Three cases of Bacillus cereus infection or colonization occurred in the same region in France, and milk from the milk bank was suspected as a possible common source of contamination. All Batches delivered to the three cases complied with the requirements of the bacteriological reference method recommended by good practices guidelines. Still, a retrospective analysis with a more sensitive method showed one batch to contain B. cereus, however straincomparison revealed no epidemiological link betweenisolates from patients and those from the milk. Consequently, in accordance with the precautionary principle, we developed a new sensitive method for the screening of pasteurized milk for pathogenic bacteria. From January 1 to August 31, 2017, 2526 samples of pasteurized milk were prospectively included in the study. We showed that a 20 mL sample of pasteurized milk incubated for 18 h at 37 °C under aerobic conditions was favoring the detection of B. Cereus. The nonconformity rate was 6.3% for the reference method and 12.6% for the improved method (p < 0.0001). Nonconformity was due to the presence of B. cereus in 88.5% of cases for the improved method and 53% of cases for the reference method (p < 0.0001). Thus our new method is improves the microbiological safety of the product distributed and only moderately increases the rate of bacteriological nonconformity .
Keywords: Bacillus cereus; Human milk; Milk bank; Pasteurization; Preterm; Reinforced method.