Total Knee Arthroplasty versus Mesenchymal Stem Cell Therapy for Moderate Knee Osteoarthritis: An Umbrella Analysis of Clinical Outcomes, Quality of Life, and Safety

Background. Moderate knee osteoarthritis (Kellgren-Lawrence [KL] Grade 2-3) represents a critical therapeutic crossroads where patients must choose between regenerative strategies and surgical replacement. This umbrella analysis synthesizes current evidence regarding the clinical efficacy, quality of life, and safety profiles of mesenchymal stem cell (MSC) therapy relative to total knee arthroplasty (TKA).
Material and methods. Following PRISMA 2020 guidelines, we systematically searched seven databases (2016-2026). Screening of 1,211 records identified 242 unique papers for full-text assessment. Eligible studies included systematic reviews, meta-analyses, and large prospective cohorts (>50 participants) evaluating intra-articular MSCs or TKA in KL Grade 2-3 populations with ≥12-month follow-up.
Results. Our analysis reveals a profound evidence gap: zero studies directly compared MSC therapy to TKA in this population. Data from 30 high-quality studies demonstrate that MSC therapy significantly outperforms placebo and conservative management, yielding VAS pain reductions of 20-40 points and WOMAC functional improvements of 15-35 points at 12-24 months. Adipose-derived MSCs showed superior analgesia, while umbilical cord-derived MSCs optimized functional recovery. MSC therapy significantly reduced the hazard of TKA conversion (HR 0.52, p<0.001). At 36-60 months, conversion rates were 12.4% for MSCs versus 22.7% for control groups. Safety profiles remained favorable with serious adverse event rates <5%. 
Conclusions. 1. While MSC therapy is a demonstrably effective bridge for delaying TKA in selected KL Grade 2-3 patients, its comparative effectiveness and long-term durability relative to arthroplasty remain uncharacterized. 2. High-quality randomized controlled trials directly benchmarking MSCs against TKA are urgently required to resolve this clinical impasse and inform treatment algorithms.