Lyme disease: laboratory diagnosis of infected and vaccinated symptomatic dogs

Serological assays for detection of canine antibodies to the Lyme agent generally have been difficult to validate because an acceptable standard of comparison such as unequivocal proof of infection status has not been available. For practical and logistical reasons, it has not been possible to use culture of organism from infected animals, seroconversion in a large number of field dogs, or clinical criteria as the standard of comparison for validation of assays. Therefore, estimates of diagnostic sensitivity and specificity based on an appropriate gold standard have not been available. When it was discovered how to infect laboratory dogs via ticks infected with Borrelia burgdorferi, it was possible to define the kinetics and magnitude of the antibody response that might be expected in nature. ELISA and Western immunoblot data from experimental dogs were then compared and correlated with results of the same tests on dogs from endemic and nonendomic areas. Coupled with studies on cross-reactive antibodies elicited from other infectious agents or autoimmune phenomena, it was possible to account for interfering antibodies and to establish estimates of diagnostic sensitivity and specificity for the ELISA based on objective criteria. Such validated assays can predict, with a relatively high degree of proficiency, the infection and/or vaccinal status of animals. These assays have shown that some dogs, vaccinated with the commercially available whole-cell Lyme bacterins develop typical signs of Lyme disease but have no evidence of an underlying infection; antibody elicited only by the vaccine and not by infection is detectable in these animals. Western immunoblot can also confirm infection in animals of equivocal ELISA status if their bands have been evaluated for specificity of antibodies to B burgdorferi. Serology can be a very useful aid in the diagnosis of Lyme disease, but it requires that the assays used have been subjected to rigorous validation criteria. When that is not performed, an unacceptable level of false-positive and false-negative test results is virtually assured.