Scoring prevalence and severity in gonarthritis: the suitability of the Kellgren & Lawrence scale

In clinical and epidemiological studies the Kellgren & Lawrence (K&L) scale was mainly used in the past to determine the radiological prevalence and severity of gonarthritis. The dependency of this scoring system on the presence of osteophytes has been criticised, and owing to its varying reliability, its appropriateness in cross-sectional and longitudinal studies is questioned. It was the purpose of this study to test whether the dependency of the K&L scale on the presence of osteophytes is reflected in a higher correlation of the overall score with a separate scoring of 'osteophytes' than with other radiological features. Furthermore we compared the inter- and intrarater reliability with the reliability of individual radiological features. Knee radiographs of 40 patients were graded according to the K&L scale and for the presence and severity of five separate radiological features of osteoarthritis. We found a moderate correlation of the K&L scale with the radiological features of 'osteophytes', 'joint space narrowing' and 'flattening of condyles'. The intra- and inter-rater reliability of the K&L scale was higher than the reliability of the individual radiological features. The K&L scale seems not to be strongly dependent on the presence of osteophytes, i.e. there is not a higher correlation of the overall score with that feature than with other radiological features.