Long-term trends in the Oxford knee score following total knee replacement 1. D. P. Williams, MBChB(Hons), MRCS, Specialist Registrar 2. D. W. Murray, BChir, FRCS (Orth), MD, Professor of Orthopaedic Surgery 3. A. J. Price, BSc, MSc, PhD, Professor of Orthopaedic Surgery Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, BOTNAR Research Centre, University of Oxford, Windmill Road, Oxford OX3 7LD, UK. 4. C. M. Blakey, MBChB, MRCS, Specialist Registrar Sheffield Teaching Hospitals NHS Trust, Herries Road, Sheffield S5 7AU, UK. 5. S. G. Hadfield, PhD, Business Research Manager 6. R. E. Field, PhD, FRCS, FRCS(Orth), Consultant Orthopaedic Surgeon, Director of Research The South West London Elective Orthopaedic Centre, Epsom General Hospital, Dorking Road, Epsom KT18 7EG, UK Cite this article: Bone Joint J 2013;95-B:45–51. Discussion The OKS is widely accepted as a reliable and valid measure of outcome after knee replacement and exhibits superior reliability and validity to similar outcome measures in terms of response rate and ease of use. Although post-operative OKS values are reported in the literature, these usually include measurements reported at singular time points, therefore providing little indication of the changes that may have occurred over time. Consequently, defining what constitutes a ‘normal’ or ‘expected’ OKS value is often based on an individual’s experience, which makes auditing and comparison of results between surgeons difficult. One of the largest reported series relating to the OKS is based on 20,885 patients from the recently established United Kingdom arthroplasty database but so far only preoperative and six-month data have been published. However, the reported data closely resemble our pre-operative (19.0 vs 19.5), and initial post-operative OKS scores (33.6 (six-month) vs 34.3 (one-year)), suggesting that our cohort is representative of the general population. Further data will become available over the next few years as this database continues to grow. In line with the trend for the overall OKS, most of the individual components also demonstrated an initial postoperative peak at one to two years, followed by a subsequent gradual decline. The greatest post-operative changes were seen for pain severity and the presence of a limp. Difficulty in washing and drying appeared to be the least specific component, with minimal change from pre- to post-operative scores. Despite exhibiting the greatest improvement, pain severity remained one of the lowest-scoring components post-operatively. These observations mirror those seen in a study by Lim et al examining pre-operative and two-year OKS data. We observed in our series that over two-thirds of patients reported residual pain following surgery. Interestingly, we also found that maximal improvement in the night pain component was not achieved until four years post-operatively. We cannot fully explain why this should be the case. One of the key objectives of knee arthroplasty is relief of pain, and residual pain is considered one of the factors associated with dissatisfaction. Patients should therefore be made aware that pain is unlikely to be completely abolished following surgery, and that night pain in particular may persist for a prolonged period. However, it should be noted that the pain component scores observed reflect only mild or very mild pain. Kneeling ability was the worst-scoring component at almost all post-operative assessments. Low scores for the kneeling component have previously been described. Pynsent et al found no difference between kneeling component scores pre- and post-operatively, and suggested that patients cannot kneel before surgery and still cannot kneel afterwards. Conversely, Rothwell et al reported significant, albeit small, improvements between six months and five years, but had no pre-operative values for comparison. It has been suggested by some that these problems are not as a result of ‘real’ functional issues but rather to perceived difficulties reflecting patient apprehension, or advice against such activity by their surgeon. The initial and significant improvement observed in our study suggests that some attempt to kneel is made by patients until at least one year after surgery. Although avoidance of kneeling may partly explain the lower scores, if this was the only cause a more stable trend in kneeling scores might be anticipated. The initial improvement in the first 12 months, followed by a rapid decrease over the first five years, suggests that there may also be a functional element to kneeling problems with time. The pre-operative OKS has been shown to correlate with the subsequent post-operative outcome. This is supported by our study, with greater post-operative OKS evident in those with the best pre-operative OKS. It should be emphasised that this does not necessarily represent a poor outcome for those in the lowest quartiles, as it still represents a significant improvement from pre-operative state as indicated by the post-operative change in scores. Nevertheless it is important that surgeons recognise the impact of pre-operative OKS on post-operative outcomes when evaluating their results and in informing patients of their likely outcome. Our results show that women have poorer scores both pre- and post-operatively, which has previously been observed with objective outcome measures. Surgery is also more likely to be delayed in women until their symptoms are more severe. In the presence of a worse preoperative OKS, it is therefore not surprising that lower post-operative OKS values were also observed in women. We found that the overall OKS values were significantly lower for those aged <60 years, with a suggestion of a more rapid decline over time. However, the overall linear trend in OKS across different age groups was not significantly different. Higher revision rates are recognised following TKR in younger patients. Unfortunately, we did not have accurate data relating to revision status for our cohort but this suggestion of a trend for lower scores over time in younger patients may be a reflection of a tendency for higher revision rates in this group. The association between BMI and the outcome from arthroplasty is controversial, with some studies suggesting inferior clinical outcomes in the morbidly obese. BMI data were not routinely collected at the beginning of our outcome programme, and so in this study data were limited. However, we observed that a BMI >35 kg/m2 appeared to be associated with a trend for lower OKS at all time points. At first glance this could be used to support the theory that high BMI leads to a worse outcome, and in turn to justify restricting surgery to non-obese patients. However, caution is required in interpreting these data. Although our data show inferior scores following surgery, the actual post-operative improvements in OKS were similar in all the groups. There are limitations to this study, primarily related to the diverse nature of the population analysed and the incomplete data. We have, however, accounted for the subject variations within our cohort using mixed model analyses, and believe the results provide a valid representation of the OKS trend with time in the general population. Factors additional to those examined have been shown to influence the OKS, but owing to the limitations of our database we could not evaluate the possible confounding effects of these other variables in this study, and this should be considered when interpreting these results. Further information will become available as the national United Kingdom database continues to grow, but in the absence of any other significant published evidence, we believe that this study still provides a valid and meaningful insight into the ‘expected’ trends for the OKS, and provides a benchmark for surgeons to compare and understand patterns in their longer-term outcomes. Summary We have described the trends observed in the overall score, as well as the individual components of the OKS over a ten-year period following TKR. The maximum OKS is not achieved until two years post-operatively, following which there is a gradual decline. Problems with kneeling are extremely common, and mild persistent pain may be anticipated in the majority of patients, particularly at night. This information may assist surgeons in advising patients of their expected outcomes, as well as providing a comparative benchmark for evaluating longer-term outcomes following knee replacement.