BACKGROUND: Serological surveillance, based on the measurement of the presence of specific antibodies in a given population, can be used in addition to traditional and routine disease surveillance methods. The added value of this has been largely documented for vaccine-preventable diseases, but to a lesser extent for vector-borne diseases. This study aimed to evaluate the utility of seroprevalence data as additional source of information on the epidemiology of Lyme borreliosis in Belgium. METHODS: In total, 3215 residual blood samples collected in 2013-2015 were analysed with Liaison® Borrelia IgG kit (DiaSorin S.p.A, Saluggia, Italy). Positive and equivocal results were further examined with immunoblotting (recomLine Borrelia IgG kit, Mikrogen, Neuried, Germany). Crude prevalence estimates of equivocal and seropositive results were calculated and further adjusted accounting for clustered sampling and standardized for age, sex and population per province, according to the Belgian population structure in 2014. The effect of age, sex and region on seropositivity was assessed using log-binomial regression. RESULTS: The overall weighted national seroprevalence for Borrelia burgdorferi sensu lato, adjusted for clustered sampling, age, sex and province was 1.06% (95%CI 0.67-1.67). Although not statistically significant, the highest prevalences were observed in men and in those younger than 15 years or older than 59 years of age. At provincial level, the seroprevalence estimates do not follow the geographical distribution of tick bites and diagnoses of Lyme borreliosis as detected through other surveillance systems. CONCLUSIONS: Although the use of residual samples for seroprevalence estimates has several advantages, it seems to be a limited tool for serological surveillance of Lyme borreliosis in Belgium, other than follow-up of trends if repeated over time. A population-based sampling strategy might provide a more representative nationwide sample, but would be very time intensive and expensive. Seroprevalence ...