Recombinant Lyme disease vaccines based on purified preparations of outer surface protein A (OspA) have been shown to be effective in preventing transmission of Borrelia burgdorferi in experimental animal models and are now being tested in humans. Since the most widely used screening tests for Lyme disease are based on a whole-cell sonicate of B. burgdorferi, serologic false positivity in vaccinated persons could result from reactivity to OspA within the antigen preparation. In order to avoid serologic false positivity in vaccinated subjects, we developed an immunoassay based on a low-passage-number, naturally occurring variant of B. burgdorferi which lacks the plasmid encoding OspA and OspB. The use of an antigen preparation derived from this organism provided sensitive and specific detection of B. burgdorferi seropositivity in experimental animals and in human Lyme disease cases. The OspA-B-negative enzyme-linked immunosorbent assay (ELISA) also appeared to be capable of discriminating the vaccinated state from vaccine failure and natural infection in experimental animals. Sera from human subjects participating in a vaccine trial gave false-positive results with an ELISA based on an OspA-containing strain, but no such reactivity was observed when the OspA-negative ELISA was used. We conclude that low-passage-number OspA-B-negative isolates in immunoassays may become useful for the immunologic discrimination of the vaccinated state, natural infection, and vaccine failure.