Share this article on:

Heating-induced Bacteriological and Biochemical Modifications in Human Donor Milk After Holder Pasteurisation

Segura, Aránzazu Gómez de*; Escuder, Diana; Montilla, Antonia; Bustos, Gerardo; Pallás, Carmen; Fernández, Leonides*; Corzo, Nieves; Rodríguez, Juan M.*

Journal of Pediatric Gastroenterology & Nutrition: February 2012 - Volume 54 - Issue 2 - p 197–203
doi: 10.1097/MPG.0b013e318235d50d
Original Articles: Hepatology and Nutrition

Objectives: The objectives of the present study were to enumerate and characterize the pathogenic potential of the Bacillus population that may survive holder pasteurisation of human milk and to evaluate the nutritional damage of this treatment using the furosine and lactulose indexes.

Materials and Methods: Milk samples from 21 donors were heated at 62.5°C for 30 minutes. Bacterial counts, lactose, glucose, myoinositol, lactulose, and furosine were determined before and after the heat treatment. Some B cereus isolates that survived after pasteurisation were evaluated for toxigenic potential.

Results: Nonpasteurised milk samples showed bacterial growth in most of the agar media tested. Bacterial survival after pasteurisation was observed in only 3 samples and, in these cases, the microorganisms isolated belonged to the species B cereus. Furosine could not be detected in any of the samples, whereas changes in lactose, glucose, and myoinositol concentrations after holder pasteurisation were not relevant. Lactulose was below the detection limit of the analytical method in nonpasteurised samples, whereas it was found at low levels in 62% of the samples after holder pasteurisation. The lactation period influenced myoinositol content because its concentration was significantly higher in transition milk than in mature or late lactation milk samples.

Conclusions: Holder pasteurisation led to the destruction of bacteria present initially in donor milk samples, except for some B cereus that did not display a high virulence potential and did not modify significantly the concentration of the compounds analyzed in the present study.

*Departamento de Nutrición, Bromatología y Tecnología de los Alimentos, Universidad Complutense de Madrid

Servicio de Neonatología, Hospital Universitario 12 de Octubre

Departamento de Bioactividad y Análisis de Alimentos, Instituto de Investigación en Ciencias de la Alimentación, Madrid, Spain.

Address correspondence and reprint requests to Dr Juan M. Rodríguez, Departamento de Nutrición, Bromatología y Tecnología de los Alimentos. Facultad de Veterinaria, Universidad Complutense de Madrid, Avenida Puerta de Hierro, s/n, 28040 Madrid, Spain (e-mail:

Received 12 April, 2011

Accepted 8 July, 2011

The present study was supported by the 110AC0386 (CYTED), CSD2007-00063 (FUN-C-FOOD, Consolider-Ingenio 2010), and AGL2010-15420 projects from the Ministerio de Ciencia e Innovación (Spain), and by projects FIS PS09/00040 (Ministerio de Sanidad y Consumo, Spain) and S2009/AGR-1469 (Comunidad de Madrid, Spain).

The authors report no conflicts of interest.

See “Heat Treatment of Human Milk” by Moro and Arslanoglu on page 165.

Human milk is widely recognized as the optimal feeding option for human term and preterm infants because of the wide spectrum of short-, medium-, and long-term potential benefits that it provides (1). Unfortunately, there are cases in which a mother's own milk is not available or enough to cover the requirements of the newborn. Therefore, there is a worldwide increasing demand for donor breast milk, particularly for preterm infants and older infants experiencing diverse medical problems (2). In such situations, clinicians value the importance of banked human milk, not only as a nutritional option but also as a potentially lifesaving therapy.

To date, there are no worldwide uniform guidelines for the screening, processing, storage, and handling of donor milk among milk banks and, in fact, protocols may vary even in banks operating in the same country. The potential mother-to-child transmission of certain viruses, such as human immunodeficiency virus, human T-lymphoma virus, or cytomegalovirus, through breast-feeding, together with the difficulties in an exhaustive surveillance of donors’ health (including repetitive serum screening), has led to the systematic pasteurisation of donor milk in the vast majority of human milk banks. Human pasteurised milk is considered the best alternative to nonheated frozen or fresh milk, and has been shown to reduce the incidence of necrotizing enterocolitis, sepsis, and other infections in premature and high-risk infants, resulting in shorter hospital stays (3–5).

Although some nutrients and bioactive compounds present in fresh human milk remain active after such heat treatment, the biological activity of other compounds is affected to a variable degree (6–10). As a consequence, questions arise concerning the effects of heat processing on some of the unique components of human milk.

Because of the content of important nutrients in lactose and proteins, heating of human milk can induce chemical changes in them, leading to adverse nutritional effects (11). The extent of damage produced by heating can be measured through the use of chemical indexes, such as the furosine (2-furoylmethyl-lysine) and lactulose (4-O-β-D-galactopyranosyl-D-fructofuranose) levels. Furosine is used as an indirect measurement of Amadori compounds formed in the early stages of Maillard reaction between proteins (ε-amino group of protein-bound lysine) and sugar components (carbonyl group of reducing sugar as lactose) during processing (12). Lactulose is a synthetic sugar, which does not occur naturally, and is produced from lactose by isomerization in basic media. This disaccharide is absent in raw milk, but the dissolved salt system of milk is a buffered solvent favorable to the formation of lactulose from lactose during the heat treatment of milk (13). Both furosine and lactulose are useful markers for evaluating the extent of heat damage in milk and infant formulas (14,15).

