Variants of SLE--a statistical approach for discrimination of a group of SLE cases into different subgroups sharing symptomatology. A pilot study.

Symptoms of two groups of Systemic lupus erythematosus (SLE) (Group 1 constituted 65 patients examined from 1975 to 1981; Group 2 constituted 104 patients examined from 1980 to 1988), which were diagnosed according to the American Rheumatism Association (ARA) preliminary criteria, were statistically analyzed with the use of a package of statistical programs which included computation of a matrix of correlation and cluster analysis. In both groups similar frequency and associations of SLE symptoms were seen. Hence, there was a positive correlation between kidney involvement and hematological abnormalities with nDNA Abs what was a hallmark of a severe SLE. In contrast, the symptoms of severe SLE disease was rarely seen in cases with Raynaud's phenomenon and discoid lesions. Our cluster analysis further distinguished groups on the basis of renal involvement, skin symptomatology, and polyserositis. These results also were similar in both groups of patients. That gave further credence to our results and continued to support the concept of SLE variants being distinguished on the basis of the clinical picture.