The proportion of osteoarthritic knees deemed suitable for medial unicompartmental arthroplasty depends on the selection criteria. Early in our experience, using the indications given above, we found that about a quarter of knees with OA severe enough to need joint replacement were suitable for OUKA. Now we find that about half of the knees are suitable, presumably because patients are presenting with less severe arthritis. The criteria are evidence-based in that they have been employed with little variation since the 1980s and are attested by a number of long term survival studies of several independent cohorts (Svard & Price 2001). In addition we have reviewed 1000 cases with 15 year follow up and found that two thirds have a Kozinn and Scott (1989) contraindication yet there is no difference in outcome or survival between those with and without the contraindications. This confirms that the Kozinn and Scott contraindications do not apply to the OUKA.
In contrast, Stern et al. (1993) applied the often quoted criteria of Kozinn and Scott (1989) to knees with OA severe enough to warrant joint replacement and found only 6% suitable for UKA.
Ritter et al. (2004) examined retrospectively the preoperative and intraoperative records of 4021 osteoarthritic varus knees treated by TKA to determine how many of them would have been ‘ideal candidates’ for UKA. They first applied anatomical criteria, i.e. ‘… a normal ACL, a normal lateral meniscus, no more than mild osteophytes in the patellofemoral and/or lateral compartment, no more than grade 2 chondromalacia in the patellofemoral compartment and no more than grade 1 chondromalacia in the lateral compartment’. Only 6% met these criteria. They then applied clinical criteria attributed to Kozinn and Scott (1989), i.e. patients should be more than 60 years old and weigh less than 82 kg, the flexion contracture should be less than 5°, and varus deformity less than 10°. This reduced the number to 2% of the original cohort.
If, in fact, UKA were useful in as few patients as these studies suggest, few surgeons would ever see enough suitable patients to become adept in its use. However, the detailed data provided by Ritter et al. (2004) reveal that the criteria that they applied retrospectively and those that we have used prospectively are so dissimilar that the 15-fold difference between our practice and their prediction is readily explained.
Strikingly, they found that 45% of their cohort of knees had a normal ACL and an undefined percentage would have had an ACL that was functionally normal although slightly damaged; both of which are appropriate for an OUKA. However, this large group was diminished to 6.1% because of the presence of lateral compartment chondromalacia (74%), a degenerative lateral meniscus (69%), patellofemoral chondromalacia (55%), lateral compartment osteophytes (33%), or patellofemoral osteophytes (30%), all of which can occur in middle-aged people with no knee complaints and none of which has played a part in our selection process. The clinical criteria (overweight (53%), more than 5° flexion deformity (32%), age (12%)) that further reduced the number of suitable knees to 2.2% have also all been ignored in our practice.
The contraindications criticised above were elaborated by surgeons whose experience was with fixed-bearing prostheses. Their criteria have been widely accepted and few knees outside these narrow limits may have been treated by UKA. Consequently, there is little evidence other than ours to contradict them. However, our experience has been with a meniscal-bearing prosthesis implanted (since 1987) by a fully instrumented method. The proportion of suitable cases may not be the same for fixed-bearing devices that are susceptible to medium-term polyethylene wear and patella problems and, until recently, have been implanted with little instrumental assistance.
Summary
Patient selection is simple when it is considered in the light of positive indications for OUKA rather than its contraindications. If a patient has only anteromedial OA and symptoms severe enough to justify the surgery and associated risks, we would do an OUKA and there are no contraindications. AMOA is clearly defined and to make this diagnosis a patient should be shown to have medial osteoarthritis with bone-on-bone, a functionally intact ACL and MCL and full thickness cartilage laterally.
For completeness, the following list brings together all the contraindications, most of which have already been discussed or implied.
General
All the general contraindications recognised in the practice of joint replacement surgery apply to unicompartmental replacement and are not rehearsed exhaustively here. Among them, OUKA (like TKA) is likely to fail in limbs with sensory and/or motor neurological impairment.
The use of a tourniquet during the operation, almost a necessity for OUKA (Phase 3) performed through a small incision, is dangerous in limbs with vascular insufficiency. It may also be dangerous to apply tourniquets to both lower limbs, and for this reason we seldom perform bilateral OUKA in one session, preferring to stage the procedures with an interval of not less than 6 weeks.
Particular
Unicompartmental arthroplasty is contraindicated in the inflammatory forms of arthritis because they are diseases of the synovium and therefore cannot be limited to one compartment.
We have already noted the anatomical contraindications listed below:
➤ absent or severely damaged ACL (or PCL or MCL)
➤ failure to demonstrate eburnated bone-on-bone contact in the medial compartment
➤ intra-articular varus not fully correctable
➤ mediolateral subluxation, not corrected on valgus-stressed films
➤ flexion deformity >15°, which is suggestive of ACL deficiency
➤ flexion range <100° (under anaesthesia)
➤ erosion of central cartilage in the lateral compartment
➤ bone loss with eburnation and grooving in the lateral part of the patellofemoral joint
➤ previous valgus tibial osteotomy (see Chapter 8).
