Abstract
A conventional indirect immunofluorescence assay (IFA) and an anticomplement indirect immunofluorescence assay (ACIF) for detecting serum antibodies to Borrelia burgdorferi in humans were evaluated during a prevalence survey in northern California. Sera obtained from 119 current or former residents of an area in which Lyme disease is endemic were split and tested by the IFA in two laboratories and the ACIF in a third. The seropositivity rate ranged from 15 to 20% with 88 to 93% agreement among laboratories. Interlaboratory agreement was statistically highly significant in each of the three pairwise comparisons and was positively associated with clinical manifestations of Lyme disease. Intralaboratory agreement ranged from 93 to 96% in two laboratories and was also statistically highly significant. Immunoblotting confirmed 100 of 101 of the nondiscrepant immunofluorescence test results and likewise was positively correlated with the degree of interlaboratory agreement. The ACIF was found to be a highly specific test (100% specificity) with a much lower cutoff titer (1:8) than the conventional IFA (determined to be 1:128 or 1:256 in two laboratories) for detecting antibodies to B. burgdorferi. It also appeared to be more sensitive (80 versus 68%) than the IFA as determined by comparative immunoblotting, though the absolute sensitivity of the ACIF for serodiagnosis of early Lyme disease has yet to be determined. Significant serologic cross-reactivity was demonstrated between B. burgdorferi, Borrelia coriaceae, and Borrelia hermsii by the IFA, which may confound spirochetal serosurveys in California where all three spirochetes are known to coexist.