How to assess ankle osteoarthritis: comparison of the Kellgren and Lawrence scale with functional outcome and digital image analysis

Introduction Currently, a validate scale of ankle osteoarthritis (OA) is not available and different classifications have been used, making comparisons between studies difficult. In other joints as the hip and knee, the Kellgren-Lawrence (K&L) scale, chosen as reference by the World Health Organizations is widely used to characterize OA. It consists of a physician based assessment of 3 radiological features: osteophyte formation, joint space narrowing and bone end sclerosis described as follows: grade 0: normal joint; grade 1: minute osteophytes of doubtfull significance; grade 2: definite osteophytes; grade 3: moderate diminution of joint space; grade 4: joint space greatly impaired, subchondral sclerosis. Until now, the K&L scale has never been validated in the ankle. Our objective was to assess the usefulness of the K&L scale for the ankle joint, by determining its reliability and by comparing it to functional scores and to computerized minimal joint space width (minJSW) and sclerosis measurements. Additionally we propose an atlas of standardized radiographs for each of the K&L grades in the ankle. Methods 73 patients 10 to 20 years post ankle ORIF were examined. Bilateral ankle radiographs were taken. Four physicians independently assessed the K&L grades and evaluated tibial and talar sclerosis on anteroposterior radiographs. Functional outcome was assessed with the AOFAS Hindfoot score. Bone density and minJSW were measured using a previously validated Ankle Image Digital Analysis software (AIDA). Results The interobserver reliability, for the K&L stages was 0.60 (intraclass correlation coefficient) indicating moderate to good agreement. The mean AOFAS hindfoot score decreased substantially (p = 009) and linearly from 99.3 in K&L grade 0 to 79.5 points in K&L grade 4. The minJSW assessed by AIDA was similar among grades 0 to 2 (between 2 and 2.5mm), but significantly lower in grade 3 (1.8mm) and in grade 4 (1.1mm). A decreased minJSW less than 2mm, commonly used as a threshold for the assessment of hip and knee OA, was found in 77% of K&L grades 3–4 compared to 33% of grades 0–2, sensitivity 77.4% and specificity 66.7%. Physician based assessment revealed that subchondral sclerosis was present in 16% of K&L grade 1 patients, 52% of grade 2, 70% of grade 3 and 100% of grade 4 patients. No correlation could be found between physician based assessment and digital image analysis of subchondral sclerosis. Conclusions Interobserver reliability in assessment of ankle OA using the K&L scale was similar to other previously described joints. OA progression correlated with functional diminution. Joint space narrowing assessed AIDA as well as the cut-off of 2mm correlated well with the K&L scale. Overall, we recommend the use of the K&L scale for the radiographic assessment of ankle OA