Validation of American College of Rheumatology classification criteria for knee osteoarthritis using arthroscopically defined cartilage damage scores

OBJECTIVE To validate the ability of the American College of Rheumatology (ACR) clinical classification criteria and the ACR clinical plus radiographic classification criteria for osteoarthritis of the knee to predict articular cartilage damage. METHODS Ninety subjects with knee osteoarthritis (OA) who were enrolled in a prospective study determining the therapeutic efficacy of arthroscopic irrigation were characterized as to whether they fulfilled the ACR clinical classification criteria or the ACR clinical plus radiographic classification criteria. Ten rheumatoid arthritis (RA) patients were included as controls. Cartilage damage was defined using the ACR/Knee Arthroscopy Osteoarthritis Scale (ACR/KAOS) system, which is a validated outcome instrument for knee OA based on arthroscopic visualization. Mean values of the damage scores in each group were calculated and compared by t-test to determine statistical significance between the 3 groups. RESULTS The mean ACR/KAOS score for the 10 RA patients was 1.8 [SD 1.22; range 0 to 4]. Of the 90 OA patients who underwent arthroscopy, only 73 patients had sufficient videotape to make an accurate assessment by the blinded assessor. The mean ACR/KAOS score for the 6 OA patients who fulfilled only the ACR clinical classification was 17.4 [SD 11.3; range 5 to 34.3] and the mean ACR/KAOS score for the 67 patients who fulfilled the ACR clinical plus radiographic classification criteria was 42.0 [SD 29.1; range 5.1 to 118.4]. These differences were statistically significant (RA versus OA clinical P = 0.02; RA versus OA clinical + radiographic P <= 0.01). Nonparametric multivariate analysis did not reveal significant correlations between ACR/KAOS scores and WOMAC global scores (r = 0.11, P = 0.39), patient VAS (r = 0.29, P = 0.022), and age (r = 0.29, P = 0.08).