Abstract
In neonatal care, donor human milk (DHM) is used when maternal milk is unavailable or insufficient. In several countries, including Germany, raw (i.e., unpasteurised) DHM is occasionally administered under specific clinical conditions. However, the lack of standardised, evidence-based microbiological testing protocols raises concerns about the reliability of safety assessments for this high-risk patient group. The objective of this study was to assess the performance of four culture-based microbiological methods for detecting Enterobacterales in donor human milk, using both spiked samples and raw milk. We compared the detection limits of four culture-based microbiological methods, with and without enrichment, using spiked DHM samples and 93 raw DHM samples from a single donor (limited generalisation). Artificially inoculated samples contained defined concentrations of E. coli, K. pneumoniae, and S. ureilytica. Detection limits varied by several orders of magnitude (2.86 × 10(2) CFU/mL to 4.90 × 10(0) CFU/mL). In real samples, enrichment-based methods detected Gram-negative pathogens in four out of ninety-three samples (three S. ureilytica, one P. juntendi); direct plating detected none. Increasing the sample volume and applying enrichment improved detection sensitivity. Whole-genome sequencing confirmed species identity and showed that the S. ureilytica isolates from a single donor were clonally related, indicating a recurring detection pattern and underscoring the need for longitudinal microbiological monitoring. In view of the new EU SoHO Regulation classifying DHM as a Substance of Human Origin, these findings highlight the urgent need for standardised, sensitive protocols to ensure neonatal safety.