OBJECTIVE--To determine whether Borrelia burgdorferi is implicated in the pathogenesis of dilated cardiomyopathy in the United Kingdom. DESIGN--A controlled prospective study. Patients' notes were reviewed for evidence of Lyme disease and serum samples were tested by enzyme linked immunoadsorbent assay (ELISA) for antibodies to B burgdorferi. Samples with raised antibody concentrations were subsequently analysed by immunoblotting to determine their antibody binding specificity. SETTING--Tertiary referral centre. PATIENTS--97 consecutive patients with dilated cardiomyopathy diagnosed according to World Health Organisation criteria were studied. Serum samples were taken from two matched control groups. The first group (n = 38) was age, sex, and geographically matched. The second control group (n = 39) was environmentally matched and consisted of members of the patients' own households. MAIN OUTCOME MEASURES--Clinical evidence of Lyme disease. Presence of raised antibody concentrations to B burgdorferi. RESULTS--No patients had a previous illness compatible with Lyme disease. Analysis of the ELISA data showed eight of 97 patients with dilated cardiomyopathy (8.2%) and two of 77 controls (3.9%) had raised antibody concentrations. Immunoblot analysis, however, did not show binding patterns consistent with the presence of IgG specific for B burgdorferi in any of these samples. CONCLUSIONS--There was no clinical or serological evidence to implicate B burgdorferi in the pathogenesis of idiopathic dilated cardiomyopathy in the United Kingdom. In the absence of specific symptoms or likely exposure to B burgdorferi routine serological testing for Lyme disease in this group of patients is not recommended. Furthermore, raised antibodies to B burgdorferi are not diagnostic of active infection and ELISA results should be interpreted with caution unless specific B burgdorferi antibody bands have been found by immunoblot analysis.