van Gestel A M, Anderson J J, van Riel P L, Boers M, Haagsma C J, Rich B, Wells G, Lange M L, Felson D T
Department of Rheumatology, University Hospital Nijmegen, The Netherlands.
J Rheumatol. 1999 Mar;26(3):705-11.
We compared the validity of the American College of Rheumatology (ACR) and the European League of Associations for Rheumatology (EULAR) definitions of response in rheumatoid arthritis (RA) clinical trials. US: ACR and EULAR improvement criteria were calculated in 7 large randomized RA clinical trials. The discriminant validity of the response criteria between treatment groups was studied using the Mantel-Haenszel chi-squared value. To compare both sets of criteria the chi-squared ratio was determined for each trial. Europe: In 2 large randomized RA clinical trials, ACR and EULAR criteria were calculated, once with extensive and once with 28 joint counts. The classification of patients with these 4 criteria were compared with each other using cross tables. We further studied the difference in response between treatment groups per trial, the association of response with patient and investigator assessment of improvement, and the association of response with radiological progression. US: The chi-squared ratio for most trials was close to 1. There was no clear pattern suggesting that the discriminant validity of the ACR criteria was stronger than the discriminant validity of the EULAR definition of response or vice versa. Europe: Conflicting results between ACR and EULAR were present in only 3% of patients in both trials. The discriminant validity of all 4 criteria (ACR and EULAR with reduced and extensive joint counts) was comparable. All criteria were related with the overall assessment of improvement by both investigator and patient. The association with radiographic progression was comparable for EULAR and ACR improvement criteria. There is a high level of agreement between ACR and EULAR improvement classification, and their validity is equivalent. The discriminating potential of the criteria between treatment groups is comparable, as is the association with patient's and investigator's overall assessment and with radiographic progression.
我们比较了美国风湿病学会（ACR）和欧洲抗风湿病联盟（EULAR）在类风湿关节炎（RA）临床试验中反应定义的有效性。美国：在7项大型随机RA临床试验中计算了ACR和EULAR改善标准。使用Mantel-Haenszel卡方值研究治疗组之间反应标准的判别效度。为比较两组标准，计算了每项试验的卡方比。欧洲：在2项大型随机RA临床试验中，计算了ACR和EULAR标准，一次采用广泛关节计数，一次采用28个关节计数。使用交叉表比较了这4种标准下患者的分类情况。我们进一步研究了每项试验中治疗组之间反应的差异、反应与患者及研究者对改善情况评估的关联，以及反应与放射学进展的关联。美国：大多数试验的卡方比接近1。没有明确模式表明ACR标准的判别效度强于EULAR反应定义的判别效度，反之亦然。欧洲：两项试验中仅有3%的患者出现ACR和EULAR之间的矛盾结果。所有4种标准（ACR和EULAR，关节计数减少和广泛）的判别效度相当。所有标准均与研究者和患者对改善情况的总体评估相关。EULAR和ACR改善标准与放射学进展的关联相当。ACR和EULAR改善分类之间高度一致，且它们的效度相当。治疗组之间标准的区分潜力相当，与患者和研究者的总体评估以及与放射学进展的关联也相当。
