Partial or Total Knee for Medial Compartment Arthritis
A partial knee replacement trades a lower chance of complications and early death for a higher chance of a second operation, so the choice changes which risks the patient carries into old age.
Key terminology
- Medial compartment
- The medial compartment is the joint between the inner femoral condyle and the inner tibial plateau. It is the most common site of knee osteoarthritis.
- Unicompartmental knee arthroplasty (UKA)
- This operation resurfaces only the worn compartment and leaves the rest of the knee and both cruciate ligaments in place. A total knee arthroplasty (TKA) replaces all three compartments.
- Anterior cruciate ligament (ACL)
- This ligament runs from the front of the tibia to the inner side of the lateral femoral condyle. It stops the tibia sliding forward, and a medial UKA depends on it.
- Varus deformity
- The leg below the knee angles toward the midline, giving a bowlegged stance. In medial arthritis the knee tips into varus as the medial cartilage wears away.
- Anteromedial osteoarthritis
- This wear pattern erodes the front of the medial tibial plateau and spares the back. It is the classic pattern that suits a medial partial knee.
- Revision
- This is a second operation that removes or exchanges part or all of a joint replacement. Registries count it as their main marker of implant failure.
- Joint registry
- This national database records every joint replacement and any later revision. It shows routine results across many surgeons and cannot balance patients the way randomization does.
- Oxford Knee Score
- This 12 question patient survey scores knee pain and function from 0 to 48. A higher score means a better knee.
Takeaways
- A medial partial knee suits anteromedial osteoarthritis with an intact ACL. Medial unicompartmental knee arthroplasty (UKA) resurfaces only the medial compartment. In anteromedial osteoarthritis the wear sits on the front of the medial tibial plateau and the back is spared, so the varus deformity corrects in flexion. Once the anterior cruciate ligament ruptures, wear spreads posteriorly, the varus becomes fixed and the lateral compartment degenerates (White et al, J Bone Joint Surg Br 1991;73:582-586, PMID 2071640).
- Medial UKA is reserved for full-thickness, bone-on-bone cartilage loss. In 94 UKAs done for partial-thickness loss, function was worse at 1, 2 and 5 years than in matched bone-on-bone knees. Reoperation was 3 times higher, mostly arthroscopy for persistent pain. MRI did not predict who would do well (Hamilton et al, Bone Joint J 2017;99-B:475-482, PMID 28385936).
- The randomized trial found equal function after partial and total knee replacement. TOPKAT randomized 528 patients with isolated medial osteoarthritis at 27 UK sites. The Oxford Knee Score did not differ at 5 years (mean difference 1.04, p=0.159) or 10 years (0.27). Revision at 10 years was 5% in both arms by allocation, and the partial knee was cheaper and more cost-effective (Beard et al, Lancet 2019;394:746-756, PMID 31326135; Beard et al, Lancet Rheumatol 2026;8:e116-e126, PMID 41270774).
- Joint registry data show partial knees are revised about twice as often as total knees. In 101,330 propensity-matched patients from the England and Wales joint registry, UKA had a revision subhazard ratio of 2.12 at 8 years. A later matched analysis of 108,430 knees found 10-year revision of 12% vs 5% (HR 2.31). Propensity matching cannot remove selection the way randomization does (Liddle et al, Lancet 2014;384:1437-1445, PMID 25012116; Mohammad et al, J Arthroplasty 2022;37:243-251, PMID 34619307).
- A total knee has more medical complications and early deaths than a partial knee. In the matched registry, 30-day mortality was lower after UKA (HR 0.23), and thromboembolism, MI, stroke, readmission and length of stay were all higher after TKA. Switching 100 patients from TKA to UKA meant about 1 fewer death and 3 more reoperations in 4 years. A 60-study meta-analysis found the same pattern (VTE risk ratio 0.39 in registries) (Liddle et al, Lancet 2014;384:1437-1445, PMID 25012116; Wilson et al, BMJ 2019;364:l352, PMID 30792179).
- Revision rate overstates partial knee failure because failed partial knees are revised more readily. In the New Zealand registry, among knees with a very poor Oxford Knee Score (under 20), about 63% of UKAs were revised vs about 12% of TKAs. The authors argued revision is not an objective way to compare the two implants. This argument comes from the group that designed the Oxford partial knee (Goodfellow et al, J Bone Joint Surg Br 2010;92:1628-1631, PMID 21119165).
- Surgeons who rarely use partial knees have the highest partial knee revision rates. In 41,986 registry UKAs, most surgeons offered UKA to 5% or fewer of their knee replacement patients. Results were acceptable at 20% usage or more and best at 40-60%, with 5-year survival of 96% vs 90% at 5% usage (Liddle et al, Bone Joint J 2015;97-B:1506-1511, PMID 26530653).
- Age, weight and patellofemoral wear are no longer firm contraindications to medial UKA. Kozinn and Scott (1989) excluded patients over 82 kg, under 60, in heavy labor, or with exposed patellofemoral bone or chondrocalcinosis. In 1000 Oxford UKAs, 68% had at least one of these, and 10-year survival was 97.0% vs 93.6% in ideal patients. This is a single designer-center series, so other surgeons may not match it (Pandit et al, J Bone Joint Surg Br 2011;93:622-628, PMID 21511927).