Unicompartmental knee replacement (UKR) is technically challenging, but has the advantage over total knee replacement (TKR) of conserving bone and ligaments, preserving knee range of movement and stability. Computer navigation allows for accurate placement of the components, important for preventing failures secondary to mal-alignment. Evidence suggests an increase in failure rates beyond 3 degrees of coronal mal-alignment.

Our previous work has shown superior functional scores in those patients having undergone UKR, when compared with those having had TKR. However, to a certain extent, this is likely to be due to differences in the two cohorts. Those selected for UKRs are likely to be younger, with less advanced and less widespread degenerative disease. It is almost inevitable, therefore, that functional outcomes will be superior. We aimed to compare the functional and radiological outcomes of UKR vs TKR in a more matched population.

Ninety-two patients having had one hundred consecutive computer navigated UKRs were reviewed both clinically and radiographically. The Smith & Nephew Accuris fixed-bearing modular prosthesis was used in all cases, with the ‘Brainlab’ navigation system. For our comparative group we identified patients who had actually undergone navigated TKR several years ago, but who, in retrospect, would have now been offered a UKR in line with our current practice. These patients were identified following review of pre-operative radiographs and operation notes, confirming degenerative disease confined mainly to one compartment of the knee, in the absence of any concern as to the integrity of the anterior cruciate ligament. This sub-group of patients were also assessed clinically and radiographically. Mean follow-up for the UKR group was 25 months, (range 8–45.) For our TKR sub-group, nineteen patients were identified. Average length of stay for the UKR group was 3.7 days, (range 2 to 7,) and for the TKR group this was 5.2 days, (range 3 to 10.)

Functional scores (Oxford Knee Score) were good to excellent for the majority of patients in both groups, although they were significantly better in the UKR group. Mean Oxford Knee Score in the UKR group was 7.5, (0–48, with 0 being best.) Mean score in the TKR sub-group was 12.1. (p = 0.02)

Reliably comparing TKR with UKR is difficult, due to the fundamental differences in the two groups. We have endeavoured to match these two cohorts as best possible, in order to compare the outcomes of both. Our use of computer navigation in both groups allows for accurate prosthesis placement. When measuring component position, there were no ‘outliers,’ outside of the widely accepted three degrees of deviation.

We propose that, with the correct patient selection, UKR gives a better functional outcome than TKR. Longer-term follow-up of our UKR group is required to monitor the onset of progressive arthrosis in other joint compartments, although our early results are very encouraging. Furthermore, we advocate the use of computer navigation to firstly allow for more accurate component positioning, and secondly to make challenging UKR surgery less technically demanding.

Computer navigation has the potential to revolutionise orthopaedic surgery, although according to the latest 7^th Annual NJR Report, only 2% of the 5 800 unicompartmental knee replacements (UKRs) performed in 2009 were carried out using ‘image guidance.’ The report also states an average 3-year revision rate for UKRs of 6.5%. Previous NJR data has shown that this figure rises up to 12% for certain types of prosthesis. We suspect that a significant proportion of these revisions are due to failure secondary to component malpositioning. We therefore propose that the use of computer navigation enables a more accurate prosthesis placement, leading to a reduction in the revision rate for early failure secondary to component malpositioning. Our early results of one hundred consecutive computer navigated UKRs are presented and discussed.

Ninety-two patients having had one hundred consecutive computer navigated UKRs were reviewed both clinically and radiographically. The Smith & Nephew Accuris fixed-bearing modular prosthesis was used in all cases, with the ‘Brainlab’ navigation system. Pre-operative aim was neutral tibial cut with three degrees posterior slope. Post-operative component alignment was measured with PACs web measuring tools. Patients were scored clinically using the Oxford Knee Score.

Our patient cohort includes 54 male knees and 46 female knees. Average age is 66.6yrs. Average length of stay was 3.7 days, (range 2–7.) With respect to the tibial component, average alignment was 0.7° varus, and 2.32° posterior slope. All components were within the acceptable 3 degrees deviation. Functional scores are very satisfactory, with an overall patient satisfaction rate of 97%.

To date, only one UKR has required revision. This was due to ongoing medial pain due to medial overhang, not related to computer navigation. There was one superficial infection, with full resolution following a superficial surgical washout, debridement and antibiotics. Unlike complications reported in the NJR, we report no peri-prosthetic fractures or patella tendon injuries.

Our results demonstrate accurate prosthesis placement with the use of computer navigation. Furthermore, clinical scores are highly satisfactory. Our current revision rate is 1% at a mean of 27 months post-op. Although longer-term follow-up of our group is required, our results compare very favourably to statistics published in the NJR, (average 3-year revision rate 6.5%.) The only major differences appear to be the type of prosthesis used and the use of computer navigation. It is our proposal that computer navigation reduces the number of revisions required due to component malpositioning and subsequent failure. Furthermore, we believe that this challenging surgery is made easier with the use of computer navigation. We expect our longer-term results to show significant benefits of computer navigation over conventional techniques.