Introduction Anti-double stranded DNA (dsDNA) antibodies are useful for the diagnosis of systemic lupus erythematosus (SLE). However, these antibodies are very heterogeneous and low avidity anti-dsDNA antibodies may occur in other rheumatic diseases as well. As a consequence of their heterogeneity, detection of anti-dsDNA antibodies partly depends on the type of assay used.

Aim of the study To compare five different anti-dsDNA antibody assays with the current one used in our routine diagnostics in order to set up an algorithm for better differentiation of high avidity anti-dsDNA antibodies from low avidity ones.

Methods From sera of 1446 consecutive patients sent to our laboratory for anti-dsDNA antibody measurements during a 4 months period, all sera that were found positive with our assay (ELISA1) were included in the study: 92 samples from SLE, 62 from other connective tissue disease (CTD) and 36 from patients with other inflammatory diseases (190 in total). All selected sera were tested for anti-dsDNA antibodies with two additional enzyme-linked immunoassays (ELISA2 and a high avidity ELISA3), with two Crithidia luciliae immunofluorescence tests (CLITF1 and CLITF2), and with Farr method used as the gold standard. Sera of 20 healthy blood donors were used as control.

Results In the selected group, out of the six methods the Farr, CLIFT1 and CLIFT2 had the highest specificity (93.87%, 90.81% and 81.63%) but lowest sensitivity (30.43%, 31.52% and 39.13%) for SLE. ELISA2 had a lower specificity (38.77%), but still high sensitivity (92.39%). The highest agreement by the Cohen’s Kappa was seen between CLIFT1 and CLIFT2 (0.558), Crithidia tests and Farr (0.531 and 0.489) and ELISA2 and ELISA3 (0.472). All Farr positive samples (n = 23) were also positive in ELISA’s.

The levels of anti-dsDNA antibodies detected with ELISA2 and ELISA3 correlated significantly (Spearman rank correlation coefficient Rho: 0.755; p<0.05). All other quantitative methods had a lower correlation coefficient.

Conclusions Based on these results, a two-step algorithm is suggested for the detection of anti-dsDNA antibodies: after screening samples with an ELISA test as the primary method (preferably ELISA2), CLIFT1 should be used to confirm positive results obtained in step one.