Predictive values of WOMAC, KOOS, and SF-12 score for knee arthroplasty: data from the OAI.

Abstract

A number of validated scores are available to monitor clinical outcome after knee arthroplasty. The current study examines whether WOMAC, KOOS, or SF-12 scores have a predictive value to identify patients that require total knee arthroplasty within 1year. The Osteoarthritis Initiative includes 4684 patients in two groups: incidence subcohort (Group A, n=3284) and progression subcohort (Group B, n=1400). A total of 278 patients reached the endpoint "unilateral knee replacement" (Group A=96; Group B=182). WOMAC, KOOS, and SF-12 scores at the maximum of 12months prior to surgery were used to analyse predictive values. ROC analysis and the diagnostic accuracy for these scores were reported. Group A: WOMAC and KOOS score as well as the "physical component summary" of SF-12 showed an area under the curve (AUC) between 0.8 and 0.9 (high diagnostic evidence) to predict total knee replacement. The KOOS "symptoms" (0.67) and SF12 "function" (0.79) showed lower values. The KOOS QoL score showed the highest predictive value with an AUC of 0.84 (CI 95% 0.77-0.91, p<0.001) resulting in a sensitivity of 0.88 and a specificity of 0.72. Similar numbers were achieved by the total WOMAC score with an AUC of 0.85 (CI 95% 0.82-0.88, p<0.001) resulting in a sensitivity of 0.77 and a specificity of 0.82. The SF-12 physical component score had an AUC of 0.83 (CI 95% 0.79-0.87, p<0.001) resulting in a sensitivity of 0.87 and a specificity of 0.68. Different combinations of scores increase the positive likelihood ratio (up to 18) and specificity (up to 0.97). Group B scores showed lower AUCs, sensitivities and specificities. Combinations of WOMAC, KOOS, or SF-12 scores accurately predict patients that require knee arthroplasty within 1year. In addition to other clinical parameters (physical examination, radiographs), they can guide patient and surgeon during the treatment of arthritis of the knee. II.