High avidity anti-dsDNA antibodies are more specific for SLE diagnosis, and more closely associated with renal involvement than intermediate or low-affinity anti-dsDNA antibodies. ELISA methods are largely used to detect anti-dsDNA, but their high sensitivity is inversely related to specificity because they also detect low avidity antibodies. We developed an ELISA assay based on the law of mass action and the competitive binding of dsDNA in solution and coated to microwells with anti-dsDNA antibodies. A simplified Scatchard plot analysis system was used to measure anti-dsDNA antibody avidity which was expressed as apparent affinity constant (Kaa), and quantified in liters per unit (I/U). We prospectively studied 101 consecutive SLE patients, who were followed for 3 years; three serum samples were sequentially collected from each patient during follow-up for determination of IgG anti-dsDNA antibody concentration, and anti-dsDNA avidity. SLE disease activity was estimated using the European Consensus Lupus Activity Measure (ECLAM) index. Sera from 100 healthy subjects and 133 patients with other connective tissue diseases or infectious diseases were also assayed as controls. The mean Kaa in SLE patients was 65.2 +/- 47.3 l/U, with no variations over time. Anti-dsDNA-positive SLE patients had higher Kaa values (79.1 +/- 46.8) than anti-dsDNA negative patients (27.2 +/- 20.1; P < 0.001). No correlation emerged between anti-dsDNA avidity and the ECLAM activity index score. Avidity was significantly higher in patients with renal involvement vs patients without this complication (78.2 +/- 50 vs 59.9 +/- 45.6 l/U; P = 0.0013). This simple ELISA method could be very useful in the diagnostic phase to differentiate high avidity anti-dsDNA autoantibodies that are characteristically found in SLE patients from low avidity antibodies that can also be found in other inflammatory diseases. Moreover, our data confirm the predictive value of high avidity anti-dsDNA antibodies for the development of lupus nephritis.