In treatment of isolated medial unicondylar osteoarthritis of the knee (MU-OA), it is possible to choose between surgery with a unicondylar knee arthroplasty (UKA), or a total knee arthroplasty (TKA).

Supporters of TKA suggest that this treatment gives a more predictable and better result, whereas supporters of UKA suggest that it is unnecessary to remove decent and functional cartilage in other compartments, and also that generally the UKA gives better results under certain circumstances. If the UKA is worn out or loosens, revision surgery will be relatively easy, whereas revision-surgery after a TKA can be much more problematic.

Also, it is of great interest to measure the direct costs of these treatments. Both general hospital costs, but also costs in societal in terms of sick-days, pain-killer expenditure, and physiotherapy.

The aim of this study is to compare the results, in terms of 1) patient-reported outcomes, 2) clinical results including prosthetic survival and 3) costs.

- Purpose OA of the knee can be debilitating and may require insertion of a knee prosthesis. In many cases there will be widespread degenerative changes of the knee, and in these cases it is necessary with the insertion of a complete prosthesis (total knee replacement, TKA) replacing all of the surfaces of the knee. However, there are cases in which the degenerative arthritis is found only between the thigh bone and the shin bone on the inner side of the knee, termed medial unicompartmental osteoarthritis (herein abbreviated to mu-OA). In situations of isolated mu-OA, it is possible to insert a complete prosthesis, as in the case of widespread OA of the knee, but it is also possible to insert a partial knee replacement (medial unicompartmental knee replacement, mUKA). There is not today enough information to decide which of the two types of prosthesis gives the best result in the event of mu-OA.

Supporters of TKA state that the treatment provides a predictable and good result, while supporters of mUKA say that it is not necessary to remove the good cartilage and mUKA provide even better results. If the inserted prosthesis become worn or loose, then it will be relatively easy to carry out a new operation after a mUKA, while a second surgery after TKA can be problematic. This is a further argument for mUKA. The costs of the two types of treatments are only poorly studied, and there are no comparative analyzes.

The aim of this study is to compare the results, in terms of 1) patient-reported outcomes, 2) clinical results including prosthetic survival of and 3) costs.

• Subjects All Danish citizens with isolated MU-OA referred to an orthopedic department at Gentofte, Næstved, Vejle, Århus and Aalborg hospital will be offered participation in the trial. The following exclude from participation in the study: under 18years of age, non-Danish speaking, alcohol or drug abuse, severe psychiatric disorder, severe systemic illness, employment in an orthopedic department, and major hip or back condition.

There will be 350 patients in the study. In 2014, the total numer of knee replacements at the participating centres was 3005 distributed at Gentofte (796 (112)), Århus (410 (152)), Vejle (629(138)), Aalborg/Farsø (381 (9)) Slagelse / Næstved (789(86)). Assuming an unchanged number of MUKA, 90% of which are suitable and a 50% participation accept-rate, we can expect an annual inclusion of approximately 220 patients. With an alternative assumption that about 25% of primary operations are suitable for mUKA, that 90% of these are suitable for the trial, and that 50% accept participation we can expect an annual inclusion of approximately 340. It is reasonable to expect that the total number of patients will be included within 12 to 19 months.

• Method The study will be conducted as a double blinded multicenter randomized clinical trial (RCT) where each participant by lot will have decided whether he should be treated with mUKA or TKA. During surgery, a midline incision in the skin will be performed on every participant, regardless of prosthesis type, securing the blinding of the participant. After the midline incision of the skin, the knee capsule is opened using the regular technique for each prosthesis. Participants will subsequently be carefully followed to determine whether there is a difference in the outcome of the two types of prosthesis.