In addition, breast milk is a source of commensal and potentially probiotic bacteria (16,17), which seem to play an important role in gut colonization of the healthy infant (17,18). Such bacteria are killed by the pasteurisation process. It is important to note that spore-forming bacteria that may survive the heating process, such as Bacillus cereus, or microorganisms that could contaminate milk after pasteurisation can grow faster than in raw milk because of the heat damage to the milk bacteriostatic systems, including the absence of natural competitors (6).

In this context, the objectives of the present study were, on the one hand, to enumerate and characterize the pathogenic potential of the Bacillus population that may survive holder pasteurisation, and, on the other hand, to evaluate the potential nutritional damage of this thermal treatment using the furosine and lactulose indexes.

Back to Top | Article Outline


Breast Milk Samples

Breast milk samples (8 mL) were obtained from the human milk bank located at the Hospital Universitario 12 de Octubre (Madrid, Spain). Milk collection was performed following a specific protocol for donor mothers approved by the local ethics committee. The samples were obtained from 21 donors who fulfilled the requirements of the bank, and informed consent was obtained from each donor. Milk was collected at home using electric (Lactaline; Ameda, Lincolnshire, IL) or manual (Harmony or Lactaset models; Medela, Baar, Switzerland) pumps. An aliquot from each milk sample was separated before pasteurisation, whereas the rest of the sample was pasteurised by heating it at 62.5°C for 30 minutes; then, it was cooled in a shaking water bath (Lab Companion, Seoul, Korea) filled with ice-cold water and provided temperature control. Once the temperature reached 4°C (always within the first 15 minutes of cooling), it was stored at −20°C until analysis. A thermometer, coupled to an external sensor of temperature (DT 132; Fourier, Fairfield, CT), was introduced in a control bottle (cow's milk) and used as a probe to monitor the temperature of the milk batch during the whole heating/cooling process.

Back to Top | Article Outline

Bacterial Cultures and Identification of Isolates

Proper peptone water dilutions of 21 pairs of milk samples (21 before and 21 after pasteurisation) were plated onto brain heart infusion (BHI, Oxoid, Basingstoke, UK; a general-purpose medium suitable for the cultivation of nonfastidious bacteria, yeasts, and moulds), Columbia nadilixic acid agar (CNA, BioMerieux, Marcy l’Etoile, France; a highly nutritious, general-purpose medium for the isolation and cultivation of fastidious microorganisms), Baird Parker (BP, BioMerieux; a selective medium for the isolation of staphylococci), MacConkey (MCK, BioMerieux; a selective medium for the isolation of enterobacteria), polymyxin-pyruvate-egg yolk-mannitol with bromothymol blue (PEMBA, Oxoid, a selective medium for the isolation of Bacillus), and de Man, Rogosa, and Sharpe (MRS, Oxoid, a medium for the isolation of lactic acid bacteria) agar plates, which were aerobically incubated at 37°C for 24 to 48 hours. Parallel, the samples were also cultured on Wilkins Chalgren (WCh, Oxoid, a general medium for isolating anaerobic bacteria), which were incubated anaerobically (85% nitrogen, 10% hydrogen, 5% carbon dioxide) in an anaerobic workstation (MINI-MACS, DW Scientific, Shipley, UK) at 37°C for 48 hours. Colonies, from the plates in which bacterial growth was detected, were isolated and stored at −20°C in the presence of glycerol (20%, vol/vol).

Identification of the isolates was performed by polymerase chain reaction (PCR) sequencing of a 470-bp fragment of the 16S rRNA gene as described elsewhere (19). The amplicons were purified using the Nucleospin Extract II kit (Macherey-Nagel, Düren, Germany) and sequenced at the genomics unit of the Complutense University (Madrid, Spain). The resulting sequences were used to search sequences deposited in the EMBL database using BLAST algorithm, and the identity of the isolates was determined on the basis of the highest scores (>98%).

Because the genomes of the B cereus group of species, including B cereus and B anthracis, are closely related in both gene content and synteny (20) and their 16S rRNA gene sequences share >99% similarity (21), those isolates identified as B cereus were submitted to a repetitive element polymorphism PCR assay to ensure that they did not belong to the B anthracis species. For this purpose, the BOX-A1R primer, which originates a 390-bp fragment if B anthracis DNA is present in the sample, was used as described previously (22).

Back to Top | Article Outline

Genetic Profiling of the B cereus Isolates and Evaluation of Their Toxigenic Potential

A collection of 49 B cereus isolates, including those obtained in the present study and those obtained from other samples of pasteurised donor milk that were previously rejected by the same milk bank because of the presence of this bacterial species, were typed by random amplification polymorphic DNA (RAPD), using primer OPL5 (5′-ACG CAG GCA C-3′) as described elsewhere (23).

