Two-tier serology is often used to confirm a diagnosis of Lyme disease. One hundred and four patients with physician diagnosed erythema migrans rashes had blood samples taken before and after 3 weeks of doxycycline treatment for early Lyme disease. Acute and convalescent serologies for Borrelia burgdorferi were interpreted according to the 2-tier antibody testing criteria proposed by the Centers for Disease Control and Prevention. Serostatus was compared across several clinical and demographic variables both pre- and post-treatment. Forty-one patients (39.4 %) were seronegative both before and after treatment. The majority of seropositive individuals on both acute and convalescent serology had a positive IgM western blot and a negative IgG western blot. IgG seroconversion on western blot was infrequent. Among the baseline variables included in the analysis, disseminated lesions (p < 0.0001), a longer duration of illness (p < 0.0001), and a higher number of reported symptoms (p = 0.004) were highly significantly associated with positive final serostatus, while male sex (p = 0.05) was borderline significant. This variability, and the lack of seroconversion in a subset of patients, highlights the limitations of using serology alone in identifying early Lyme disease. Furthermore, these findings underline the difficulty for rheumatologists in identifying a prior exposure to Lyme disease in caring for patients with medically unexplained symptoms or fibromyalgia-like syndromes.