Two operations, two different jobs

Choosing between a high tibial osteotomy (HTO) and a unicompartmental (partial) knee replacement for medial knee osteoarthritis is not a matter of one operation being better than the other — it is a matter of which operation matches the problem in front of it.

The two procedures do fundamentally different jobs. An HTO works on the bone itself: a precise wedge is created in the upper tibia to shift body weight away from the worn medial compartment, correcting the varus (bow-legged) alignment that is driving the damage in the first place. No joint surface is removed or replaced. A unicompartmental knee arthroplasty (UKA) takes a different approach entirely — it resurfaces the diseased medial compartment with metal and polyethylene components, relieving pain at the source of the wear without touching the rest of the joint.

Both operations share the same overarching goal: to reduce pain, restore function, and delay or avoid total knee replacement for as long as possible. Where they diverge is in the patient profile each suits best — particularly around age, activity level, and the condition of the remaining cartilage. The sections below set out how those factors guide the decision.

How each operation works

Stand with your feet together and look at the line from your hip to your ankle. In a neutral knee, that line passes through the centre of the joint, spreading weight evenly across both compartments. In varus — the bow-legged alignment that commonly accompanies medial knee OA — that mechanical axis tilts inward, directing the majority of the body's load through the inner (medial) compartment with every step. Both operations in this comparison address the consequences of that pattern, but through entirely different means.

What the HTO surgeon does

In a medial opening-wedge HTO, the surgeon makes a careful cut in the upper tibia and gently opens a precise gap, inserting a bone graft and securing a low-profile plate (commonly a Tomofix plate) to hold the new angle. The tibia is essentially tilted slightly outward, shifting the mechanical axis away from the worn medial side. Nothing about the joint surface itself is removed or replaced — the cartilage, menisci, and ligaments remain exactly where they are. Because the procedure targets the bone beneath the joint, it can be thought of as correcting the structural cause of the wear rather than treating its effects. The trade-off is that pushing load toward the lateral compartment may, over time, accelerate wear on that side.

What the UKA surgeon does

In a unicompartmental knee arthroplasty, the medial compartment alone is resurfaced: a thin metal cap is fixed to the femoral condyle, a matching tibial tray is placed, and a plastic insert sits between them. The ACL, the lateral compartment, and the patellofemoral joint are left entirely intact. Medial UKA achieves approximately 90% survivorship at 15 years in well-selected patients, with revision rates in registry data most often attributed to component malpositioning rather than implant failure itself. The procedure relieves pain at its source without addressing the underlying varus — which matters when choosing between the two.

What the evidence actually shows

The numbers from the largest comparative analysis look clear enough. A 2025 meta-analysis of 39 studies and 56,686 patients found UKA significantly outperformed HTO on complication rates (relative risk 0.37), revision-to-TKA rates (RR 0.64), post-operative pain, and HSS functional scores; range of motion and walking speed were equivalent between the two procedures.

That aggregate advantage, however, reflects a pooled population that skews older — UKA's dominance among elderly, lower-demand patients inflates its overall numbers.

The age-stratified data complicates the summary. A concurrent 2025 narrative review of 114 studies found that patient-reported outcome improvements can be broadly equivalent between the two procedures in selected younger cohorts. The Knee OUTcome Study — confined to men under 55 with medial OA — reported ten-year survivorship of 99% for UKA against 58% for HTO, with HTO's overall reoperation rate reaching 50% (including union failures and hardware problems, not only conversion to TKA). Survivorship here means still functioning without conversion to total knee replacement. That gap is difficult to dismiss; it is also a single-centre retrospective study, making it impossible to separate technique effects from inherent procedural differences.

Held together, the evidence is genuinely uncertain rather than simply awaiting a definitive trial. For patients under 55, the evidence is split — and the state of the remaining cartilage does more work than age alone in guiding the decision. Where early medial cartilage loss accompanies significant varus, HTO's case for structural correction remains plausible; where bone-on-bone wear dominates, UKA's survivorship advantage is harder to argue against.

No large prospective randomised trial comparing HTO and UKA with contemporary surgical techniques has been completed. What that means practically is that no single headline figure settles the choice — cartilage grade, alignment severity, activity demands, and patient preference all have to be weighed in a consultant assessment rather than read off an aggregate number.

Who is the right candidate for each procedure

The age-and-cartilage pairing does most of the sorting work here — though neither variable operates alone.

For patients in their late forties or early fifties who are still physically active, running, or doing heavy manual work, HTO is typically the starting point of the conversation. The reasoning is biological: if the varus deformity is the primary driver of the medial load, and the cartilage on that side is damaged but not entirely gone, reshaping the bone to redistribute force addresses the root problem. The procedure preserves every native structure and leaves the joint available for future intervention without the tissue compromises that arthroplasty would create. Importantly, a relative contraindication to HTO is severe medial cartilage loss — a paradox, given that 'early OA' is exactly the indication the procedure is designed for. That is why cartilage grading, through MRI and sometimes direct arthroscopic assessment, is often the decisive step before the choice is finalised: a scan showing significant residual cartilage supports the HTO pathway; one showing bone-on-bone contact shifts the balance toward resurfacing.

