrolled series and active-comparator trials can overstate product-specific effect.

Limitations

  • One PRP preparation and schedule; formulation heterogeneity limits class-wide inference.
  • The selected mild-to-moderate medial-compartment population may not represent other phenotypes.
  • Twelve months may be insufficient for some structural effects, but it was adequate to reject a large symptomatic claim.

Why it matters

RESTORE is the pivotal rigorous placebo-controlled counterweight to positive PRP-versus-hyaluronate meta-analyses.

Cited by wiki pages

  • overview
  • injections and orthobiologics