The presence of genes involved in the biosynthesis of the main B cereus toxins was evaluated. In relation to the toxins responsible for food poisoning by B cereus, cereulide is associated with the emetic symptoms and is encoded by the cereulide synthetase (ces) gene cluster, whereas 3 pore-forming toxins appear to be responsible for the diarrhoeal symptoms: hemolysin BL (Hbl), nonhaemolytic enterotoxin (Nhe), and cytotoxin K (CytK-1 or CytK-2) (24). Hbl consists of the 3 proteins L2, L1, and B, encoded by the genes hblC, hblD, and hblA, respectively; Nhe is composed of the proteins NheA, NET, and NheC, encoded by the nheABC operon; finally, CytK-1 or -2 are single-component toxins. To detect the presence of toxin gene determinants, total genomic DNA from each B cereus strain was extracted by disrupting colonies in deionized water and in a (chloroform:isoamylic alcohol):water (1:1, vol:vol) solution. Then, 3 multiplex PCR assays were used for the detection of the hblCDA and nheABC operons and the ces gene cluster using primer sets and PCR conditions described elsewhere (25). At the same time, a duplex PCR assay was performed to detect genes encoding CytK-1 and CytK-2 (26). The presence of toxins Hbl and Nhe in culture supernatants of same B cereus isolates was also analyzed with the gold-labeled immunosorbent assay-rapid test for the qualitative detection of B cereus enterotoxins (Merck, Darmstadt, Germany) following the instructions of the manufacturer.

Back to Top | Article Outline

Determination of Furosine

Determination of furosine in the 21 pairs of milk samples was performed by ion-pair reverse phase-high performance liquid chromatography (RP-HPLC) following the method of Resmini et al (27). Before analysis, milk samples (2 mL) were hydrolysed with 6 mL of 10.6 N HCl under inert conditions at 110°C for 24 hours in a Pyrex screw-cap vial with PTFE-faced septa. The hydrolysate was filtered through Whatman No. 40 filter paper, and 0.5 mL of filtrate was applied to a previously activated Sep-Pak C18 cartridge (Millipore, Milford, MA). Furosine was eluted with 3 mL of 3 N HCl and 20 μL was injected into the chromatograph.

RP-HPLC analysis of furosine was carried out in a C8 column (250 mm × 4.6 mm, 5 μm) (Alltech furosine-dedicated; Alltech Associates, Laarne, Belgium) maintained at 35°C using a linear binary gradient at a flow rate of 1.2 mL/min. Mobile phase was made up of solvent A, 0.4% acetic acid, and solvent B, 0.3% KCl in phase A. Detection was performed using a variable-wavelength UV detector at 280 nm (LDC Analytical, SM 4000). Acquisition and processing of data were achieved with a HPChem Station (Hewlett-Packard Co, Wilmington, DE). Calibration was performed by external standard method using commercial standard of pure furosine (Neosystem Laboratories, Strasbourg, France). The detection limit of the RP-HPLC method was 1.16 mg/100 g of protein. The determination of protein concentration was done following the Bradford procedure (BioRad, Hercules, CA) using albumin as external standard.

Back to Top | Article Outline

Gas Chromatography Analysis of Carbohydrates

Lactose, glucose, lactulose, and myoinositol were also determined by gas chromatography in the 21 pairs of milk samples, following the method of Montilla et al (28). For this purpose, 0.2 mL of sample was made up to 2 mL with methanol in a volumetric flask to remove proteins and fat. Mixtures were vigorously stirred, followed by standing for at least 1 hour. The supernatant was used for carbohydrate analysis, and a solution of 0.1% (wt/vol) phenyl-β-D-glucoside in methanol/water (70:30, vol/vol) was added as internal standard.

Before derivatization, equal volumes (0.5 mL) of supernatant and internal standard solution were mixed and dried at 38°C to 40°C in a rotary evaporator. The dried mixtures were treated with 100 μL N,N-dimethylformamide and held at 70°C for 1 hour to obtain a constant anomeric composition. Then, 100 μL of N-trimethylsilylimidazole were added to silylated carbohydrates, and the reaction was completed in 30 minutes at 70°C. Silylated carbohydrates were extracted with 0.1 mL of hexane and 0.2 mL of water. Volumes in the range of 0.2 to 1 μL of the organic phase containing silyl derivatives were injected into the column.

The trimethylsilyl ethers of carbohydrates were analysed in an Agilent Technologies 7890A gas chromatograph equipped with a commercial 30 m × 0.32 mm inside diameter, 0.5 μm film-fused silica capillary column SPBTM-17, bonded, cross-linked phase (50% diphenyl/50% dimethylsiloxane) (Supelco, Bellefonte, PA). Separation was performed at 235°C for 9 minutes, followed by an increase up to 270°C at a rate of 15°C/min and maintenance of this temperature for 15 minutes. Temperatures of the injector and flame ionization detector were 300°C during the analysis. Injections were carried out in split mode 1:30, using 1 mL/min of nitrogen as carrier gas. Data acquisition and integration were performed using Agilent Chem-Station Rev. B.03.01 software (Wilmington, DE). To study the response factor relative to the internal standard, solutions containing lactose, lactulose, glucose, and myoinositol were prepared over the expected concentration range in milk samples. The identity of carbohydrates present in milk samples was confirmed by comparison with relative retention times of standard samples.

