Abstract Objective Disease activity monitoring in SLE includes serial measurement of anti-double stranded-DNA (dsDNA) antibodies, but in patients who are persistently anti-dsDNA positive, the utility of repeated measurement is unclear. We investigated the usefulness of serial anti-dsDNA testing in predicting flare in SLE patients who are persistently anti-dsDNA positive. Methods Data were analysed from patients in a multinational longitudinal cohort with known anti-dsDNA results from 2013 to 2021. Patients were categorized based on their anti-dsDNA results as persistently negative, fluctuating or persistently positive. Cox regression models were used to examine longitudinal associations of anti-dsDNA results with flare. Results Data from 37 582 visits of 3484 patients were analysed. Of the patients 1029 (29.5%) had persistently positive anti-dsDNA and 1195 (34.3%) had fluctuating results. Anti-dsDNA expressed as a ratio to the normal cut-off was associated with the risk of subsequent flare, including in the persistently positive cohort (adjusted hazard ratio [HR] 1.56; 95% CI: 1.30, 1.87; P < 0.001) and fluctuating cohort (adjusted HR 1.46; 95% CI: 1.28, 1.66), both for a ratio >3. Both increases and decreases in anti-dsDNA more than 2-fold compared with the previous visit were associated with increased risk of flare in the fluctuating cohort (adjusted HR 1.33; 95% CI: 1.08, 1.65; P = 0.008) and the persistently positive cohort (adjusted HR 1.36; 95% CI: 1.08, 1.71; P = 0.009). Conclusion Absolute value and change in anti-dsDNA titres predict flares, including in persistently anti-dsDNA positive patients. This indicates that repeat monitoring of dsDNA has value in routine testing.