The purpose of this study was to identify causes of failure and rates of revision of the Oxford prosthesis (OXF) in New Zealand, by reviewing and comparing the uni-compartmental (UKA) and total knee arthroplasty (TKA) data from January 2000 to December 2005, as recorded in the New Zealand National Joint Registry.

Eighty one orthopaedic surgeons performed 2006 Oxford UKAs (64% of all UKAs). The revision rate was 4.7%. This compared with a revision rate of 4.8% for all UKAs combined, and 1.6% for TKA. UKA (3122) made up 13% of all knee arthroplasties (24 260). The most common reasons for revision of the OXF were aseptic loosening (45%), unexplained pain (33%) and bearing dislocation (12%). Unexplained pain as the only reason for revision (33%) was significantly different (p = 0.001) from the TKA rate (23%).

Deep infection as a cause for revision was 0.20% for the OXF compared with 0.48% for TKA (p=0.07). The patient- generated Oxford scores at six months after operation were rated excellent or very good (Field et al, 2004) in 68% of OXF compared with 62% TKA patients (p = 0.001).

Five higher-use OXF surgeons (12 or more/year) performed 25.1% of the operations with a revision rate of 0.99%. Ten high- use surgeons (eight to 11/year) performed 28.1 % of operations with a revision rate 4.6%. Thirty medium-use surgeons (two to seven/year) performed 39.0% of the operations with a revision rate of 6.4%. Thirty-six low-use surgeons (one or less/ year) performed 7.8% of the operations with a revision rate of 8.3%. The difference in revision rate between the higher-use surgeons (one operation/month) and all the other three lower use groups was significant (e.g. p=0.0006 higher/low)

The early revision rate for the OXF was 2.9 times greater than that for TKA. However, higher-use surgeons (i.e. those performing one/month or more) had a revision rate comparable to TKA. Deep infection was lower and six month function scores were higher for OXF compared with TKA. Unexplained pain as the only reason for revision was significantly higher for OXF compared with TKA.

To identify frequency and patterns of Oxford Phase 3 UKA failure in New Zealand through analysis of national primary and revision data.

Retrospective audit examining all revision Oxford Phase 3 UKAs recorded in the New Zealand National Joint Register from January 2000 to October 2003 were analysed along with surgeons’ clinical notes and patient x-rays.

Seventy-three Orthopædic Surgeons performed 1216 Oxford UKAs. The average age was 66.4 years (range 35–94). Osteoarthritis was the primary diagnosis for 1163 (96%) patients. Mean time to revision was 437 days (14.4 months). The early revision rate was 2.2% (n=27). The most common reasons for revision were aseptic loosening (n=7, 26%), bearing dislocation (n=5, 19%) and pain (n=4, 15%). The deep infection rate was 0.16% (2/1216). Eighteen surgeons (high use > 8 UKAs/year) performed 787 (64%) operations, with a revision rate of 1.5%. Twenty-two surgeons (low use ≤ 1 UKA/year) performed 38 (3%) operations, with a revision rate of 8%. This was statistically significant, p= 0.03 (odds ratio 5.7).

The early revision rate for the Oxford UKA is 1.4 times greater than TKA. High use surgeons revision rate is lower than TKA. An inverse relationship between failure and surgeon experience exists. This confirms Swedish Knee Arthroplasty Register findings.