Back to Top | Article Outline

Statistical Analysis

Microbiological data, recorded as colony-forming units per milliliter of milk, were transformed to logarithmic values before statistical analysis. Quantitative biochemical data were expressed as mean ± standard deviation (SD) and 95% confidence interval (CI) of the mean. Values were tested for normality of distribution. Correlations between lactose, glucose, and myoinositol concentration and lactation time were determined by the Spearman method. The effect of holder pasteurisation on sugar concentrations was evaluated with paired Student t tests. The influence of the lactation period (transition milk, mature milk, and late lactation milk) in lactose, glucose, and myoinositol concentration was analyzed by 1-way ANOVA followed by Bonferroni multiple comparison tests. Statistical tests were considered significant at P < 0.05. The SAS system (SAS Institute, Cary, NC) was used to perform these analyses.

Back to Top | Article Outline


Bacterial Counts in the Milk Samples and Identification of the Isolates

To evaluate the effect of holder pasteurisation on the viability of the milk bacteria, 21 samples of donor milk were cultured before and after heating at 62.5°C for 30 minutes. In all of the cases, inoculation of nonpasteurised milk samples in BHI, CNA, BP, MRS, WCh, and PEMBA agar plates led to bacterial growth (Table 1). In contrast, bacteria could be isolated from only 13 (62%) of the same samples when inoculated on MCK plates (Table 1). Globally, the bacterial counts in nonpasteurised samples oscillated between 2.60 and 5.22 log10 CFU/mL in BHI medium, with a mean (SD; 95% CI) value of 3.93 (0.85, 3.54–4.31) log10 CFU/mL. The lowest mean bacterial counts of donor milk samples were found in MRS medium, and were 0.88 log10 CFU/mL lower than in BHI agar plates (Table 1). In the samples in which growth was observed on MCK agar plates (n = 13), the counts oscillated between 1.70 and 4.92 log10 CFU/mL.

Most of the bacteria isolated from the raw milk samples, both qualitatively and quantitatively, belonged to the genera Staphylococcus (BHI, CNA, MRS, WCh agar plates), Streptococcus (CNA, MRS, WCh agar plates), and Bacillus (PEMBA), or to the Lactobacillus group (MRS). When growth was observed on MCK plates, all of the isolates belonged to the coliform group and, most of them, to the species Escherichia coli or to the genus Enterobacter.

Pasteurisation had a radical effect on the bacterial population of the samples (Table 1); in fact, bacterial growth could not be detected from most pasteurised samples after culturing onto BHI, MCK, CNA, BP, MRS, or WCh agar plates. Bacterial survival after pasteurisation was observed in 1 sample when cultured on BHI (1.7 log10 CFU/mL) and PEMBA (3.44 log10 CFU/mL) agar plates, and in 2 additional samples when cultured on PEMBA agar plates (both at a concentration of ∼2.0 log10 CFU/mL) (Table 1). In all of these cases, the microorganisms isolated belonged to the species B cereus. A low number of colonies (n = 1–3; dilution 0) of staphylococci or propionibacteria were observed in 2 samples, but their presence seemed to be the result of postprocessing contamination because these isolates did not survive holder pasteurisation when they were inoculated in sterile milk at an initial concentration of ∼4.0 log10 CFU/mL (results not shown).

Back to Top | Article Outline

Genetic Profiling of the B cereus Isolates and Evaluation of Their Toxigenic Potential

RAPD profiling of the B cereus isolates showed the presence of 6 different band patterns. Interestingly, each RAPD profile was coincident with 1 of the 6 toxin gene profiles observed among the 49 isolates (Table 2).

The isolates harbored neither the ces gene, associated with the biosynthesis of the emetic toxin, nor cytK1, whereas all carried those required for Nhe production. The cytK2 gene was present in approximately 50% of the isolates. Finally, only 1 strain (∼2%) harbored the complete hblCDA operon. The results obtained with the gold-labeled immunosorbent assays (detection of toxins in culture) were in agreement with the presence of the genes as determined by multiplex PCR. The Nhe toxin could be detected in cultures of the nheABC-positive strains, whereas Hbl toxin could be detected only in the strain that harbored the complete hblCDA operon (Table 2).

Back to Top | Article Outline

Effect of Pasteurisation and Lactation Period on the Concentrations of Furosine and Carbohydrates

In the present study, no peak of furosine was detected in any of the samples, neither before nor after the pasteurisation process; therefore, holder pasteurisation did not favor Maillard reaction.

