OCCURRENCE OF ANTIBODIES TO BORRELIA-BURGDORFERI IN PATIENTS WITH NONSPIROCHETAL SUBACUTE BACTERIAL-ENDOCARDITIS

Objective: To determine the prevalence and specificity of antibodies t o Borrelia burgdorferi in patients with nonspirochetal subacute bacter ial endocarditis and assess whether increased levels of antibodies to B. burgdorferi were attributable to rheumatoid factor. Design: Retrosp ective case-control study. Setting: Urban referral center in an area d evoid of infected ticks as a source of endocarditis sera. Patients: Se ra from 30 consecutive patients with culture-proven subacute endocardi tis between 1979 and 1981 were compared with 30 control sera collected between 1989 and 1990. In addition, sera from 20 consecutive patients with rheumatoid arthritis who were positive for rheumatoid factor wer e collected between 1991 and 1992. Sera were compared with a convenien ce sample from 15 patients who met the criteria for Lyme disease. Meas urements: Antibodies to B. burgdorferi were assessed by enzyme-linked immunosorbent assay (ELISA) and immunoblot analysis, IgM rheumatoid fa ctor was quantified using solid-phase radioimmunoassay or latex agglut ination techniques. Results. Thirteen of 30 patients with endocarditis (43%) compared with 3 of 30 normal controls (10%) had increased level s of antibodies to B. burgdorferi (P < 0.01). Of these 13 patients, on ly 1 had an immunoblot consistent with previous infection. The other's had nonspecific immunoblots: 5 showed isolated 60-kd reactivity; 1 pa tient had isolated 41-kd reactivity; and 6 had no bands of reactivity. Immunoblots of the 3 controls with increased antibodies showed only i solated 41-kd reactivity. Thus, the specificity of the B. burgdorferi antibody test in patients with endocarditis was only 60% (95% Cl, 42% to 78%), compared with 90% (CI, 79% to 100%) in controls. No correlati on was noted between IgM rheumatoid factor and antibodies to B. burgdo rferi in patients with endocarditis (r = 0.2; P > 0.2). Only 1 of 20 p atients with rheumatoid arthritis without known bacterial infections h ad antibodies to B. burgdorferi. Conclusions: Although a positive ELIS A test for B. burgdorferi may be a ''true positive,'' a positive serol ogic test alone does not ensure that the clinical problem is due to Ly me borreliosis. Cross-reactive antibodies to shared epitopes between B . burgdorferi and the endocarditis organism may account for the high f alse-positive results.