Abstract: The sensitivity of the 1971 preliminary criteria for the classification of systemic lupus erythematosus (SLE) was 90% according to Cohen et al (1), who assembled those criteria. This high degree of sensitivity was not found by other investigators (2-6). The1982 revised criteria proposed by Tan et al (7) showed a 96% sensitivity among their patients. In a recent work, Levin et al (8) compared the 1971 and the 1982 American Rheumatism Association (ARA) criteria in 156 of their patients and found 88% sensitivity with the 1971 preliminary criteria and 91% with the 1982 revised criteria. The increased sensitivity was attributed to e addition of the fluorescent antinuclear antibody (FANA) test as a criterion, whereas lowering the value required for fulfilling the proteinuria criterion did not affect sensitivity. We compared the 2 criteria as did Levin et at (8). In 135 of our SLE patients, we found 81% sensitivity using the 1971 preliminary criteria and 90% with the 1982 revised criteria. In our group of patients, the improvement in sensitivity was 9%. This improvement was due to the decrease in the level of proteinuria to fulfill that criterion in 4% of the patients, to the addition of the anti-native DNA criterion in 2% of the patients, and to the inclusion of FANA in 9%. The elimination of alopecia and Raynaud's phenome-non as criteria, the selection of cellular casts as an alternative to proteinuria, and the use of the false-positive reaction for syphilis as an alternative to LE cells (not as a separate criterion) did not change the overall sensitivity. Although we observed a decrease in the total number of the criteria present in our patients, the total number was never fewer than the necessary 4 criteria. The gain in sensitivity with the 1982 ARA criteria was higher in our patients (9%) than it was the patients of Levin and coworkers (3%) (8). This gain is essentially due to the addition of FANA as a criterion.