Highlights
Both partial (unicompartmental) knee replacement (PKR) and total knee replacement (TKR) produced similar patient-reported outcomes at 10 years (mean difference in Oxford Knee Score [OKS] 0.27, 95% CI -1.59 to 2.13), with no clinically meaningful differences. Reoperation and revision rates were comparable between groups. Economic analysis favoured PKR (mean QALY difference 0.322, 95% CI -0.069 to 0.712; mean cost difference -£731, 95% CI -1352 to -110), making PKR the more cost-effective strategy in the TOPKAT cohort.
Background and Clinical Context
Late-stage medial compartment osteoarthritis (OA) of the knee is a common, disabling condition affecting older adults and is a leading indication for knee arthroplasty. Treatment options for isolated medial compartment disease include total knee replacement (TKR) and partial (unicompartmental) knee replacement (PKR or UKA). PKR preserves non-diseased compartments and native ligaments, offering theoretical benefits of faster recovery, more physiological knee kinematics, and lower perioperative morbidity, while concerns remain about higher revision rates in some registry analyses. Longevity of the implant and long-term patient-reported outcomes are critical in choosing between procedures; randomized long-term data are therefore essential to inform practice and policy.
Study Design
Overview
The Total or Partial Knee Arthroplasty Trial (TOPKAT) was a multicentre, pragmatic, randomized controlled trial with an expertise component conducted across 27 UK NHS hospitals. Between Jan 18, 2010, and Sept 30, 2013, 528 patients with symptomatic medial compartment knee OA were randomized 1:1 to PKR (n=264) or TKR (n=264). Surgeons were either able and willing to perform both procedures (generalists) or were designated experts in the allocated procedure; allocation was unblinded to surgeons and patients.
Endpoints and Follow-up
The primary long-term endpoint reported here was patient-reported function using the Oxford Knee Score (OKS) at 10 years in the intention-to-treat population. Secondary outcomes included area under the curve (AUC) of OKS over 10 years, complications, reoperations (including revisions), and a within-trial cost-effectiveness analysis estimating quality-adjusted life years (QALYs) and healthcare costs. Patient representatives were involved in study design. Trial registration: ISRCTN03013488; ClinicalTrials.gov NCT01352247.
Key Findings and Interpretation
Primary and Longitudinal Patient-Reported Outcomes
At 10 years, follow-up response rate among eligible participants (excluding deaths and withdrawals) was 326/444 (73%). Mean OKS difference between PKR and TKR at 10 years was 0.27 (95% CI -1.59 to 2.13), a result that rules out any clinically meaningful difference between procedures. The 10-year AUC analysis, capturing outcomes across the entire follow-up period, likewise showed no meaningful cumulative difference (mean difference 0.45, 95% CI -0.98 to 1.88).
Clinical implication: from a patient-reported function standpoint, both operations delivered sustained, comparable improvements up to 10 years for patients with isolated medial compartment OA when applied within the trial’s inclusion criteria and surgical context.
Surgical Safety, Reoperations and Revisions
By treatment received at 10 years, rates were as follows: complications 53/245 (22%) for PKR versus 74/270 (27%) for TKR; reoperations (including revision) 21/245 (9%) for PKR versus 23/270 (9%) for TKR; and revisions 15/245 (6%) for PKR versus 11/270 (4%) for TKR. In the intention-to-treat analysis (by allocation) complication rates were 21% (PKR) versus 29% (TKR), reoperations 8% versus 10%, and revisions 5% versus 5%.
Interpretation: absolute differences in reoperation and revision were small and not convincingly different between groups. Numerically, PKR showed a modest excess in revisions by treatment received (6% vs 4%) but this did not translate into greater reoperation rates overall or worse patient-reported outcomes at 10 years.
Cost-effectiveness
Within-trial economic evaluation demonstrated that PKR was more cost-effective than TKR over the 10-year horizon. PKR was associated with a mean QALY gain of 0.322 (95% CI -0.069 to 0.712) and lower mean healthcare costs by -£731 (95% CI -1352 to -110). Although the QALY confidence interval crosses zero, the combination of increased mean QALYs and lower costs made PKR the dominant strategy in the trial’s analysis.
Clinical and policy implication: where patient selection permits use of PKR, health systems with constrained resources may realize modest cost savings alongside at least equivalent patient outcomes.
Strengths of the Trial
– Randomized, multicentre pragmatic design reflecting real-world NHS practice increases external relevance for similar health systems.
