Problems in the use of serologic tests for the diagnosis of Lyme disease

Lyme disease can be reliably diagnosed in the presence of erythema migrans. When erythema migrans is absent, serologic tests are often used to confirm the diagnosis. To choose a test for our Lyme disease diagnostic center, serum samples were obtained from 34 patients and tested for antibodies to Borrelia burgdorferi. We evaluated five enzyme-linked immunosorbent assays from Stony Brook (NY) University Hospital, Cambridge Bioscience (Worcester, Mass), Hillcrest Biologicals (Cypress, Calif), Sigma Diagnostics (St Louis, Mo), and Zeus-Wampole Scientific Inc (Raritan, NJ) and two fluorescent antibody tests (3M [Diagnostic Systems Inc, Santa Clara, Calif] and FIAX [Whittaker M.A. Bioproducts Inc, Walkersville, Md]). A positive sample by any test was further analyzed by Western blot. Using the Centers for Disease Control (Atlanta, Ga) epidemiologic case definitions, patients were classified into those with clinical Lyme disease, patients not meeting the Centers for Disease Control definitions, and asymptomatic patients. Sensitivities of Lyme serologies varied from 13% to 73%, with Hillcrest showing the highest value and Sigma the lowest value. False-positive test results were found in 0% to 27% of patients. Western blot analysis was positive in six of 15 patients with clinical Lyme disease. These results emphasize the need for better serologic testing for Lyme disease and underline their usefulness only as adjuncts in the clinical diagnosis of Lyme disease.