Isolated medial compartment osteoarthritis accounts for nearly half of all surgical knee osteoarthritis presentations. Historically, registry data showed substantially higher revision rates for unicompartmental knee replacement (UKR/PKR) compared to total knee replacement (TKR), discouraging its uptake despite theoretical advantages such as bone conservation, preservation of native cruciate kinematics, reduced surgical morbidity, and faster rehabilitation. However, registry observations suffer from significant confounding, including lower revision thresholds for partial knee implants. The TOPKAT trial was designed to provide definitive Level I evidence comparing the clinical effectiveness and cost-effectiveness of TKR versus PKR in patients eligible for both procedures.

TOPKAT was a pragmatic, multicentre randomised controlled trial conducted across 27 National Health Service (NHS) hospital sites in the United Kingdom. It included both a standard randomisation track and an expertise-based track to account for surgeon preference and technique proficiency. A total of 528 patients with end-stage medial compartment osteoarthritis suitable for either TKR or PKR were randomised 1:1 to undergo either TKR (n=264) or PKR (n=264). The primary outcome was the Oxford Knee Score (OKS) at 5 years post-randomisation. Secondary endpoints included American Knee Society Scores (AKSS), EQ-5D-3L health utility, complication rates, reoperations, implant revision, and comprehensive health economic evaluation.

TOPKAT challenged the historical dogma that registry-derived revision rates make partial knee arthroplasty an inferior option. It established that when performed by trained surgeons in appropriately selected patients with isolated medial compartment disease, PKR offers equal clinical function, equivalent short-to-medium-term survivorship, and superior cost-effectiveness compared to TKR. These findings supported guidelines advocating that PKR be considered a primary surgical option—rather than a niche alternative—for isolated medial unicompartmental osteoarthritis.

Strengths include the multicentre pragmatic design across 27 centres, excellent 5-year follow-up retention, integration of an innovative expertise-based randomisation model to mitigate surgeon learning-curve bias, and robust parallel health-economic analysis. Limitations include the lack of patient and surgeon blinding (inherent to surgical trials comparing distinct incisions), potential under-representation of younger high-demand patients due to recruitment criteria, and the use of implants and cemented fixation techniques that preceded more modern cementless or robotic-assisted modalities.