– Long-term (10-year) follow-up addresses an essential question about durability and sustained patient benefit.
– Inclusion of an expertise component mirrors clinical realities where surgeon experience varies and may reduce bias from surgical skill differences.
– Embedded economic evaluation provides actionable information for clinicians and decision-makers.
Limitations and Considerations
– Lack of blinding for patients and surgeons is unavoidable in surgical trials but can influence subjective outcomes; however, the primary endpoint is a patient-reported measure designed for knee arthroplasty and results were consistent across longitudinal analyses.
– Follow-up response at 73% among eligible participants leaves potential for attrition bias; the authors performed appropriate intention-to-treat analyses but missing data remain a limitation for rare events and subgroup analyses.
– The trial’s sample size—while large for a randomized surgical trial—was not designed specifically to detect small differences in rare outcomes such as revision beyond 10 years; registry data remain complementary for late implant survivorship.
– Generalizability is tied to surgeon experience and case selection: outcomes for PKR are sensitive to patient selection criteria (isolated medial compartment disease, intact ligaments) and the operator’s volume and skill. Health systems or centers with low UKA volumes may not replicate these results.
How These Results Fit with Other Evidence
Registry data have historically suggested higher revision rates with PKR than TKR in some settings, although registries reflect broad real-world practice with varying surgeon volumes and implant choices. TOPKAT’s randomized evidence indicates that when PKR is used for appropriately selected patients within experienced or supported contexts, clinical outcomes and reoperation rates are comparable to TKR up to 10 years, and PKR may confer economic advantages. This supports selective use of PKR rather than defaulting to TKR for all medial compartment disease.
Clinical Implications and Recommendations
– Patient selection remains paramount: PKR is appropriate for patients with isolated medial compartment OA, preserved lateral compartment and patellofemoral joint where clinically indicated, and intact collateral ligaments.
– Surgeon experience matters: centres and surgeons considering offering PKR should ensure adequate training, caseload, and governance to achieve outcomes comparable to trials.
– Shared decision-making should incorporate the comparable functional outcomes, similar reoperation risk in this trial, and the economic advantages of PKR, while discussing the potential for higher revision rates reported in some registries and the contingency planning for conversion to TKR if needed.
– Health services should consider PKR as a cost-effective option in eligible patients, but implementation should include quality assurance measures (audit, registry reporting, competence standards).
Research and Policy Gaps
– Longer follow-up beyond 10 years and linkage to national registries will clarify late implant survivorship and inform lifetime cost-effectiveness.
– Subgroup analyses (age, activity level, BMI, comorbidity) to define patients who derive the most benefit from PKR vs TKR.
– Comparative effectiveness across implant designs and surgical techniques (robotic-assisted UKA, patient-specific instrumentation) requires randomized and registry-coupled evidence.
Conclusion
The 10-year TOPKAT results provide high-quality randomized evidence that, for patients with isolated medial compartment knee OA, PKR and TKR deliver similar long-term patient-reported outcomes and comparable reoperation/revision rates when applied in a pragmatic, expertise-inclusive surgical setting. Economic analysis favours PKR, which produced modestly greater mean QALYs and lower healthcare costs over 10 years. These findings support the considered use of PKR in well-selected patients within systems that ensure surgical expertise and ongoing outcome monitoring.
Funding and Trial Registration
The TOPKAT trial was funded by the National Institute for Health and Care Research Health Technology Assessment Programme. Trial registration: ISRCTN03013488; ClinicalTrials.gov NCT01352247.
References
1. Beard DJ, Davies LJ, Cook JA, MacLennan G, Hudson J, Price AJ, Carr AJ, Little M, Leal J, Fitzpatrick R, Murray DW, Campbell MK; TOPKAT Study Group. Assessing clinical and cost effectiveness of total versus partial knee replacement (TOPKAT): 10-year follow-up of a multicentre, randomised controlled trial. Lancet Rheumatol. 2025 Nov 18:S2665-9913(25)00250-4. doi: 10.1016/S2665-9913(25)00250-4.
2. Dawson J, Fitzpatrick R, Carr A, Murray D. Questionnaire on the perceptions of patients about total knee replacement. J Bone Joint Surg Br. 1998 Jan;80(1):63-9. (Original description and validation of the Oxford Knee Score.)
3. National Joint Registry for England, Wales, Northern Ireland and the Isle of Man. NJR 20th Annual Report 2023. (Annual registry data provide broader context on implant survivorship and national trends.)
