Systemic lupus erythematosus is characterized by antibodies to a variety of intracellular self-antigens, such as dsDNA and Sm, and these serve as hallmarks in the diagnosis of systemic autoimmune diseases. Several studies have shown that SmD1 and SmD3 synthetic peptides represent highly functional antigens for autoantibody detection and thus for diagnostic applications. The present study analysed the technical and clinical accuracy of an anti-SmD1 (amino acids 83-119) and an anti-SmD3 (amino acids 108-122) ELISA for the detection of anti-Sm antibodies. Depending on the cut-off value of the SmD1 ELISA, we found a high degree of concordance between the two tests. At an optimized cut-off value of 100 units for SmD1 we found the same clinical sensitivity (12.5%) and specificity (100%) in a group of systemic lupus erythematosus patients (n = 48) and in controls (n = 99). The concordance at this cut-off value was 100% (P < 0.0001; chi2 = 127.61). Using a second panel of sera (n = 65) preselected based on positive anti-Sm results, we confirmed the high degree of concordance between the two assays. Using dsDNA-coated ELISA plates and biotinylated peptides we confirmed the high dsDNA binding properties for SmD1, which were significantly higher than the SmD3-derived peptide. However, no cross-linking of anti-dsDNA antibodies to SmD1 was observed after adding increasing amounts of dsDNA to anti-dsDNA positive, anti-SmD1 negative serum. We therefore conclude that the reported difference in the sensitivity is related to the different cut-off levels and not to the detection of anti-dsDNA antibodies bridged via dsDNA to the SmD1 peptide. Moreover, we found that a subpopulation of anti-Sm antibodies cross-reacted with SmD1 and SmD3. Taken together, the data indicate that both SmD peptide ELISAs represent accurate assays and may be used as important standards for the detection of anti-Sm antibodies.