OBJECTIVE:To determine the degree of correlation between Farr and ELISA methods of detecting anti-dsDNA antibodies in patients with systemic lupus erythematosus (SLE), and their association with measures of disease activity. METHODS:Anti-dsDNA antibodies were assayed using the Farr and ELISA methods in patients followed between January 1, 2000, and December 31, 2002. Statistical correlations between Farr and ELISA were determined. Relationships between the 2 assays and measures of disease activity [SLE Disease Activity Index 2000 (SLEDAI-2K-DNA), renal, central nervous system (CNS), and vasculitis] were determined for the same clinic visit. RESULTS:550 patients with 2940 clinic visits met the inclusion criteria. Correlation between Farr and ELISA levels was 0.46 using the first visit for each patient. When the Farr was abnormal, the ELISA was equally likely to be normal or abnormal. Abnormal Farr results were associated with higher SLEDAI-2K scores than normal Farr results (6.2 vs 4.3, respectively; p < 0.0001). There was less of a distinction with ELISA results (5.9 vs 4.8; p = 0.04). Farr levels were significantly associated with the presence of renal disease and vasculitis, while ELISA levels were not. Neither Farr nor ELISA results correlated with the presence of active CNS involvement. CONCLUSION:Farr and ELISA techniques for the detection of anti-dsDNA antibodies in patients with SLE are poorly correlated. The Farr is superior to the ELISA in correlating with measures of global disease activity, as well as renal and vasculitis involvement. The Farr technique should continue to be used in clinical practice. The ELISA adds no additional information.