Lactose, glucose, myoinositol, and lactulose concentrations in donor milk samples are presented in Table 3. Mean (SD; 95% CI) concentration of lactose in nonpasteurised milk samples was 64.08 g/L (6.14; 61.28–66.88). Glucose and myoinositol were found in all of the samples at approximately 3 orders of magnitude lower concentration than lactose and showed considerable sample-to-sample variation; that is, 206.45 mg/L (78.79; 170.59–242.32) for glucose and 196.45 mg/L (104.40; 148.93–243.97) for myoinositol. Glucose and myoinositol concentrations were not related in each individual milk sample; in fact, the ratio (glucose/myoinositol) varied from 0.33 to 3.05, and only 5 samples showed similar amounts of both compounds (glucose/myoinositol = 0.91–1.17). There was no correlation between lactose, glucose, and myoinositol concentrations in milk samples (data not shown).

Changes in lactose, glucose, and myoinositol concentrations in milk samples after holder pasteurisation were not relevant, although mean lactose concentration (expressed as mean ± SD) increased by 1.42 ± 2.89 g/L (paired Student t test, P = 0.036) and mean glucose and myoinositol concentrations decreased by 6.82 ± 14.75 and 1.61 ± 19.95 mg/L, respectively (paired Student t test, P = 0.047 and P = 0.716, respectively). Lactulose was below the detection limit of the analytical method (10 mg/L) in nonpasteurised milk samples, and it was found in 62% of the samples after holder pasteurisation with a mean concentration of 18.96 ± 6.14 mg/L (Table 3).

Because milk samples were donated by women at different lactation periods, between 6 days and 1.9 years, Spearman correlation coefficients were calculated to compare the relation between the concentration of lactose, glucose, and myoinositol in human milk samples and the lactation time (in days). Lactose and glucose concentration had poor correlation with the length of lactation (Spearman r = −0.302, P = 0.1836 for lactose and r = 0.155, P = 0.503 for glucose), but there was a strong negative and significant correlation between myoinositol concentration and the lactation time (Spearman r = −0.752, P ≤ 0.0001).

Lactose, glucose, and myoinositol concentration in human milk samples as a function of 3 different lactation periods is shown in Figure 1. Lactose and glucose concentrations were slightly higher in mature milk (15–180 days) samples than in transition milk (<15 days) and in late lactation milk (180–250 days of lactation), but these differences were not statistically significant. In contrast, the lactation period had a significant effect on myoinositol concentration (1-way ANOVA, F = 11.65, P = 0.0006). The concentration of myoinositol in transition milk (341.35 ± 95.22 mg/L, n = 4) was significantly higher than that in mature (198.04 ± 80.14 mg/L, n = 8, P < 0.05) and late lactation (130.64 ± 53.44 mg/L, n = 9, P < 0.001) milk samples (Fig. 1).

Back to Top | Article Outline


In the present study, bacteria could be isolated from nonpasteurised human milk in different culture media. This finding is not strange because fresh human milk contains a number (<3 log10 CFU/mL) of viable bacteria and a wide range of free bacterial DNA signatures, which may program the neonatal immune system (29). In fact, breast milk has been shown to be a continuous source of commensal and potentially probiotic bacteria to the infant gut, including staphylococci, streptococci, bifidobacteria, and lactic acid bacteria (16,17,30).

The fact that, in the present study, donors extracted the milk using pumps may explain why many samples had counts >3 log10 CFU/mL and why growth was observed in 62% of samples when cultured on MCK agar plates. It has been shown that the use of milk pumps to collect the samples is associated with a higher level of bacteria, and particularly enterobacteria, which are not related to the usual breast milk microbiota (31). Contamination of milk during pumping has been reported previously and seems to be of particular concern for premature infants or ill infants in neonatal intensive care units (32,33). Many milk pumps and/or their accessories cannot be properly sanitised and/or sterilised, and bacteria usually persist after the application of present cleaning protocols. Therefore, the design of new pumping devices that can be sterilised and subjected to more efficient cleaning and disinfection procedures is greatly desirable. In addition, donors should receive more education on hygienic milk extraction and storage from the milk bank staff.

Holder pasteurisation of the milk samples led to the destruction of the bacteria present in the initial fresh samples with the exception of 3 samples in which B cereus could be isolated. Similarly, a recent study revealed that 93% of milk samples submitted to holder pasteurisation showed no bacterial growth on cultures and that Bacillus was the predominant contaminant in those that were positive after pasteurisation (34). B cereus is described as being of ubiquitous presence in nature; in addition to a full life cycle in soil, in which it is richly present, it is also adapted to human hosts either as a pathogen or, more frequently, as a part of the intestinal microbiota of a healthy host (24). Additionally, it has been found in breast milk of healthy rhesus monkeys (35) and in the udders of cows (36). The possible adaptation of B cereus to the environment of the animal gut could be the basis of their proposed probiotic effect. In fact, certain strains producing negligible amounts of toxin at 37°C have been approved for probiotic use by the European Food Safety Authority (24); however, because the level of virulence is greatly variable among different strains, caution is strongly required when dealing with this species.

