To evaluate the distribution of asymptomatic infection by Leishmania infantum in a metropolis in Brazil with different relative risks (RRs) for disease and risk factors associated with the infection, an ecological study was conducted using a Bayesian approach to estimate the RR of human visceral leishmaniasis (HVL) based on cases between 2008 and 2011. The areas were categorized and selected according to disease incidence: low (area-1), medium (area-2) and high (area-3). Cross-sectional study enrolling 935 children was used to estimate the prevalence of infection by L. infantum. Volunteers from these three areas were tested for L. infantum infection by ELISA (rK39 and soluble antigens). Infection prevalence rates were estimated and compared with the RR of disease. Multilevel logistic regression model evaluated the relationship between infection and the analysed variables. The RR of HVL was distributed heterogeneously in the municipality. The infection prevalence rates were: 34·9% in area-1; 29·3% in area-2; and 33·6% in area-3, with no significant differences between these areas. The variables ‘Presence of backyards in the neighbourhood’ and ‘Younger children’ were associated with L. infantum infection. We conclude that infection by L. infantum affects a significant proportion of the infant population regardless of the RR of disease.