What Actually Works for Knee OA Injections — A 2025 Evidence Breakdown

Osteoarthritis affects an estimated 30.8 million Americans and 300 million people worldwide, representing the most prevalent joint disorder globally. Despite its enormous burden, no approved injection therapy can halt disease progression — all current options target symptom management. This 2025 narrative review from the Schroeder Arthritis Institute systematically evaluates the evidence for every major intra-articular (IA) injection class, from traditional agents to cutting-edge biologics and gene therapies, with emphasis on the most recent prospective randomized trials.

Corticosteroids remain widely used for short-term pain relief, typically effective for 4–6 weeks. However, the landmark McAlindon et al. double-blind RCT (n=140, 2 years) found that triamcinolone 40mg every 12 weeks produced greater cartilage thickness loss than saline (−0.21mm vs. −0.10mm) without superior long-term pain relief. This critical finding suggests repeated corticosteroid injections may accelerate the very disease they are meant to treat. Hyaluronic acid (viscosupplementation) fares no better: a meta-analysis of 24 placebo-controlled trials (n=8,997) found only a modest pain reduction (SMD −0.08; 95% CI −0.15 to −0.02) that failed to meet the minimal clinically important difference threshold, while a separate analysis of 15 trials (n=6,462) found a significantly elevated risk of serious adverse events (RR 1.49; 95% CI 1.12–1.98).

Platelet-rich plasma (PRP) has generated the most clinical research among biologics. Multiple meta-analyses and RCTs report superior pain and functional outcomes versus HA and placebo at 3, 6, and 12 months. A Shen et al. meta-analysis of 14 RCTs (n=1,423) found PRP consistently outperformed controls including saline, HA, ozone, and corticosteroids, with effects most pronounced at 6 and 12 months. A double-blind RCT (Patel et al., n=156 knees) showed significant WOMAC and VAS improvements at 6 weeks, 3 months, and 6 months with no serious adverse events. Leukocyte-poor PRP appears superior to leukocyte-rich formulations. However, the largest and most rigorous trial — Bennell et al. (n=288, triple-blinded, 1 year) — found no statistically significant difference between leukocyte-poor PRP and saline in pain reduction (2.1 vs. 1.8 points) or cartilage volume loss (1.4% vs. 1.2%). PRP's heterogeneous preparation methods remain a major barrier to definitive conclusions.

Bone marrow aspirate concentrate (BMAC) and mesenchymal stromal cells (MSCs) represent the frontier of biologic IA therapy. BMAC contains a mix of growth factors, cytokines, and progenitor cells. Early RCTs show promising pain and function improvements, but sample sizes remain small and long-term structural data are limited. MSC injections — both autologous and allogeneic — have shown signals of cartilage preservation in phase I/II trials, but no large phase III RCT has yet confirmed efficacy or safety at scale. Standardization of cell sourcing, dosing, and preparation remains unresolved.

Emerging therapies include sprifermin (fibroblast growth factor 18), which has shown dose-dependent cartilage thickness preservation in phase II trials, and gene therapy approaches targeting pro-inflammatory pathways such as IL-1 and TNF-α. These represent the most disease-modifying potential of any IA strategy studied to date, but remain investigational. The review concludes that while IA injections are a cornerstone of conservative OA management, no current therapy offers proven disease modification, and treatment selection must weigh short-term symptom relief against potential joint harm.