The Oxford medial unicompartmental arthroplasty.

Murray DM, Goodfellow JW and O'Connor JJ. J Bone Joint Surg (Br) 1998;80-B:983-9.

This is the editor's interpretation of a paper published in the orthopaedic literature in 1988 - our attempt to make relevant medical articles accessible to lay readers. If you wish to read the original it is easy to ask your librarian to obtain a reprint for you from any medical library.

This is a report of the 'survival rate' of a particular partial knee replacementOxford unicompartmental arthroplasty - within the patient group of one of the authors (JWG) over a period of several years, and where inclusion criteria were carefully specified. Also, via a literature review, the durability of this particular knee replacement is compared with others, both partial and total.

[Ed: Just to clarify terminology - a prosthesis is the artificial implant itself, while the term arthroplasty is the procedure inclusive of the implant. The term arthroplasty could mean an implant of any joint, but a knee arthroplasty means an implant of the knee, or knee replacement. A knee replacement can be partial (only part of the bony contact surfaces of the joint replaced) or total (replacing all the bony contact surfaces). A unicompartmental knee replacement ('uni') is a partial knee replacement, the word compartment being applied to a single point of contact of bones within the joint. Where the femur bone makes contact with the tibia bone there are in fact two discrete points of contact, as the femur has two rounded ends with a gap between them. The inner contact area is called the medial compartment, while the outer contact area is called the lateral compartment.

The authors start their paper by pointing to previous literature that had demonstrated that for carefully selected cases of knee osteoarthritis (OA) -

  • a uni is less invasive than a total knee replacement (TKR)
  • in a uni the cruciate ligaments are preserved
  • a uni gives a better range of movement and more physiological function
  • there are fewer complications of the surgery itself
  • the implant is cheaper

but past studies had also shown that the uni had a higher failure rate than TKR, commonly due to wear of the polyethylene component. [Ed: the polyethylene component is the bit sandwiched between the all-metal bit on the femur and the metal-polythene bit on the tibia. Prior to the Oxford uni, the polythene was 'fixed' to the metal part of the tibial component and could not move - a 'fixed bearing'.]

The Oxford uni was different from the designs that preceded it in having a mobile bearing - the polyethylene central bit acting  more like a meniscus in being able to move in relation to femur and tibia. The mobile bearing was thus also called a 'meniscal bearing'. The Oxford uni was also 'fully congruent' - the polyethylene bit was shaped to fully accomodate the fact that the lower ends of the femur are rounded while the upper end of the tibia is flat. 

The first Oxford unicompartmental arthroplasty was performed in 1982, and by 1991 research had established that it was suitable when the OA was antero-medial in its distribution and the anterior cruciate ligament was intact. By the time of this report (1998) Oxford unis had been around for over a decade, and surgeons were comparing the 10-year survival rate of one type of uni against another, as well as against TKR.


This Study

The patients included in this study were all those of one of the authors (JWG) and all had had a uni some time between 1982 and 1992. They had been followed up annually since the original operation. Several papers had already been published about the findings along the way, and the paper gives details of the references if you want to follow this up.

In this particular study the authors had selected only patients who had had medial-sided implants. All the patients included in the study had -

  • had their operations by or under the direct supervision of one surgeon
  • full thickness loss of cartilage on the medial side
  • normal thickness of cartilage on the lateral side
  • any bowing of the knee (varus deformity) was correctable when the bones were subjected to stress under X-ray
  • normal anterior and posterior cruciate ligaments

[Ed: This group of criteria had been previously established in the literature as descriptive of 'anteromedial OA of the knee'. There is quite an important issue here. As the cartilage is destroyed anteriorly (in the front) the patient's knee may tend to bow ('go into varus'), but if the cartilage posteriorly (at the back) is still intact then this bowing usually is not fixed but can be corrected - abeit temporarily - by applying force while observing the alignment correction under X-ray. This is called a 'stress X-ray'. If the posterior cartilage, however, is also destroyed then the bowing (varus) may become fixed and uncorrectable under stress X-ray. This is an important concept because if the varus is fixed then there will be no room to accomodate the prosthesis and its bearing.]

The authors point out that they did not exclude from the study patients with OA of the patellofemoral joint. Neither was obesity an exclusion criterion, nor the presence of crystals in the joint.


The Group

  • 114 patients were included in this study
  • Of them, the fate of 143 prostheses was determined
  • All patientss had had their operations between 1982 and 1992
  • The average age was 70.2 years (range 34.6 - 90.6)
  • The ratio of male to female was 1:1.2
  • The mean follow-up was 7.6 years (maximum 13.8)



Of the 143 procedures, five had to be revised to TKR for the following reasons -

  • one was infected with loosening and pain
  • two had cartilage loss in the lateral compartment with pain
  • one had a loose femoral compartment with pain
  • one had unexplained pain that continued after the revision (no cause established)

Of these five revisions, three had a good result after TKR, but the other two continued to have problems - the one with infection and the other with unexplained pain. Besides the five revisions, there was one patient in whom the bearing dislocated, but it was resolved with a manipulation under anaesthesia with no further problem.

Taking into account any of the knees that the authors thought may yet be 'at risk', they gave a predictive survival in this group of 97% of the unis - the best recorded to that date. But, the authors stress that these results have been obtained in a group meeting stringent selection criteria and that it is important that surgeons pay heed to these selection criteria if they are to obtain equivalent results. They also point out that the uni is a more technically demanding procedure than a TKR, with a smaller margin for error. The fact that the Oxford has an unconstrained bearing means that it merits particular attention to ensure that dislocation does not occur - the surgeon has to choose between several sizes of tibial components and several thicknesses of meniscal bearing - and dislocation of the bearing is more common where the surgeon is inexperienced.

The authors go on to describe the benefits of the Oxford meniscal bearing prosthesis compared to other designs of uni. Its unconstrained bearing - which can slide around freely on the flat shelf of the tibial component - allows more movement of the tibia but with less wear of the polyethylene than prostheses where the polyethylene is constrained. Also they pointout that the bearing is fully congruent - with its upper end concave to snug the rounded femoral component while the lower end is flat to slide freely on the flat tibial component - and this congruency also results in less polyethylene wear. [Ed: the issue of wear of the polyethylene component is important, as debris can build up in the joint and also as the prosthesis can loose and fail.]

The article goes into these issues in greater depth and has many references at the end and the interested reader my benefit from obtaining a copy of the original article via any library.