Date of Award


Degree Name

Master of Philosophy (School of Health Sciences)

Schools and Centres

Health Sciences

First Supervisor

Professor Naomi Trengove

Second Supervisor

Associate Professor Robert Trengove

Third Supervisor

Dr Joel Gummer


Background: Human milk is considered the best source of nutrition for all newborns as it contains important growth, developmental and immunological factors. The WHO (2003) recommends exclusive breastfeeding for the first six months of age, with complementary breastfeeding up to two years and beyond. However, some women experience complications of the breast that lead to early cessation of breastfeeding, which can adversely affect the well being of the developing infant and her own health.

Nipple pain is the most commonly cited reason for weaning in the first week postpartum. Nipple pain is also linked to mastitis from milk stasis and possible bacterial infection, although the influence of bacteria is still largely unknown. However, it is known that the presence of bacteria and fungi along with their metabolites contribute to the composition of the milk as the baby receives it.

Metabolomics is increasingly being utilised in the dairy industry to determine spoilage as a result of teat trauma and mastitis. Given the current diagnostic application of metabolomics in clinical medicine uses blood and urine samples, it has been proposed as a potential tool for detecting biomarkers and determining compositional changes in human milk. Measuring the composition of milk from human mothers experiencing persistent nipple pain, with or without evidence of trauma, and identifying the influence of this condition on endogenous and exogenous metabolites may determine the relationship between milk composition and nipple pain.

Aims: The aims of this study were to source the appropriate human and bovine milk samples; to identify and quantify bacterial and fungal species using traditional culture and microscopy techniques; to measure the effect of nipple pain on the paracellular pathway of the breast by measuring the sodium and potassium concentration and ratio in the milk; to optimise GC-MS methodology for the measurement of milk metabolites; and to use untargeted metabolomics to identify compositional differences in the metabolite profile in human milk from mothers presenting with nipple pain compared to healthy control mothers.

Results: Two groups were recruited; a control group of mothers not experiencing nipple pain (n=22 samples) and a group of mothers experiencing persistent nipple pain during breastfeeding (n=11 samples); mothers with unilateral nipple pain supplied a milk sample from their affected and non-affected breast (n=4). The nipple pain group (n=11) was divided into two subgroups; persistent nipple pain without evidence of trauma (PG) (n=6) and persistent nipple pain with evidence of trauma (TG) (n=5). Additionally 9 bovine samples were collected, 3 from healthy cows (control), 4 from cows presenting with mastitis and 2 from a single storage vat, to be used as positive controls throughout the study.

All 42 samples were tested for the presence of microbial and fungal species, sodium and potassium concentrations and ratio were determined and untargeted metabolomics analysis of the milk metabolome was performed.

Overall there was no significant difference in microbe content between the human control and nipple pain group (1, 623 CFU/ml vs. 1, 503 CFU/ml); the TG subgroup had the highest colony count of 2, 778 CFU/ml. The bovine mastitis group had a higher colony count than the bovine control group, 2, 173 CFU/ml vs. 473 CFU/ml. Coagulase negative staphylococcus ssp. were the most frequently isolated microorganisms and was found in 91% of human milk samples and 100% of bovine milk samples. Staphylococcus aureus were identified in one human milk sample from a mother in the PG subgroup and in one bovine sample from a cow suffering from untreated mastitis as well as both pooled bovine vat samples. Streptococcus ssp. And yeast were only found in bovine samples.

The TG subgroup had the highest Na+ concentration of the human milk samples (8.04 ± 2.40 mM), significantly highly than the control group (4.32 ± 1.18 mM; p

Untargeted metabolomic analysis found compositional differences between the human control and nipple pain groups, in particular samples from the TG subgroup. Compositional variations between milk from the control and nipple pain subgroups was identified using principal component analysis and PC4 best represented the differences in metabolite composition between the groups. This result is consistent with the subtlety of the nipple pain condition. A list of the most influential metabolites based on their correlation loadings (explained within 50-100% of the model) was determined. The most influential metabolites with respect to the TG milk samples were included isoleucine, proline, galactose and some as yet unidentified metabolites.

Conclusion: As nipple pain is often a precursor to mastitis the results from this study will form a basis for further development using metabolomics as a tool for more efficient detection and treatment of breast infection and inflammation within the nipple and breast.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."