Most donor human milk (HM) banks use Holder pasteurization (HoP) to ensure microbiological safety, although it can degrade essential bioactive factors for newborns. This study evaluates the innovative ultra-high-pressure homogenization (UHPH) technology as a potential alternative. Listeria innocua, Staphylococcus carnosus, Franconibacter helveticus (formerly named Cronobacter helveticus) and Escherichia coli strains were used as surrogates for common HM pathogens according to European Milk Bank Association (EMBA) guidelines, to evaluate the efficacy of new technologies. A reconstituted powder milk formula inoculated with these strains was used to determine the most efficient conditions (those to achieve a lethality of ≥5 Log), applying treatments from 150 to 300 MPa. These treatments were later validated using inoculated HM with the same strains. Immunoglobulin (sIgA, IgG, IgM) retention was also evaluated and compared with HoP. Results showed that UHPH treatments at 200 MPa achieved a lethality > 5 Log for all strains, except for St. carnosus, which required 250 MPa for complete inactivation in HM. Unlike HoP, UHPH at 200 and 250 MPa did not significantly reduce the basal concentration of sIgA, IgG, or IgM compared with raw HM. These findings suggest UHPH as a promising alternative to HoP, maintaining both microbiological safety and immunological quality.

Donor human milk (DHM) can harbor microbial contaminants that cause serious infections in premature infants. Bacillus cereus is a pathogen frequently found in DHM, capable of forming spores that can resist Holder pasteurization (62.5 °C, 30 min). Since no microbial growth is acceptable in post-pasteurized DHM, microbiological testing of pre-pasteurized DHM provides information about its contamination level to determine if it should be accepted for pasteurization. Culture is the gold standard in microbiological control but it requires 24–48 h to provide results. In this study we developed and validated a non-commercial real-time PCR assay for the detection of Bacillus cereus (BC test) in DHM specimens on a fully automated high-throughput platform, the cobas^® 6800 system. The BC test showed excellent sensitivity and specificity, repeatability and linearity over an 8-log range and a low limit of detection in milk specimens, as well as good agreement with selective culture methods. BC test was then used to systematically control all milk donations (3439) over a 24-month period. Bacillus cereus was detected in 14.2% of DHM, with monthly rates ranging from 6 to 29% and a significantly higher incidence in warmer months. Incorporating this assay into our laboratory workflow improved efficiency and reduced turnaround time.

Microbiological testing protocols are key components of human milk bank operations, yet the lack of consensus on screening schedules and microbiological criteria leads to variability in discard rates and uncertainty about safety standards. This study provides a large-scale characterization of pre-pasteurization donor human milk microbiota and evaluates the impact of testing protocols on operational outcomes. Over 59 months, 7,921 donor human milk batches were analysed. Pre-pasteurization analyses were dominated by skin staphylococci, with environmental Gram-negative bacilli and enteric genera detected at lower prevalence. Among pathogens of concern, only Staphylococcus aureus , Escherichia coli and Bacillus cereus sensu lato were identified. A selective 24-h workflow, incorporating BACARA® plates with a detection limit of 1 CFU/mL, was validated and implemented to target these organisms within the pre-pasteurization 24-h storage interval. This protocol significantly increased B. cereus s.l. detection (2.2% vs. 8.8%), without affecting S. aureus or Enterobacteriaceae detection, indicating improved analytical sensitivity. B. cereus s.l. exhibited a reproducible warm-season excess (5.5% vs. 2.4%) and accounted for 72% of post-pasteurization positives. In a sub-study of 491 batches, agreement between pre- and post-pasteurization B. cereus s.l. results was good ( κ ≈ 0.61) with a high negative predictive value (95%), supporting the predictive value of early screening. Protocol design was a major determinant of discard rates, and using higher microbiological thresholds (2 × 10 ⁵ CFU/mL) reduced avoidable losses without compromising safety. These findings hightlight the importance of integrating rapid, selective methods and advancing towards harmonized, risk-based microbiological criteria in human milk banking to improve both operational efficiency and microbiological safety.