There are many aspects in the outcome of an operation, and in this case the result will be assessed in terms of 1) complications during hospitalization, 2) rehabilitation rate, 3) Participant assessment of their own health, knee function and pain level, 4) surgical assessment knee condition, 5) radiographic findings, 6) long-term complications and 7) cost. To avoid bias in some of these targets, the participant will not know what type of prosthesis they have received the first year after surgery. To avoid bias based on doctors, physiotherapists and nurses' preferences, these groups, with the exception of the operating physician, will be blinded for the type of prosthesis.

There will be an analysis of results at 25% inclusion of patients (total of 350, therefore analysis at 85 patients included) with operation per protocol, and if undesirable results should be acknowledged, the study will be stopped by statistically significant difference (2p <0.01) between the two treatment groups. A Datareview-board has been set in place for this, whom are neutral to the study. There are planned short-term publications of early results at about 1-2 years after inclusion of the last patient, medium- and long-term results after 5 and 10 years.

• Risks and benefits Every participant having surgery with insertion of a knee prosthesis, runs the risk of complications. The main and most serious of which is infection in the knee, blood clot in the leg, and loosening of the prosthesis. Every participant is subjected to these risks, but they do not take greater risks than participants treated outside the trial.

Participants would, as the main disadvantage, appear for additional outpatient controls after 1, 2, 5 and 10 years. Participants will also be requested to complete a questionnaire on the knee function before surgery and after 1, 2, 3, 6, 9, 12, 18 months, annually up to 8 years, and every second year until 20years after surgery. There is a slight risk that these controls will keep participants in the patient-role, but the extra follow-up appointments may also be reassuring to participant that their knees are checked regularly, and that they have access to a specialist, if it would not go as expected after surgery.

Participants will, compared to a normal course of treatment, have made one extra unconventional special radiograph (0.2 mSv) of the knee prior to surgery. Postsurgical conventional radiography will be performed (0.056 mSv) after 2, 5 and 10 years. This represents a total additional radiation dose of 0, 2168 mSv. This extra radiation dose results in an increased risk of 0.00125% (1: 80.000) of developing fatal cancer.

• Research ethics The two treatments being compared in this study are considered in advance to be equal. Each treatment has, however, clear advantages and disadvantages. Clarifications of these factors are of obvious interest for future patients and the society. Neither the coordinating investigator nor local investigators have economic interests in the study results.

Referring to the risks and disadvantages mentioned above, there are in this study minimal risks associated with the study itself, and the disadvantages are also estimated as minimal. Overall considered, these risks and disadvantages are small enough that the study's results fully justify the conduct of the study. The results of the study should be of benefit to future MU-OA patients and thereby also for the use of resources in healthcare both at home and abroad.

Inclusion Criteria:

Anteromedial OA of the knee as diagnosed on posteroanterior and lateral projections of the knee, and symptoms of sufficient severity to justify arthroplasty. Each patient can only participate with one knee, but may be eligible with both knees. Eligibility is independent of possible previous procedures to index or contralateral knee, and to age, sex, occupation and etiology of the anteromedial OA.

Bone og Bone medially on pre-operative radiographs.

Exclusion Criteria:

1. knee instability caused by either cruciate or collateral ligament insufficiency,
2. rheumatoid or other inflammatory arthropathy, 3) lateral patellar subluxation on skyline views with articular cartilage thinning of the patellofemoral joint. Specific exclusion criteria for study participation are: 1) age below 18 years of age, 2) limited possibilities of understanding and assessing study information (e.g. severe psychiatric or mental conditions and non-Danish speaking), 3) occupational or personal relation to surgeons responsible for treatment (e.g. employment in participating orthopedic departments).

This will be decided at a meeting with fellow surgeons once inclusion has started.

Stratified, permuted block randomization with a 1:1 allocation ratio. Block sizes will be 4, 6, 8 and 10. Stratification will be two-dimensional, where surgeon is one dimension and patient sex the other. The randomization will be done immediately preceding surgery. Randomization is blinded to everyone (patient, staff, GP, physiotherapists etc.) except theatre staff, for the first year following surgery.