A new quantitative measure for radiologic osteoarthritis of the lateral knee compartment distinguishes patients with longstanding lateral meniscectomy from non-pathological knees.

Abstract

Common radiologic scores to evaluate knee osteoarthritis (OA) have been widely used but are descriptive and may lack objectivity. The aim of this study was to develop a quantitative and objective radiologic measure for the evaluation of lateral knee compartment OA. Furthermore, we tested the reliability of this new measure and its correlation to well-accepted radiologic scores. This retrospective study was performed within the context of a multi-centre long-term follow-up (15-25years) of a patient cohort after partial or total lateral meniscectomy (n=36 knees). In addition, 99 radiographs of clinically and radiologically normal knees were obtained from a control group. Anteroposterior weight-bearing views (0°) and Schuss (45°) views were analysed. The joint height of the lateral knee compartment was measured on the lateral border (lateral joint space height) and in the centre (central joint space height, CJSH) and normalized with respect to the width of the lateral compartment (X). All measurements were taken independently by two observers, allowing for inter- and intra-observer reliability assessments. The results of the two groups were compared using an analysis of covariance. Finally, the correlations between the results and the Fairbank and Kellgren & Lawrence scores were determined using a Spearman ρ rank order correlation. The normalized joint space height in the centre of the compartment on the Schuss view (CJSH/X) showed the highest intra- (ICC=0.980) and inter-observer reliability (ICC=0.982). There was a significant difference in CJSH/X between the control (0.19±0.05) and the meniscectomized knees (0.08±0.07) (p<0.001). CJSH/X showed a significant decline of 11% per 10years in the meniscectomy group. A negative correlation could be found between CJSH/X and the Fairbank (ρ=-0.751; p<0.001) and Kellgren & Lawrence scores (ρ=-0.712; p<0.001). A cut-off value of 0.14 of CJSH/X was defined-representing one standard deviation below the mean of the control group-from which measurements were considered as pathologic. The normalized joint space height measured in the centre of the lateral knee compartment from a Schuss view, CJSH/X, was highly reproducible and showed a significant correlation to established radiologic scores. This new measure has the advantage of being objective and dimensionless and thus independent of the size of the radiograph. The normative values provided by our healthy control knees are useful to help establish an early diagnosis of radiologic lateral knee compartment OA. Retrospective diagnostic study, Level III.