The sensitivity of the American Rheumatism Association's preliminary and revised criteria for the classification of systemic lupus erythematosus (SLE) was tested in 156 of our SLE patients. Eighty-eight percent met the 1971 preliminary criteria. Eighty-three percent fulfilled the 1982 revised criteria when arthritis was strictly defined (nonerosive arthritis) and 91 9' 0 when arthritis was more liberally defined (nondeforming arthritis). Analysis revealed that of the 3 serologic tests added in the revised criteria (antinuclear antibody, antiSm, anti-DNA), the antinuclear antibody test accounted for the increased sensitivity of the revised criteria.The American Rheumatism Association (ARA) revised criteria for the classification of systemic lupus erythematosus (SLE) have recently been published (1). The ARA preliminary criteria for the classification of systemic lupus erythematosus had been in use since publication in 1971 (2); the revision was considered necessary in order to add newer laboratory tests such as anti-DNA and antinuclear antibody (ANA) and to regroup laboratory tests and clinical findings into specific sets so as to increase the specificity and sensitivity of the preliminary criteria. Although many studies have been reported on testing of the preliminary criteria, no such studies have been reported on the revised criteria. We have therefore tested the revised criteria on our own group of SLE patients, in order to evaluate sensitivity and compare it with that of the preliminary criteria.The results of our study reveal that the revised criteria are minimally more sensitive than the preliminary criteria and that the addition of the antinuclear antibody manifestation is the major contributor to this increased sensitivity.
PATIENTS AND METHODSPatients. Outpatients or hospitalized patients followed by faculty members of the Division of Rheumatic Diseases during the period January 1, 1980 through December 31, 1982 at the University of Connecticut Health Center were considered for study. Patients were considered by their physician to have SLE, based on the presence of multisystern disease considered to be compatible with SLE. The diagnosis of SLE was made by the faculty member whether or not the patient fulfilled the preliminary criteria.Data entry and analysis. Data on the 156 patients were reviewed and information entered into a database booklet by the fellow and/or faculty member caring for the patient. Data entered by fellows were checked by the faculty member who had been responsible for the patient. The database booklet listed the 36 elements of the 2 sets of criteria exactly as defined in the literature (Tables 1 and 2).For each element, the reviewer entered one of the following: "yes," "no," "unclear," or "unknown." "Unclear" entries were reviewed by one of the authors (NFR) and were