Summary

The anatomical distribution of osteoarthritis influences Oxford Knee Score in patients with end stage osteoarthritis, so that despite better pre-operative function, patients with isolated patellofemoral arthritis have a smaller improvement in function in the first 12 months postoperatively.

Abstract

Background

The prevalence of knee Osteoarthritis (OA) is increasing globally and is characterised by knee pain and radiological changes including joint space narrowing, subchondral sclerosis and osteophyte formation. With no known cure, total knee replacement (TKR) remains a reliable treatment for end stage OA. However, not all TKR patient patients benefit equally. Determining the pre-operative anatomical pattern of OA of the knee might be associated with post-operative outcomes such as knee pain and function, but clinical studies looking at this aspect are limited.

Objectives:
1) To develop a valid reliable system for categorising the anatomical distribution of radiological OA.
2) To characterise how the anatomical distribution of OA impacts OKS in patients before and after TKR.

Methods

We devised a novel anatomical categorisation of OA, combining Ahlbaeck (tibiofemoral) and Iwano (patellofemoral) radiological classifications to create 4 categories: isolated medial, lateral, patellofemoral (PFJOA) and multicompartmental OA. We analysed the association between anatomical distribution of OA in a cross-section of patients planned for TKR and their pre-operative OKS, and 12-month post-operative OKS, using logistic regression analysis adjusting for age, sex and body mass index (BMI).

Results

Of 214 patients, 53.6% were females, average age 69.7 years, average BMI of 30.4. The study included 351 knees with an average pre-operative OKS of 26.1. Of these, 178 had medial, 58 had lateral, 43 had PFJOA, and 72 had multicompartmental OA. In adjusted analyses prior to TKR, participants with PFJOA showed higher OKS [ßeta 95 % 1.87: Confidence Interval (CI):-0.60, 4.32]), whilst participants with OA in the medial compartment had lower OKS (ßeta -0.62 CI:-2.19, 0.94) compared to participants with OA in other compartments. When we stratified data by sex: OKS score was negatively associated with medial OA in males (ßeta -2.27 CI: -4.51, -0.02) and multicompartmental OA in females (ßeta -2.60 CI: -4.93, -0.22). Conversely OKS was positively associated in males with PFJOA (ßeta: 4.90 CI:0.54, 9.26) and younger age (<69 years) (ßeta: 27.6, CI:5.75, 49.5). However, in older participants, OKS was negatively associated with PFJOA (ßeta: -14.2 CI: -26.6, -1.74). Lastly, in those with PFJ OA and higher BMI trended towards lower OKS score (ßeta -11.7 CI: -24.6, 1.20). Postoperatively, there was no statistically significant difference in the average improvement in OKS scores of 43.2, however there was a trend to poorer OKS scores amongst patients with PFJOA.

Conclusion

This anatomical classification of radiographic OA demonstrated high inter- and intra-observer correlation. This study demonstrated association between the anatomical distribution of OA and pre- and post-operative OKS. Pre-operative scores were substantially modified by sex and age, less by BMI and mostly for patients with PFJOA. Patients with isolated PFJOA demonstrate better preoperative function, however the OKS in this group improves less following TKR. This information is helpful for surgeons counselling their patients about expectations following TKR.

Level II evidence: Development of diagnostic criteria on consecutive patients (all compared to “gold” standard) and no negative criteria.