OBJECTIVE Positive effects of breast milk feeding for preterm newborns have been demonstrated but bacterial contamination of mechanically expressed breast milk may cause neonatal sepsis. Pasteurizing breast milk reduces the bacterial count but impairs helpful nutrients. The aim of the study was (1) to analyze bacterial concentration in expressed breast milk and (2) to clarify if procedures after extraction cause an increase of bacterial contamination. METHODS Routine bacterial breast milk cultures of newborns < 32 weeks of gestational age were analyzed retrospectively from the year 2010. Serial milk cultures from both breasts of 50 mothers, the collecting bottle, and the in-house milk bank were tested prospectively. RESULTS Forty percent of milk samples from 2010 contained more than 100,000 colony-forming units (CFU) per milliliter or more than 10,000 CFU/mL potential pathogenic bacteria. In the prospective study no significant differences in bacterial concentration after breast milk extraction were observed. CONCLUSIONS High concentrations of bacteria are frequently observed in expressed breast milk. The study provides evidence that these bacteria are in the breast milk directly after emission. The pumping procedure and further transportation to the milk bank are of no influence.