A 2‐year longitudinal study of enzyme‐linked immunosorbent assay (ELISA) antibodies against Phlebotomus perniciosus and Phlebotomus papatasi (Diptera: Psychodidae) sandfly saliva was performed in 32 Beagle dogs treated preventively with an imidacloprid–permethrin topical insecticide in an endemic area in Spain. Dogs were grouped into three sandfly exposure groups according to the time of inclusion in the study. Assays analysed immunoglobulin G (IgG) against salivary gland homogenates (SGH) of both species and recombinant P. papatasi rSP32 and P. perniciosus rSP03B proteins in serum. The dogs were participating in a Leishmania infantum (Kinetoplastida: Trypanosomatidae) vaccine trial and were experimentally infected with the parasite in the second year. No dog acquired natural L. infantum infections during the first year, but most developed anti‐saliva antibodies, and median log‐transformed optical densities (LODs) were seasonal, mimicking those of local sandflies. This indicates that the repellent efficacy of the insecticide used is below 100%. Multi‐level modelling of LODs revealed variability among dogs, autocorrelation and differences according to the salivary antigen and the dog's age. However, dog seroprevalence, estimated using pre‐exposure LODs as cut‐offs, was relatively low. This, and the fact that dogs did not become naturally infected with L. infantum, would support the efficacy and usefulness of this imidacloprid–permethrin topical insecticide in canine leishmaniasis control.