Abstract
Currently, the only method for identifying infective hosts with Leishmania infantum to the vector Lutzomyia longipalpis is xenodiagnosis. More recently, quantitative polymerase chain reaction (qPCR) has been used to model human reservoir competence by assuming that detection of parasite DNA indicates the presence of viable parasites for infecting vectors. Since this assumption has not been proven, this study aimed to verify this hypothesis. The concentration of amastigotes in the peripheral blood of 30 patients with kala-azar was microscopically verified by leukoconcentration and was compared to qPCR estimates. Parasites were identified in 4.8 mL of peripheral blood from 67% of the patients, at a very low concentration (average 0.3 parasites/mL). However, qPCR showed 93% sensitivity and the estimated parasitaemia was over a thousand times greater, both in blood and plasma, with higher levels in plasma than in blood. Furthermore, the microscopic count of circulating parasites and the qPCR parasitaemia estimates were not mathematically compatible with the published proportions of infected sandflies in xenodiagnostic studies. These findings suggest that qPCR does not measure the concentration of circulating parasites, but rather measures DNA from other sites, and that blood might not be the main source of infection for vectors.