In contrast to vegetative cells, spores of B cereus can survive different heat treatments, including holder pasteurisation. As a consequence, this species is a common inhabitant of milk (36,37) and it can cause a defect known as sweet curdling in dairy products. Considering the nonfastidious nature of this microorganism, no type of food with pH <4.8 can be excluded as a risk of food spoilage or foodborne disease (38). Failure to follow basic food preparation rules, such as slow or inadequate cooling, storage at ambient temperature, or prolonged heat-keeping at approximately 60°C, may allow the growth of B cereus. Therefore, these hygienic rules are critical in a milk bank that provides milk to preterm neonates. It should be understood that a negative result for Bacillus in a postpasteurisation culture does not mean that this microorganism is absent; it only means that this species is under the detection limit of the technique (eg, 100 CFU/mL if 10 μL of milk were cultured).

Two distinct foodborne disease types, emetic and diarrhoeal, are associated with B cereus. For both types, 3 to 8 log10 CFU cells or spores have been indicated as the infective dose (24,38). The count of B cereus in a confirmed foodborne outbreak in Norway was as low as 2 log10 CFU/g of food (39), although further research showed that the actual number was closer to 4 log10 CFU/g, and that the underestimation was because the bacilli were present as aggregated spores (24). Although the role of cereulide in causing the emetic syndrome of B cereus is well established, that of the cytotoxins as aetiological agents of diarrhoeal disease is not so clear. Strong evidence indicates that Hbl, Nhe, and CytK cytotoxins are virulence factors usually involved in B cereus foodborne diarrhoeal disease, but there are difficulties in establishing a single factor as the aetiological agent of gastroenteritis due to this species; this fact reflects that the disease is most probably multifactorial and that a number of additional virulence factors may contribute to overall cellular damage, possibly in a strain-dependent manner. In the present study, no strain harboured genes responsible for the biosynthesis of the emetic toxin, whereas all carried those required for Nhe production. Genes encoding Nhe are, however, now thought to be present in all of the known B cereus group strains (24). In relation to cytK genes, cytK1 could not be detected in any of the 49 strains, but cytK2 was present in approximately 50% of them. Finally, only 1 strain (2%) harboured the complete hblCDA operon. Hbl- and CytK-related genes are present in <50% of randomly sampled strains (40–42). Hbl is a 3-component toxin complex and all 3 components are necessary for maximal biological activity (43,44). Nhe and Hbl toxins could be detected in pure cultures of those strains harbouring the nheABC or the hblCDA operons, respectively, using a qualitative technique. Because the counts of such strains in the positive milk samples were low, however, the actual presence of the toxins in the donors’ milk is extremely improbable. In fact, quantitative measurement of the toxins, directly in the milk samples, would be more relevant from the clinical point of view. B cereus strains isolated from pasteurised milk in the present study do not seem to possess a high virulence potential.

Holder pasteurisation did not significantly modify the concentration of any of the biochemical parameters analysed in the present study. Furosine and lactulose values are used to determine the effects of thermal treatment applied to milk or the addition of reconstituted milk powder to raw, pasteurised, or ultra–high-temperature milk. In the present study, furosine could not be detected in any of the samples, in contrast with levels previously reported in holder and high-temperature (72°C, 15 seconds) pasteurised milks, ranging from 6.9 to 0.0 and from 6.7 to 20.3 mg/100 g protein, respectively (45). This could be caused by the low protein concentration present in human milk compared with cows’ milk. Furosine determination has gained attention from food chemists and biomedical researchers because its formation upon heat treatment is well characterised. Moreover, it represents the Amadori products from early Maillard reactions in which amino acids react with reducing carbohydrates, resulting in a loss of their bioavailability. This is of importance for the essential amino acid lysine, which is also the limiting amino acid in many proteins.

In the pasteurised samples in which lactulose was detected, its content was higher than that found in Spanish pasteurised milks (45), a fact that may be attributed to the high content of lactose present in human milk. The lactulose concentrations were well below the limits considered acceptable for infant formulas. A correlation between lactulose and furosine exists (46,47) because both parameters are influenced by the intensity of the heating process and by the storage conditions (48); however, the concentrations found here do not seem enough to negatively affect protein quality.

No differences were found in the concentrations of lactose and glucose when the fresh samples of donor milk were divided into 3 groups on the basis of the duration of lactation (transition milk, mature milk, late lactation milk). Interestingly, there was a statistically significant decrease in the concentration of myoinositol in the samples as the lactation period increased. This finding is relevant because administration of myoinositol to premature infants with respiratory distress syndrome who are receiving parenteral nutrition is associated with increased survival without bronchopulmonary dysplasia or neural developmental handicap and with a decreased incidence of retinopathy (49). Serum myoinositol concentration increases after birth in premature breast-fed infants, whereas it tends to fall in those receiving parenteral nutrition (49). This reflects the fact that concentrations of myoinositol are significantly higher in human milk than in infant formulas or parenteral nutrition solutions (50). Although the observation of a higher myoinositol concentration in early milk is interesting, the comparison of the 3 lactation periods has limited value because of the reduced number of samples within each group. Therefore, more work is required to establish firm conclusions regarding its influence on preterm health.

Globally, the results of the present study showed that holder pasteurisation led to the destruction of bacteria present initially in donor milk samples, with the exception of some B cereus strains that did not display a high virulence potential; in addition, the thermal treatment did not modify significantly the concentration of furosine and lactulose, 2 compounds that are used as markers for evaluating the extent of heat damage in cows’ milk and infant formulas.

