Measurement of anti-DNA antibodies: a reappraisal using five different methods.

One hundred and thirty coded sera, 60 from patients with systemic lupus erythematosus (SLE) and 70 from patients with other autoimmune rheumatic diseases were tested for deoxyribonucleic acid (DNA) binding activity by five different types of assay. These were enzyme linked immunosorbent assay (ELISA) (distinguishing IgG and IgM anti-ssDNA and anti-dsDNA), Crithidia luciliae, a nitrocellulose filter assay, the Amersham kit, and another modified Farr assay, the radioimmunoassay (RIA) (UK). The Crithidia test was the most specific, none of the controls was positive, but the least sensitive (13% positive only). The RIA (UK) was the most sensitive (57% positive). In most of the assays 3-9% of the controls were positive. When the SLE sera were analysed according to disease activity the IgG anti-dsDNA ELISA, all three RIA values, and the Crithidia test values were raised in all the patients with severely active disease. Some patients with inactive disease, however, were positive in each of the tests. The best interassay correlations (r less than 0.49) were found between RIA (UK), and ss IgG and the Amersham kit; and between ds IgG and ss IgG. In the main, however, it was clear that different assays are dependent upon distinctive properties of DNA antibodies. It seems inevitable that most major rheumatology units will require more than one anti-DNA antibody assay.