The objective of this study was to identify if bacteraemias occur during sialography, which could predispose to endocarditis in susceptible patients. Patients requiring sialography who gave informed consent were included in the trial. Patients taking antibiotics, or who were predisposed to infective endocarditis, were excluded. Sialography was performed using a standard hand injection technique with a water-soluble non-ionic contrast agent (Niopam, Bracco, UK). An indwelling cannula was inserted into the forearm prior to the procedure and then three blood samples of at least 10 ml were taken: sample one before sialography; sample two was taken as a continuous withdrawal during the sialographic procedure; and sample three over a further 10 min period after sialography. The blood samples were cultured for bacteria using the BacT/ALERT system (BioMerieux, UK). 32 patients were included in the study. No samples showed evidence of bacteraemia. Three patients had bacterial contamination from skin commensals. Using the rule of three, we have shown a 95% confidence interval of 3/32 = 0.09. In conclusion, the results of this study suggest that sialography is not associated with bacteraemia.