Back to Top | Article Outline


1. Hanson LA. Human milk and host defense: immediate and long-term effects. Acta Paediatr 1999; 88:42–46.
2. Arnold LDW. Using banked donor milk in clinical settings. In: Caldwell K, ed. Reclaiming Breastfeeding in the United States: Protection, Promotion, and Support. Sudbury, MA: Jones & Bartlett; 2002:161–78.
3. Naranyanan I, Prakashil K, Murphy NS, et al. Randomized controlled trial of affect of raw and holder pasteurized human milk and formula supplements on the incidence of neonatal infection. Lancet 1984; 8412:111–113.
4. Lucas A, Morley R, Cole TJ, et al. A randomised multicentre study of human milk versus formula and later development in preterm infants. Arch Dis Chid 1994; 70:141–146.
5. Arslanoglu S, Ziegler EE, Moro GE, et al. Donor human milk in preterm infant feeding: evidence and recommendations. J Perinat Med 2010; 38:347–351.
6. Ford JE, Law BA, Marshall VME, et al. Influence of the heat treatment of human milk on some of its protective constituents. J Pediatr 1977; 91:29–35.
7. Fidler N, Sauerwald TU, Koletzko B, et al. Effects of human milk pasteurization and sterilization on available fat content and fatty acid composition. J Pediatr Gastroenterol Nutr 1998; 27:317–322.
8. Koenig A, Diniz EMD, Barbosa SFC, et al. Immunologic factors in human milk: the effects of gestational age and pasteurization. J Hum Lact 2005; 21:439–443.
9. Tully D, Jones F, Tully MR. Donor milk: what's in it and what's not. J Hum Lact 2001; 17:152–155.
10. Silvestre D, Ferrer E, Gaya J, et al. Available lysine content in human milk: stability during manipulation prior to ingestion. Biofactors 2006; 26:71–79.
11. van Boekel MAJS. Effect of heating on Maillard reactions in milk. Food Chem 1998; 62:403–414.
12. Erbersdobler HF, Somoza V. Forty years of furosine-forty years of using Maillard reaction products as indicators of the nutritional quality of foods. Mol Nutr Food Res 2007; 51:423–430.
13. Corzo N, Olano A. Lactulose as food ingredient. J Sci Food Agric 2009; 89:1987–1990.
14. Delgado T, Corzo N, Santamaría G, et al. Determination of furosine in milk samples by ion-pair reversed phase liquid-chromatography. Chromatographia 1992; 33:374–376.
15. Cattaneo S, Masotti F, Pellegrino L. Liquid infant formulas: technological tools for limiting heat damage. J Agric Food Chem 2009; 57:10689–10694.
16. Martín R, Langa S, Reviriego C, et al. Human milk is a source of lactic acid bacteria for the infant gut. J Pediatr 2003; 143:754–758.
17. Martin R, Jimenez E, Heilig H, et al. Isolation of bifidobacteria from breast milk and assessment of the bifidobacterial population by PCR-denaturing gradient gel electrophoresis and quantitative real-time PCR. Appl Environ Microbiol 2009; 75:965–969.
18. Martín R, Heilig HG, Zoetendal EG, et al. Diversity of the Lactobacillus group in breast milk and vagina of healthy women and potential role in the colonization of the infant gut. J Appl Microbiol 2007; 103:2638–2644.
19. Kullen MJ, Sanozky-Dawes RB, Crowell DC, et al. Use of the DNA sequence of variable regions of the 16S rRNA gene for rapid and accurate identification of bacteria in the Lactobacillus acidophilus complex. J Appl Microbiol 2005; 89:511–516.
20. Rasko DA, Altherr MR, Han CS, et al. Genomics of the Bacillus cereus group of organisms. FEMS Microbiol Rev 2005; 29:303–329.
21. Ash C, Farrow JA, Dorsch M, et al. Comparative analysis of Bacillus anthracis, Bacillus cereus, and related species on the basis of reverse transcriptase sequencing of 16S rRNA. Int J Syst Bacteriol 1991; 41:343–346.
22. Cherif A, Borin S, Rizzi A, et al. Characterization of a repetitive element polymorphism-polymerase chain reaction chromosomal marker that discriminates Bacillus anthracis from related species. J Appl Microbiol 2002; 93:456–462.
23. Ruiz-Barba JL, Maldonado A, Jimenez-Diaz R. Small-scale total DNA extraction from bacteria and yeast for PCR applications. Anal Biochem 2005; 347:333–335.
24. Stenfors Arnesen LP, Fagerlund A, Granum PE. From soil to gut: Bacillus cereus and its food poisoning toxins. FEMS Microbiol Rev 2008; 32:579–606.
25. Wehrle E, Moravek M, Dietrich R, et al. Comparison of multiplex PCR, enzyme immunoassay and cell culture methods for the detection of enterotoxinogenic Bacillus cereus. J Microbiol Meth 2009; 78:265–270.
