Microbiological evaluation of milk powdered formulae for infants sold on the Italian market

Background Powdered formulae are alternatives when breastfeeding is not possible or sufficient, but they can be contaminated by pathogenic or spoilage microorganisms. Spore-forming bacteria such as Bacillus spp. and clostridia, as well as pathogens like Cronobacter sakazakii and Salmonella spp., pose serious health risks to newborns and infants. However, microbiological studies on these products on the Italian market remain limited and often focus on only a few microbiological parameters. Methods Our study aimed to assess the microbiological quality and safety of powdered formula sold on the Italian market. A total of 83 samples were analysed, including 23 infant formulae (PIF), 42 follow-on formulae (FOF), and 18 formulae for special medical purposes (SMP). Enumeration was performed for total viable count (TVC – ISO 4833), anaerobic bacteria (Plate Count Agar in a CO2 atmosphere at 30°C for 48-72 hours), lactic acid bacteria (LAB – ISO 15214), Pseudomonas spp. (ISO 13720), Enterobacteriaceae (ISO 21528-2), Escherichia coli (ISO 16649-1), clostridia (ISO 15213-2), Enterococci (Slanetz-Bartley agar in aerobic conditions at 37°C for 48 hours), yeasts and moulds (ISO 21527-1), Bacillus cereus (ISO 7932), and coagulasepositive staphylococci (CPS – ISO 6888-1). After 24-hour enrichment in BPW at 37 °C, 10 μL of the homogenate were plated on PCA and incubated at 30 °C for 24 hours to verify any bacterial growth. The presence of Salmonella spp. (ISO 6579-1), C. sakazakii (ISO 22964), Listeria monocytogenes (ISO 11290-1), and Yersinia enterocolitica (ISO 10273) was investigated. Isolates were identified using MALDI-TOF MS and confirmed by RAPD-PCR. Results Most formulae showed low TVC, with 73 out of 83 samples (88.0%) below the detection limit (2 log CFU/g). Anaerobic bacterial counts were also low: 72 out of 83 samples (86.7%) were below the detection limit, 5 samples showed counts between 2–5 log CFU/g, and 6 samples had high counts (≥5 log CFU/g). LAB counts were below 2 log CFU/g in 73 samples. High LAB counts (≥5 log CFU/g) were detected in 3 PIF and 4 FOF samples from the same brand (Limosilactobacillus fermentum, identified by MALDI-TOF MS), as well as in 1 PIF and 2 FOF samples from different brands (Lactobacillus reuteri). Mould counts were below 2 log CFU/g in 74 samples (89.2%), while 8 samples had counts of 2 log CFU/g. One PIF sample (4.35%) showed a mould count exceeding 4 log CFU/g. All isolates were identified as Penicillium spp. Pseudomonas spp., Enterobacteriaceae, E. coli, Clostridium spp., Enterococcus spp., CPS, and yeasts were all below the detection limit (<2 log CFU/g). Presumptive B. cereus was detected in 38 samples after enrichment, but PCR confirmed its presence in 37 out of 83 samples (44.6%): among the 43 isolates collected, 42 were confirmed as B. cereus by PCR. The bacterium was enumerated in two PIF samples (from two batches of the same product, 0.9 log CFU/g) and in one SMP sample (1 log CFU/g). Conclusions The powdered formulae showed satisfactory microbiological quality, with no detection of major pathogens such as Salmonella spp. and C. sakazakii. However, B. cereus was confirmed in 44.6% of the samples, while it was enumerable in only 3 samples, suggesting generally low levels but potential risk. Future work will focus on the characterization of the 42 B. cereus isolates through virulence phenotyping, detection of toxin-encoding genes, and assessment of antibiotic resistance profiles.