The sensitivity and specificity of salivary rubella antibody detection was investigated using samples collected from 301 children after a mass vaccination campaign in the state of São Paulo, Brazil. Saliva samples were collected by 2 different methods: directly dribbling into a container or using a commercial collecting device. Corresponding finger-prick blood samples were collected on filter paper. Rubella specific immunoglobulin G (IgG) was measured in saliva by antibody capture radioimmunoassay and in blood samples by indirect enzyme-linked immunosorbent assay. The detection of salivary rubella specific IgG showed good correlation with the detection of rubella antibody in the blood samples. For both collecting techniques the predictive value for a positive saliva test was > 99% compared with the results from the blood tests. However, the predictive value for a negative saliva test was only 58.3% for a dribbled sample, compared to 100% for saliva collected using the commercial device. Moreover, collecting saliva by dribbling from children less than 4 years old was difficult. The detection of rubella specific IgG in saliva collected using a commercial device proved to be sensitive and specific in this epidemiological study, encouraging its more widespread application as a means of surveillance after mass vaccination.