Background

Plain radiographs are routinely used to diagnosis and monitor the progression of knee osteoarthritis. However multiple studies show that the severity of OA on radiographic tools does not correlate with the clinical aspects of OA. There is a paucity of studies examining intra-operative classification of knee osteoarthritis, with the Outerbridge classification being the only surgical classification in widespread general use, and it is designed asan arthroscopic classification. The aim of this prospective cohort study is to propose and validate a new intra-operative classification system for knee osteoarthritis and determine its correlation with pre-operative imaging and patient-reported outcomes.

Methods

A prospective cohort study was performed on 100 patients undergoing TKA for knee osteoarthritis by five participating surgeons. A classification system was developed evaluating extent and distribution of cartilage loss, synovium status, osteophytes, and overall impression of severity. These data were recorded by the surgeon, and photos capturing four key views of the knee were taken for evaluation by 2 other observers. Pre-operative x-rays were assessed using Ahlbäck scores for medial and lateral compartments and the IWANO score for the patella. Patient-reported outcomes measures (PROMs - Oxford Knee Score (OKS), Forgotten Joint Score (FJS)) of pre-operative symptoms were collected and then correlated against both the Ahlback score, and the new intra-operative classification. Interclass correlation coefficients were used to assess intra- and inter-rater reliability, Pearson’s correlation to determine relationships between intra-operative and radiographic classifications, and regression analyses to evaluate associations between the radiographic and intra-operative classification systems, with pre-operative PROMS.

Results

100 patients were included, 58 males and 42 females with an average age of 69 ± 9 (range 42-85). Seventy-seven patients had varus knees, 8 neutral, and 15 valgus. Intra-operative classification system demonstrated high intra and inter-rater reliability for cartilage assessment (0.83-0.92 <0.001) and overall OA grading (0.78-0.88, p<0.001). Osteophyte assessment showed moderate reliability (0.34-0.77, p<0.001), synovitis evaluation had low inter-rater reliability (0.20-0.37, p=0.02). Correlations between intra-operative classification and t Ahlbäck score were modest: 0.40, (p <0.001) on the lateral side, 0.33, (p <0.001) on the medial side, and negligible for the patella (0.09, p = 0.41). X-rays underestimated the amount of arthritis in 56% of patella cases, 40% of medial compartment cases, and 36% of lateral compartment case. Ahlback scores were not predictive of pre-operative patient-reported symptom (p>0.05) whilst intra-operative findings of medial patella cartilage wear, the presence of an effusion, medial and lateral tibial cartilage wear, and patella osteophytes were shown to be predictive of the Forgotten Joint Score (FJS) (p<0.05). Additionally, medial tibia cartilage wear, posterior medial osteophytes, and the presence of an ACL were predictive of the Oxford Knee Score (OKS)(p<0.05).

Conclusion

This study establishes a reliable and valid intra-operative classification system for knee osteoarthritis which correlates with pre-operative patient reported symptoms. In contrast, pre-operative x-ray scores often underestimate the degree of intra-operative osteoarthritis in all compartments, and do not correlate with the severity of patient reported symptoms. Future application may allow better prediction of true knee OA severity.