26. Guinebretiere MH, Fagerlund A, Granum PE, et al. Rapid discrimination of cytK-1 and cytK-2 genes in Bacillus cereus strains by a novel duplex PCR system. FEMS Microbiol Lett 2006; 259:74–80.
27. Resmini P, Pellegrino L. Analysis of food heat damage by direct HPLC of furosine. Int Chromatogr Lab 1991; 6:7–11.
28. Montilla A, Moreno FJ, Olano A. A reliable gas capillary chromatographic determination of lactulose in dairy samples. Chromatographia 2005; 62:311–314.
29. Perez PF, Doré J, Leclerc M, et al. Bacterial imprinting of the neonatal immune system: lessons from maternal cells? Pediatrics 2007; 119:e724–e732.
30. Heikkilä MP. Saris PEJ: inhibition of Staphylococcus aureus by the commensal bacteria of human milk. J Appl Microbiol 2003; 95:471–478.
31. Marín ML, Arroyo R, Jimenez E, et al. Cold storage of human milk: effect on its bacterial composition. J Pediatr Gastroenterol Nutr 2009; 49:343–348.
32. Boo N-Y, Nordiah AJ, Alfizah H, et al. Contamination of breast milk obtained by manual expression and breast pumps in mothers of very low birthweight infants. J Hosp Infect 2001; 49:274–281.
33. Brown SL, Bright RS, Dwyer DE, et al. Breast pump adverse events: reports to the Food and Drug Administration. J Hum Lact 2005; 21:169–174.
34. Landers S, Updegrove K. Bacteriological screening of donor human milk before and after Holder pasteurization. Breastfeed Med 2010; 5:117–121.
35. Jin L, Hinde K, Tao L. Species diversity and relative abundance of lactic acid bacteria in the milk of rhesus monkeys (Macaca mulatta). J Med Primatol 2011; 40:52–58.
36. Andersson A, Rönner U, Granum PE. What problems does the food industry have with the spore-forming pathogens Bacillus cereus and Clostridium perfringens? Int J Food Microbiol 1995; 28:145–155.
37. Lin S, Schraft H, Odumeru JA, et al. Identification of contamination sources of Bacillus cereus in pasteurized milk. Int J Food Microbiol 1998; 43:159–171.
38. Gilbert RJ, Kramer JM. Bacillus cereus food poisoning. In: Cliver DC, Cochrane BA, eds. Progress in Food Safety (Proceedings of Symposium). Madison: Food Research Institute, University of Wisconsin-Madison; 1986:85–93.
39. Granum PE. Bacillus cereus som næringsmiddelhygienisk problemorganisme (in Norwegian). Norsk Veterinærtidsk 1994; 106:911–915.
40. Ehling-Schulz M, Svensson B, Guinebretiere MH, et al. Emetic toxin formation of Bacillus cereus is restricted to a single evolutionary lineage of closely related strains. Microbiology 2005; 151:183–197.
41. Ehling-Schulz M, Vukov N, Schulz A, et al. Identification and partial characterization of the nonribosomal peptide synthetase gene responsible for cereulide production in emetic Bacillus cereus. Appl Environ Microbiol 2005; 71:105–113.
42. Moravek M, Dietrich R, Buerk C, et al. Determination of the toxic potential of Bacillus cereus isolates by quantitative enterotoxin analyses. FEMS Microbiol Lett 2006; 257:293–298.
43. Beecher DJ, Schoeni JL, Wong AC. Enterotoxic activity of hemolysin BL from Bacillus cereus. Infect Immun 1995; 63:4423–4428.
44. Lindbäck T, Fagerlund A, Rødland MS, et al. Characterization of the Bacillus cereus Nhe enterotoxin. Microbiology 2004; 150:3959–3967.
45. Villamiel M, Arias M, Corzo N, et al. Use of different thermal indices to assess the quality of pasteurized milks. Z Lebensm Unters Forsch A 1999; 208:169–171.
46. Pellegrino L, De Noni I, Resmini P. Coupling of lactulose and furosine indices for quality evaluation of sterilized milk. Int Dairy J 1995; 5:647–659.
47. Montilla A, Calvo MM, et al. Correlation between lactulose and furosine in UHT heated milk. J Food Prot 1996; 59:1061–1064.
48. Garcia-Baños JL, del Castillo MD, Sanz ML, et al. Maillard reaction during storage of powder enteral formulas. Food Chem 2005; 89:555–560.
49. Hallman M, Bry K, Hoppu K, et al. Inositol supplementation in premature infants with respiratory distress syndrome. N Engl J Med 1992; 326:1233–1239.
50. Cavalli C, Teng C, Battaglia FC, et al. Free sugar and sugar alcohol concentrations in human breast milk. J Pediatr Gastroenterol Nutr 2006; 42:215–221.

Cited By:

This article has been cited 1 time(s).

Journal of Pediatric Gastroenterology and Nutrition
Heat Treatment of Human Milk
Moro, GE; Arslanoglu, S
Journal of Pediatric Gastroenterology and Nutrition, 54(2): 165-166.
PDF (82) | CrossRef
Back to Top | Article Outline

Bacillus cereus; furosine; human milk; lactulose; milk bank; myoinositol; pasteurisation

Copyright 2012 by ESPGHAN and NASPGHAN