Arthritis & Rheumatism, Volume 63,
November 2011 Abstract Supplement
Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Chicago, Illinois November 4-9, 2011.
Sensitivity and Specificity of the American College of Rheumatology and Europe League Against Rheumatism Response Criteria in Rheumatoid Arthritis for Changes Important to Patients: Is An ACR20 Meaningful?
Alba1, Maria I., Guthrie1, Lori C., Ward2, Michael M.
ACR and EULAR response criteria are widely accepted outcomes in rheumatoid arthritis (RA) clinical trials, but many question whether ACR20 responses represent important improvement in RA status, and emphasize ACR50 and ACR70 responses. It is not known how well response criteria correspond to changes that are clinically meaningful to patients. We investigated the sensitivity and specificity of ACR and EULAR responses to reflect patients' subjective assessment of improvement in RA activity.
In this prospective longitudinal study, we enrolled adults with RA who had active arthritis (evidenced by six or more tender joints and physician judgment) and were either starting a new medication or undergoing a dose escalation of their anti-rheumatic medication. Participants were examined twice, at baseline and at 1 month (for those treated with prednisone) or 4 months later (all others). At each visit we performed joint counts, physician global assessment, ESR, CRP, patient global assessment, HAQ disability index and pain scales. At the follow up visit, we asked patients to report if their overall arthritis status was improved, unchanged or worse, and computed ACR and EULAR responses. We used patient-rated global improvement as the standard.
We enrolled 214 patients (75% women, median duration of RA 6.5 years, mean age 52 years). At entry, the mean swollen joint count was 16, tender joint count 24, CRP 1.9, ESR 39, and DAS28 6.09. 136 (64%) patients reported improvement, 56 (26%) reported no change, and 22 (10%) said they were worse at the follow up visit. ACR20 responses were present in 41%, but only 8% had ACR70 responses (table). Specificity was high for each response criteria, but sensitivity was no higher than 0.55 (table).
|ACR 20||41%||.55 (.46, .64)||.83 (.75, .92)||.34 (.22, .46)|
|ACR 50||16%||.24 (.17, .32)||.97 (.93, 1.0)||.17 (.10, .24)|
|ACR 70||8%||.11 (.06, .18)||.99 (.96, 1.0)||.08 (.03, .13)|
|EULAR Moderate||42%||.55 (.46, .64)||.81 (.71, .91)||.31 (.19, .43)|
|EULAR Good||9%||.11 (.05, .17)||.94 (.88, 1.0)||.04 (0, .10)|
Sensitivity and specificity of the ACR20 was similar in subgroups by age, gender, and ethnicity, but sensitivity (0.64) and specificity (0.60) was lower in those with RA of 2 years or less. Many patients who reported improvement but failed to meet ACR20 criteria did so because they lacked 20% improvement in both tender and swollen joint counts.
An ACR20 response has a high specificity to detect changes in RA status rated as important by patients, suggesting that improvements of this degree capture changes that are meaningful to patients. EULAR responses were comparable to ACR responses.
To cite this abstract, please use the following information:
Alba, Maria I., Guthrie, Lori C., Ward, Michael M.; Sensitivity and Specificity of the American College of Rheumatology and Europe League Against Rheumatism Response Criteria in Rheumatoid Arthritis for Changes Important to Patients: Is An ACR20 Meaningful? [abstract]. Arthritis Rheum 2011;63 Suppl 10 :136