The performances of five commercially available enzyme immunoassays were compared for the detection of Borrelia burgdorferi IgM and IgG antibodies. Sensitivity was assessed with European serum samples collected from 45 patients with clinically defined Lyme disease in conjunction with a positive immunoblot (n = 44) or other serological test (n = 1). Sensitivities for the detection of IgM and IgG with each test were: Dako IgM 64%; Dako IgG 53%; Serion IgM 89%; and Serion IgG 88%. The Immunetics assay makes no distinction between IgM and IgG antibodies and had a sensitivity of 91%. Specificity was calculated by testing a control group comprising 40 patients with acute Epstein-Barr virus infection, cytomegalovirus infection, syphilis or rheumatoid factor positivity. The specificities achieved for each test were: Dako IgM 78%; Dako IgG 100%; Serion IgM 52%; Serion IgG 92%; and Immunetics 92%. The discriminatory power between control and patient samples appeared highest for the Immunetics assay. Between-run variation was comparable for the five tests and did not exceed 13%. When the Immunetics assay was used as an initial screening test, with low-titre positive results confirmed by an immunoblot, a sensitivity of 91% and a specificity of 100% were achieved. To attain maximal sensitivity, the Serion IgM and IgG tests were also performed on samples with negative Immunetics results. All positive Serion IgM and IgG results were also confirmed by immunoblot. In conclusion, the Immunetics assay, based on a synthetic C6 peptide, can be used reliably as an initial screening test for the serodiagnosis of Lyme disease.