For patients in their late fifties who are less physically demanding, and whose medial compartment is essentially worn through, partial replacement tends to offer a faster and more predictable recovery. The ACL must be intact and the patellofemoral joint relatively unaffected — these are clinical prerequisites, not flexible preferences.

In the 50–60 bracket, the two procedures are genuinely in competition. HTO may be more cost-effective if it successfully delays total knee replacement by a decade or more; if it does not, conversion to UKA or TKA remains available. UKA can likewise be revised to TKA if the disease progresses.

One number that should not be assigned with false precision is the varus angle threshold. Clinical consensus does not currently specify the exact degree of deformity at which HTO becomes preferable to UKA — that assessment belongs in a consultant evaluation, not in a self-triage framework.

Complications and recovery compared

The risks diverge before the first dressing change — and so does the recovery trajectory.

For HTO, procedure-specific complications include hinge fracture at the osteotomy cut, delayed or non-union of the reshaped bone, and hardware irritation from the plate and screws, which sometimes requires removal once healing is confirmed. With contemporary implants such as the Tomofix locking plate, restricted weight-bearing from around two weeks post-operatively is now standard practice, but meaningful rehabilitation remains constrained by the bone-healing phase — typically six to twelve weeks of partial loading before full progression. For working-age patients who cannot take extended leave, or who rely on physical jobs, that timeline is a material factor in the decision rather than a footnote to it. Hinge fracture rates are low with modern technique, and non-union is uncommon; these are risks to understand rather than to be deterred by.

UKA carries a different profile. Mobile-bearing implant designs carry a specific risk of bearing insert dislocation. Beyond that, the principal long-term concern is disease progression in the lateral or patellofemoral compartments — a mechanistic consequence of the fact that UKA does not correct the underlying varus, as covered earlier. Registry data consistently show that revision after UKA is driven disproportionately by component malpositioning and limb alignment errors rather than implant failure as such, which is why surgical experience with the technique has a demonstrable bearing on durable results.

Recovery generally favours UKA: most patients bear full weight immediately, return to driving within weeks, and resume daily activities sooner than after HTO.

Prior HTO can make a subsequent total knee replacement technically more demanding — altered bone geometry and retained metalwork require careful pre-operative planning — which is worth factoring in for younger patients for whom multiple procedures over a lifetime may be realistic.

Getting the right assessment before deciding

Making this choice rests on more than reported pain level or age bracket. The decisive data comes from a structured diagnostic workup: standing, weight-bearing X-rays to quantify the varus deformity; MRI to grade cartilage loss in the medial compartment using Outerbridge or ICRS staging — meaningful residual cartilage pointing toward HTO, bone-on-bone contact shifting the case toward UKA; and clinical confirmation that the ACL is intact and the patellofemoral compartment relatively unaffected.

Objective, dynamic data on how load is distributed through the knee during movement can sharpen this picture further, particularly where the varus angle sits in the borderline range where both procedures are genuinely in competition. MAI Motion®, the group's UKCA-registered markerless motion-capture system, provides that quantified loading analysis as part of the assessment pathway, complementing static imaging rather than substituting for it.

Patients across the Lincolnshire and wider non-London catchment can access this workup without a GP referral at the Sleaford site — where an Open MRI scanner and consultation facilities sit together — or at the Grantham centre. For London-based patients, the London Cartilage Clinic offers an equivalent consultant-led pathway.

The governing principle throughout this decision is matching the operation to the individual — not defaulting to one procedure on demographic grounds alone, and not conflating early OA with an automatic indication for either technique. That judgement depends on having the right imaging, biomechanical data, and clinical expertise in the same conversation; to arrange that assessment directly, without a referral, visit mskdoctors.com.

1. [1] HTO, UKA, and PFO for Medial-Compartment Knee Osteoarthritis: A Narrative Review of Comparative Mechanisms, Clinical Outcomes, and Decision-Making. (2025). https://doi.org/10.3390/jcm14217882 https://doi.org/10.3390/jcm14217882
2. [2] Current concepts and a glimpse into the future of partial knee replacement around the knee. (2024). https://doi.org/10.71165/n2y7-hw2n https://doi.org/10.71165/n2y7-hw2n
3. [3] Comparison of Unicompartmental Knee Arthroplasty Versus High Tibial Osteotomy for Medial Knee Osteoarthritis: An Updated Meta-Analysis of 56,000 Patients. (2025). https://doi.org/10.1111/os.70049 https://doi.org/10.1111/os.70049
4. [4] TOPKAT: Assessing clinical and cost effectiveness of total versus partial knee replacement — 10-year follow-up. (2025). https://doi.org/10.1016/s2665-9913(25)00250-4 https://doi.org/10.1016/s2665-9913(25)00250-4
5. [5] Knee OUTcome Study: Comparison of HTO, UKA, and TKA for Medial Compartment OA in Men under Age 55. (2017). https://doi.org/10.1177/2325967117S00165 https://doi.org/10.1177/2325967117S00165