OBJECTIVES: To determine if the past presence of anti-double-strand (ds)DNA antibody (Ab) will predict subsequent disease activity in patients with systemic lupus erythematosus (SLE). METHODS: A longitudinal study of clinical and serological disease manifestations registered during 2,412 patient months of follow-up in a well-defined lupus cohort. Organ-specific disease manifestations, the modified SLE disease activity index (M-SLEDAI) score, disease flares (M-SLEDAI increase > or =3) and predictive value of anti-dsDNA Ab testing [by enzyme-linked immunoabsorbent assay (ELISA) and Crithidia luciliae immunofluorescence (CLIFT) assays] were related to past anti-dsDNA Ab status. RESULTS: Anti-dsDNA Ab was previously demonstrated in 54 (57%) patients (group 1), while they were not earlier detected in 40 (43%) patients (group 2). The number of patients experiencing flares (46 vs 25%, p<0.01), the total number of flares (75 vs 17, p<0,001) as well as overall (60 vs 24 per 100 patient years, p<0,001) and organ-specific flare rate were higher in group 1. After adjustment for control frequency, group 1 remained at a higher risk for renal flares [odds ratio (OR) 2.4; confidence interval (CI) 1.5-4.1], and group 2 was at a higher risk for skin flares (OR 0.7; CI 0.5-0.8). While anti-dsDNA Ab testing overall was performed slightly more often in group 1 (OR 1.45; CI 1.0-4.6), anti-dsDNA Ab testing during flares was similar in both groups. CONCLUSION: The past presence of anti-dsDNA Ab identified patients with an increased risk of subsequent renal flares. However, as a new onset of anti-dsDNA Abs occurred late in the disease course, prior anti-dsDNA status was not adequate to predict disease flares.